Synth-diy-test

> My first modular synth was a Paia 4700/j synth. Yes I know, cheesy design, but the one thing I miss is how linear response VCOs have a very different detuned chorusing effect with each interval. Just wondering if anyone has experimented with this?
> ________________________________________________________
> This is the Synth-diy mailing list
> Submit email to: Synth-diy@synth-diy.org
> View archive at: https://synth-diy.org/pipermail/synth-diy/
> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
> Selling or trading? Use marketplace@synth-diy.org

> On 16 Mar 2026, at 11:36, Roman Sowa via Synth-diy <synth-diy@synth-diy.org> wrote:
> 
> IMHO linear response VCOs are simply better and easier to use. No temperature dependency by design and easier to tune - as long as all offsets are small or trimmed, whatever you do, an octave is always an octave, not like with exp VCO where you have to turn 3 to 5 trimmers in several iterations to get them sound remotely acceptable.
> 
> One may say that summing few CV inputs is difficult for linear VCOs, but today the VCOs are controlled by microcontroller anyway, so the problem of the 1970's is no longer any concern.
> 
> Roman
> 
> W dniu 2026-03-16 o 03:49, Thomas Hudson via Synth-diy pisze:
>> My first modular synth was a Paia 4700/j synth. Yes I know, cheesy design, but the one thing I miss is how linear response VCOs have a very different detuned chorusing effect with each interval. Just wondering if anyone has experimented with this?
>> ________________________________________________________
>> This is the Synth-diy mailing list
>> Submit email to: Synth-diy@synth-diy.org
>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>> Selling or trading? Use marketplace@synth-diy.org
> ________________________________________________________
> This is the Synth-diy mailing list
> Submit email to: Synth-diy@synth-diy.org
> View archive at: https://synth-diy.org/pipermail/synth-diy/
> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
> Selling or trading? Use marketplace@synth-diy.org

> That's a thought-provoking point of view, Roman.
> 
> I thought the fundamental problem with linear VCOs was that it's hard to cover an 8 or even ten octave range in a linear fashion - that's 256 or 1024 times the base value. For a CV, that requires very small offsets at the bottom. Say we're going with 1V/KHz. We have a maximum CV somewhere up around 10 or 12V giving us an impressive 10 or 12KHz maximum frequency. But we have an system of 1mV=1Hz, so errors will need to be in the sub-mV level at the low end to keep things sounding in tune.
> 
> uPs and DACs do help in some respects, and since DACs are always always linear these days, it's kind-of an obvious way to go. But again, the accuracy requirement bites at the bottom end. Since the width of each octave doubles every time, if I have 1024 steps in a low octave, then I have 262144 steps in an octave eight octaves higher! That's a massive waste of DAC resolution, and eats up DAC accuracy in a way that makes finding a suitable device a lot harder and more expensive.
> 
> Still, you've given me something to think about, and I shall go and do some calculations to see how well this might work. Thanks.
> 
> Tom
> 
> 
>> On 16 Mar 2026, at 11:36, Roman Sowa via Synth-diy <synth-diy@synth-diy.org> wrote:
>>
>> IMHO linear response VCOs are simply better and easier to use. No temperature dependency by design and easier to tune - as long as all offsets are small or trimmed, whatever you do, an octave is always an octave, not like with exp VCO where you have to turn 3 to 5 trimmers in several iterations to get them sound remotely acceptable.
>>
>> One may say that summing few CV inputs is difficult for linear VCOs, but today the VCOs are controlled by microcontroller anyway, so the problem of the 1970's is no longer any concern.
>>
>> Roman
>>
>> W dniu 2026-03-16 o 03:49, Thomas Hudson via Synth-diy pisze:
>>> My first modular synth was a Paia 4700/j synth. Yes I know, cheesy design, but the one thing I miss is how linear response VCOs have a very different detuned chorusing effect with each interval. Just wondering if anyone has experimented with this?
>>> ________________________________________________________
>>> This is the Synth-diy mailing list
>>> Submit email to: Synth-diy@synth-diy.org
>>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>>> Selling or trading? Use marketplace@synth-diy.org
>> ________________________________________________________
>> This is the Synth-diy mailing list
>> Submit email to: Synth-diy@synth-diy.org
>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>> Selling or trading? Use marketplace@synth-diy.org
>

> On Mar 16, 2026, at 12:01 PM, Roman Sowa via Synth-diy <synth-diy@synth-diy.org> wrote:
> 
> Your point is valid, but rises fundamental question:
> who needs 8 or 10 octave continuous tuning range anyway?
> No analog synth has more than 5 and today most popular is 3-4.
> That means smaller range is enough and you can add octave switch as separate control to cover all frequencies of interesrt. Just like a knob doing that in any synth basicaly.
> And octave switch in DAC controled Hz/V VCO is just switching the resistance on the way from DAMC to VCO core. This is also exactly how it was done in vintage linear VCO synths like MS20 and many more.
> And tuning it further down with modulation or pitchbend doesn't require much of a precicion, especially when it happens at so low frequencies where ear pitch perception is rather limited.
> 
> Roman
> 
> W dniu 2026-03-16 o 16:13, Tom Wiltshire pisze:
>> That's a thought-provoking point of view, Roman.
>> I thought the fundamental problem with linear VCOs was that it's hard to cover an 8 or even ten octave range in a linear fashion - that's 256 or 1024 times the base value. For a CV, that requires very small offsets at the bottom. Say we're going with 1V/KHz. We have a maximum CV somewhere up around 10 or 12V giving us an impressive 10 or 12KHz maximum frequency. But we have an system of 1mV=1Hz, so errors will need to be in the sub-mV level at the low end to keep things sounding in tune.
>> uPs and DACs do help in some respects, and since DACs are always always linear these days, it's kind-of an obvious way to go. But again, the accuracy requirement bites at the bottom end. Since the width of each octave doubles every time, if I have 1024 steps in a low octave, then I have 262144 steps in an octave eight octaves higher! That's a massive waste of DAC resolution, and eats up DAC accuracy in a way that makes finding a suitable device a lot harder and more expensive.
>> Still, you've given me something to think about, and I shall go and do some calculations to see how well this might work. Thanks.
>> Tom
>>> On 16 Mar 2026, at 11:36, Roman Sowa via Synth-diy <synth-diy@synth-diy.org> wrote:
>>> 
>>> IMHO linear response VCOs are simply better and easier to use. No temperature dependency by design and easier to tune - as long as all offsets are small or trimmed, whatever you do, an octave is always an octave, not like with exp VCO where you have to turn 3 to 5 trimmers in several iterations to get them sound remotely acceptable.
>>> 
>>> One may say that summing few CV inputs is difficult for linear VCOs, but today the VCOs are controlled by microcontroller anyway, so the problem of the 1970's is no longer any concern.
>>> 
>>> Roman
>>> 
>>> W dniu 2026-03-16 o 03:49, Thomas Hudson via Synth-diy pisze:
>>>> My first modular synth was a Paia 4700/j synth. Yes I know, cheesy design, but the one thing I miss is how linear response VCOs have a very different detuned chorusing effect with each interval. Just wondering if anyone has experimented with this?
>>>> ________________________________________________________
>>>> This is the Synth-diy mailing list
>>>> Submit email to: Synth-diy@synth-diy.org
>>>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>>>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>>>> Selling or trading? Use marketplace@synth-diy.org
>>> ________________________________________________________
>>> This is the Synth-diy mailing list
>>> Submit email to: Synth-diy@synth-diy.org
>>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>>> Selling or trading? Use marketplace@synth-diy.org
> ________________________________________________________
> This is the Synth-diy mailing list
> Submit email to: Synth-diy@synth-diy.org
> View archive at: https://synth-diy.org/pipermail/synth-diy/
> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
> Selling or trading? Use marketplace@synth-diy.org

> On Mar 16, 2026, at 6:11 PM, Adam (synthDIY) <synthdiy@adambaby.com> wrote:
> 
> 
> 
>> On 17 Mar 2026, at 02:01, Roman Sowa via Synth-diy <synth-diy@synth-diy.org> wrote:
>> 
>> who needs 8 or 10 octave continuous tuning range anyway?
>> No analog synth has more than 5 and today most popular is 3-4.
>> That means smaller range is enough and you can add octave switch as separate control to cover all frequencies of interesrt. Just like a knob doing that in any synth basicaly.
>> And octave switch in DAC controled Hz/V VCO is just switching the resistance on the way from DAMC to VCO core. This is also exactly how it was done in vintage linear VCO synths like MS20 and many more.
> 
> 
> Agree. 
> Musically, in any given piece, many instruments are active in just the range of an octave or two.
> The most characterful-sounding synths I have here are linear and are limited to about 4 octaves or less via Hz/volt external control (Korg 700, Roland SH-2000, TB303!)
> 
> A
> ________________________________________________________
> This is the Synth-diy mailing list
> Submit email to: Synth-diy@synth-diy.org
> View archive at: https://synth-diy.org/pipermail/synth-diy/
> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
> Selling or trading? Use marketplace@synth-diy.org

From: Synth-diy <synth-diy-bounces@synth-diy.org> on behalf of Adam (synthDIY) <synthdiy@adambaby.com>
Sent: 16 March 2026 22:11
To: Roman Sowa <modular@go2.pl>
Cc: synth-diy@synth-diy.org <synth-diy@synth-diy.org>
Subject: Re: [sdiy] Linear response VCOs?



On 17 Mar 2026, at 02:01, Roman Sowa via Synth-diy <synth-diy@synth-diy.org> wrote:

who needs 8 or 10 octave continuous tuning range anyway?
No analog synth has more than 5 and today most popular is 3-4.
That means smaller range is enough and you can add octave switch as separate control to cover all frequencies of interesrt. Just like a knob doing that in any synth basicaly.
And octave switch in DAC controled Hz/V VCO is just switching the resistance on the way from DAMC to VCO core. This is also exactly how it was done in vintage linear VCO synths like MS20 and many more.


Agree.
Musically, in any given piece, many instruments are active in just the range of an octave or two.
The most characterful-sounding synths I have here are linear and are limited to about 4 octaves or less via Hz/volt external control (Korg 700, Roland SH-2000, TB303!)

A

> One more question. I have a couple of vcos based on the cem3340, which as a linear FM input. Assuming I could generate an appropriate voltage source, could I drive them through this input?
> 
>> On Mar 16, 2026, at 12:01 PM, Roman Sowa via Synth-diy <synth-diy@synth-diy.org> wrote:
>>
>> Your point is valid, but rises fundamental question:
>> who needs 8 or 10 octave continuous tuning range anyway?
>> No analog synth has more than 5 and today most popular is 3-4.
>> That means smaller range is enough and you can add octave switch as separate control to cover all frequencies of interesrt. Just like a knob doing that in any synth basicaly.
>> And octave switch in DAC controled Hz/V VCO is just switching the resistance on the way from DAMC to VCO core. This is also exactly how it was done in vintage linear VCO synths like MS20 and many more.
>> And tuning it further down with modulation or pitchbend doesn't require much of a precicion, especially when it happens at so low frequencies where ear pitch perception is rather limited.
>>
>> Roman
>>
>> W dniu 2026-03-16 o 16:13, Tom Wiltshire pisze:
>>> That's a thought-provoking point of view, Roman.
>>> I thought the fundamental problem with linear VCOs was that it's hard to cover an 8 or even ten octave range in a linear fashion - that's 256 or 1024 times the base value. For a CV, that requires very small offsets at the bottom. Say we're going with 1V/KHz. We have a maximum CV somewhere up around 10 or 12V giving us an impressive 10 or 12KHz maximum frequency. But we have an system of 1mV=1Hz, so errors will need to be in the sub-mV level at the low end to keep things sounding in tune.
>>> uPs and DACs do help in some respects, and since DACs are always always linear these days, it's kind-of an obvious way to go. But again, the accuracy requirement bites at the bottom end. Since the width of each octave doubles every time, if I have 1024 steps in a low octave, then I have 262144 steps in an octave eight octaves higher! That's a massive waste of DAC resolution, and eats up DAC accuracy in a way that makes finding a suitable device a lot harder and more expensive.
>>> Still, you've given me something to think about, and I shall go and do some calculations to see how well this might work. Thanks.
>>> Tom
>>>> On 16 Mar 2026, at 11:36, Roman Sowa via Synth-diy <synth-diy@synth-diy.org> wrote:
>>>>
>>>> IMHO linear response VCOs are simply better and easier to use. No temperature dependency by design and easier to tune - as long as all offsets are small or trimmed, whatever you do, an octave is always an octave, not like with exp VCO where you have to turn 3 to 5 trimmers in several iterations to get them sound remotely acceptable.
>>>>
>>>> One may say that summing few CV inputs is difficult for linear VCOs, but today the VCOs are controlled by microcontroller anyway, so the problem of the 1970's is no longer any concern.
>>>>
>>>> Roman
>>>>
>>>> W dniu 2026-03-16 o 03:49, Thomas Hudson via Synth-diy pisze:
>>>>> My first modular synth was a Paia 4700/j synth. Yes I know, cheesy design, but the one thing I miss is how linear response VCOs have a very different detuned chorusing effect with each interval. Just wondering if anyone has experimented with this?
>>>>> ________________________________________________________
>>>>> This is the Synth-diy mailing list
>>>>> Submit email to: Synth-diy@synth-diy.org
>>>>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>>>>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>>>>> Selling or trading? Use marketplace@synth-diy.org
>>>> ________________________________________________________
>>>> This is the Synth-diy mailing list
>>>> Submit email to: Synth-diy@synth-diy.org
>>>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>>>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>>>> Selling or trading? Use marketplace@synth-diy.org
>> ________________________________________________________
>> This is the Synth-diy mailing list
>> Submit email to: Synth-diy@synth-diy.org
>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>> Selling or trading? Use marketplace@synth-diy.org
>

> On 16 Mar 2026, at 16:01, Roman Sowa <modular@go2.pl> wrote:
> 
> Your point is valid, but rises fundamental question:
> who needs 8 or 10 octave continuous tuning range anyway?
> No analog synth has more than 5 and today most popular is 3-4.
> That means smaller range is enough and you can add octave switch as separate control to cover all frequencies of interesrt. Just like a knob doing that in any synth basicaly.
> And octave switch in DAC controled Hz/V VCO is just switching the resistance on the way from DAMC to VCO core. This is also exactly how it was done in vintage linear VCO synths like MS20 and many more.
> And tuning it further down with modulation or pitchbend doesn't require much of a precicion, especially when it happens at so low frequencies where ear pitch perception is rather limited.
> 
> Roman
> 
> W dniu 2026-03-16 o 16:13, Tom Wiltshire pisze:
>> That's a thought-provoking point of view, Roman.
>> I thought the fundamental problem with linear VCOs was that it's hard to cover an 8 or even ten octave range in a linear fashion - that's 256 or 1024 times the base value. For a CV, that requires very small offsets at the bottom. Say we're going with 1V/KHz. We have a maximum CV somewhere up around 10 or 12V giving us an impressive 10 or 12KHz maximum frequency. But we have an system of 1mV=1Hz, so errors will need to be in the sub-mV level at the low end to keep things sounding in tune.
>> uPs and DACs do help in some respects, and since DACs are always always linear these days, it's kind-of an obvious way to go. But again, the accuracy requirement bites at the bottom end. Since the width of each octave doubles every time, if I have 1024 steps in a low octave, then I have 262144 steps in an octave eight octaves higher! That's a massive waste of DAC resolution, and eats up DAC accuracy in a way that makes finding a suitable device a lot harder and more expensive.
>> Still, you've given me something to think about, and I shall go and do some calculations to see how well this might work. Thanks.
>> Tom
>>> On 16 Mar 2026, at 11:36, Roman Sowa via Synth-diy <synth-diy@synth-diy.org> wrote:
>>> 
>>> IMHO linear response VCOs are simply better and easier to use. No temperature dependency by design and easier to tune - as long as all offsets are small or trimmed, whatever you do, an octave is always an octave, not like with exp VCO where you have to turn 3 to 5 trimmers in several iterations to get them sound remotely acceptable.
>>> 
>>> One may say that summing few CV inputs is difficult for linear VCOs, but today the VCOs are controlled by microcontroller anyway, so the problem of the 1970's is no longer any concern.
>>> 
>>> Roman
>>> 
>>> W dniu 2026-03-16 o 03:49, Thomas Hudson via Synth-diy pisze:
>>>> My first modular synth was a Paia 4700/j synth. Yes I know, cheesy design, but the one thing I miss is how linear response VCOs have a very different detuned chorusing effect with each interval. Just wondering if anyone has experimented with this?
>>>> ________________________________________________________
>>>> This is the Synth-diy mailing list
>>>> Submit email to: Synth-diy@synth-diy.org
>>>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>>>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>>>> Selling or trading? Use marketplace@synth-diy.org
>>> ________________________________________________________
>>> This is the Synth-diy mailing list
>>> Submit email to: Synth-diy@synth-diy.org
>>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>>> Selling or trading? Use marketplace@synth-diy.org

On 3/17/26 4:12 AM, Tom Wiltshire wrote:
> While you might not need 8 or 10 octaves at once to play a melody, many V/Oct analog synths *do* have a lot more than five as a continuous range. Arguably that's a problem, since it makes the tuning issues worse and it necessitated the High Frequency Trimmer, but that *is* how it's been done many times. And that is probably thanks to the 3340 and its temperature compensated expo convertor, as you say!
> The Sequential Pro-One is one example. It has a three octave keyboard and an octave switch, and all the octave switch does is feed nicely calibrated doses of 1V to the oscillators to switch the octave.
>
> We certainly both agree that trying to cover less range in one go makes life technically simpler for an oscillator designer, almost whatever oscillator design is used. That would be true for a V/Oct oscillator too. If we only have to cover 4 octaves in tune, then our expo convertor has a lot less to do and scale errors are much less of an issue. That's not specific to linear oscillators. Another example where this sort of range switching makes life easier and hides the limitations is the Juno 106's divider-based DCOs. Without range switching, there wouldn't be enough divider resolution to get really low notes, and the stepping at the high end would be even worse than it is already!
>
> For a linear oscillator run from a DAC, the errors would be worse for the lowest frequencies where you have the fewest DAC steps. As you say, our pitch perception is pretty poor down there, so as long as you've got plenty of resolution by the time the oscillator has got up to the sensitive midrange frequencies, you can probably afford to lose some further down. This definitely helps, I agree!
>
> Tom
>
>
>> On 16 Mar 2026, at 16:01, Roman Sowa <modular@go2.pl> wrote:
>>
>> Your point is valid, but rises fundamental question:
>> who needs 8 or 10 octave continuous tuning range anyway?
>> No analog synth has more than 5 and today most popular is 3-4.
>> That means smaller range is enough and you can add octave switch as separate control to cover all frequencies of interesrt. Just like a knob doing that in any synth basicaly.
>> And octave switch in DAC controled Hz/V VCO is just switching the resistance on the way from DAMC to VCO core. This is also exactly how it was done in vintage linear VCO synths like MS20 and many more.
>> And tuning it further down with modulation or pitchbend doesn't require much of a precicion, especially when it happens at so low frequencies where ear pitch perception is rather limited.
>>
>> Roman
>>
>> W dniu 2026-03-16 o 16:13, Tom Wiltshire pisze:
>>> That's a thought-provoking point of view, Roman.
>>> I thought the fundamental problem with linear VCOs was that it's hard to cover an 8 or even ten octave range in a linear fashion - that's 256 or 1024 times the base value. For a CV, that requires very small offsets at the bottom. Say we're going with 1V/KHz. We have a maximum CV somewhere up around 10 or 12V giving us an impressive 10 or 12KHz maximum frequency. But we have an system of 1mV=1Hz, so errors will need to be in the sub-mV level at the low end to keep things sounding in tune.
>>> uPs and DACs do help in some respects, and since DACs are always always linear these days, it's kind-of an obvious way to go. But again, the accuracy requirement bites at the bottom end. Since the width of each octave doubles every time, if I have 1024 steps in a low octave, then I have 262144 steps in an octave eight octaves higher! That's a massive waste of DAC resolution, and eats up DAC accuracy in a way that makes finding a suitable device a lot harder and more expensive.
>>> Still, you've given me something to think about, and I shall go and do some calculations to see how well this might work. Thanks.
>>> Tom
>>>> On 16 Mar 2026, at 11:36, Roman Sowa via Synth-diy <synth-diy@synth-diy.org> wrote:
>>>>
>>>> IMHO linear response VCOs are simply better and easier to use. No temperature dependency by design and easier to tune - as long as all offsets are small or trimmed, whatever you do, an octave is always an octave, not like with exp VCO where you have to turn 3 to 5 trimmers in several iterations to get them sound remotely acceptable.
>>>>
>>>> One may say that summing few CV inputs is difficult for linear VCOs, but today the VCOs are controlled by microcontroller anyway, so the problem of the 1970's is no longer any concern.
>>>>
>>>> Roman
>>>>
>>>> W dniu 2026-03-16 o 03:49, Thomas Hudson via Synth-diy pisze:
>>>>> My first modular synth was a Paia 4700/j synth. Yes I know, cheesy design, but the one thing I miss is how linear response VCOs have a very different detuned chorusing effect with each interval. Just wondering if anyone has experimented with this?
>>>>> ________________________________________________________
>>>>> This is the Synth-diy mailing list
>>>>> Submit email to: Synth-diy@synth-diy.org
>>>>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>>>>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>>>>> Selling or trading? Use marketplace@synth-diy.org
>>>> ________________________________________________________
>>>> This is the Synth-diy mailing list
>>>> Submit email to: Synth-diy@synth-diy.org
>>>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>>>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>>>> Selling or trading? Use marketplace@synth-diy.org
>
> ________________________________________________________
> This is the Synth-diy mailing list
> Submit email to: Synth-diy@synth-diy.org
> View archive at: https://synth-diy.org/pipermail/synth-diy/
> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
> Selling or trading? Use marketplace@synth-diy.org
>

> On Mar 17, 2026, at 1:44 PM, David Manley via Synth-diy <synth-diy@synth-diy.org> wrote:
> 
> It's interesting to see how PAiA's John Simonton solved some these issues in the 1970's by having a custom laser trimmed resistor network built for their 6-bit "Equally Tempered DAC" to be used with linear VCOs.  See the bottom of the schematic on page 18, the resistor values are on the last page.
> 
> https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf
> 
> As is typical for PAiA a very low cost solution: build your own DAC with a few components.
> 
> -Dave
> 
> On 3/17/26 4:12 AM, Tom Wiltshire wrote:
>> While you might not need 8 or 10 octaves at once to play a melody, many V/Oct analog synths *do* have a lot more than five as a continuous range. Arguably that's a problem, since it makes the tuning issues worse and it necessitated the High Frequency Trimmer, but that *is* how it's been done many times. And that is probably thanks to the 3340 and its temperature compensated expo convertor, as you say!
>> The Sequential Pro-One is one example. It has a three octave keyboard and an octave switch, and all the octave switch does is feed nicely calibrated doses of 1V to the oscillators to switch the octave.
>> 
>> We certainly both agree that trying to cover less range in one go makes life technically simpler for an oscillator designer, almost whatever oscillator design is used. That would be true for a V/Oct oscillator too. If we only have to cover 4 octaves in tune, then our expo convertor has a lot less to do and scale errors are much less of an issue. That's not specific to linear oscillators. Another example where this sort of range switching makes life easier and hides the limitations is the Juno 106's divider-based DCOs. Without range switching, there wouldn't be enough divider resolution to get really low notes, and the stepping at the high end would be even worse than it is already!
>> 
>> For a linear oscillator run from a DAC, the errors would be worse for the lowest frequencies where you have the fewest DAC steps. As you say, our pitch perception is pretty poor down there, so as long as you've got plenty of resolution by the time the oscillator has got up to the sensitive midrange frequencies, you can probably afford to lose some further down. This definitely helps, I agree!
>> 
>> Tom
>> 
>> 
>>> On 16 Mar 2026, at 16:01, Roman Sowa <modular@go2.pl> wrote:
>>> 
>>> Your point is valid, but rises fundamental question:
>>> who needs 8 or 10 octave continuous tuning range anyway?
>>> No analog synth has more than 5 and today most popular is 3-4.
>>> That means smaller range is enough and you can add octave switch as separate control to cover all frequencies of interesrt. Just like a knob doing that in any synth basicaly.
>>> And octave switch in DAC controled Hz/V VCO is just switching the resistance on the way from DAMC to VCO core. This is also exactly how it was done in vintage linear VCO synths like MS20 and many more.
>>> And tuning it further down with modulation or pitchbend doesn't require much of a precicion, especially when it happens at so low frequencies where ear pitch perception is rather limited.
>>> 
>>> Roman
>>> 
>>> W dniu 2026-03-16 o 16:13, Tom Wiltshire pisze:
>>>> That's a thought-provoking point of view, Roman.
>>>> I thought the fundamental problem with linear VCOs was that it's hard to cover an 8 or even ten octave range in a linear fashion - that's 256 or 1024 times the base value. For a CV, that requires very small offsets at the bottom. Say we're going with 1V/KHz. We have a maximum CV somewhere up around 10 or 12V giving us an impressive 10 or 12KHz maximum frequency. But we have an system of 1mV=1Hz, so errors will need to be in the sub-mV level at the low end to keep things sounding in tune.
>>>> uPs and DACs do help in some respects, and since DACs are always always linear these days, it's kind-of an obvious way to go. But again, the accuracy requirement bites at the bottom end. Since the width of each octave doubles every time, if I have 1024 steps in a low octave, then I have 262144 steps in an octave eight octaves higher! That's a massive waste of DAC resolution, and eats up DAC accuracy in a way that makes finding a suitable device a lot harder and more expensive.
>>>> Still, you've given me something to think about, and I shall go and do some calculations to see how well this might work. Thanks.
>>>> Tom
>>>>> On 16 Mar 2026, at 11:36, Roman Sowa via Synth-diy <synth-diy@synth-diy.org> wrote:
>>>>> 
>>>>> IMHO linear response VCOs are simply better and easier to use. No temperature dependency by design and easier to tune - as long as all offsets are small or trimmed, whatever you do, an octave is always an octave, not like with exp VCO where you have to turn 3 to 5 trimmers in several iterations to get them sound remotely acceptable.
>>>>> 
>>>>> One may say that summing few CV inputs is difficult for linear VCOs, but today the VCOs are controlled by microcontroller anyway, so the problem of the 1970's is no longer any concern.
>>>>> 
>>>>> Roman
>>>>> 
>>>>> W dniu 2026-03-16 o 03:49, Thomas Hudson via Synth-diy pisze:
>>>>>> My first modular synth was a Paia 4700/j synth. Yes I know, cheesy design, but the one thing I miss is how linear response VCOs have a very different detuned chorusing effect with each interval. Just wondering if anyone has experimented with this?
>>>>>> ________________________________________________________
>>>>>> This is the Synth-diy mailing list
>>>>>> Submit email to: Synth-diy@synth-diy.org
>>>>>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>>>>>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>>>>>> Selling or trading? Use marketplace@synth-diy.org
>>>>> ________________________________________________________
>>>>> This is the Synth-diy mailing list
>>>>> Submit email to: Synth-diy@synth-diy.org
>>>>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>>>>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>>>>> Selling or trading? Use marketplace@synth-diy.org
>> 
>> ________________________________________________________
>> This is the Synth-diy mailing list
>> Submit email to: Synth-diy@synth-diy.org
>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>> Selling or trading? Use marketplace@synth-diy.org
>> 
> ________________________________________________________
> This is the Synth-diy mailing list
> Submit email to: Synth-diy@synth-diy.org
> View archive at: https://synth-diy.org/pipermail/synth-diy/
> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
> Selling or trading? Use marketplace@synth-diy.org

> Unfortunately this solution didn't work. I built this synth and he had to add three to four trimmers, all of them interacting to tune the thing. It was doable and I was able to get tracking pretty good. Perhaps things have gotten better, but from experience it took some tweaking.
>
>> On Mar 17, 2026, at 1:44 PM, David Manley via Synth-diy <synth-diy@synth-diy.org> wrote:
>>
>> It's interesting to see how PAiA's John Simonton solved some these issues in the 1970's by having a custom laser trimmed resistor network built for their 6-bit "Equally Tempered DAC" to be used with linear VCOs.  See the bottom of the schematic on page 18, the resistor values are on the last page.
>>
>> https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf
>>
>> As is typical for PAiA a very low cost solution: build your own DAC with a few components.
>>
>> -Dave
>>
>> On 3/17/26 4:12 AM, Tom Wiltshire wrote:
>>> While you might not need 8 or 10 octaves at once to play a melody, many V/Oct analog synths *do* have a lot more than five as a continuous range. Arguably that's a problem, since it makes the tuning issues worse and it necessitated the High Frequency Trimmer, but that *is* how it's been done many times. And that is probably thanks to the 3340 and its temperature compensated expo convertor, as you say!
>>> The Sequential Pro-One is one example. It has a three octave keyboard and an octave switch, and all the octave switch does is feed nicely calibrated doses of 1V to the oscillators to switch the octave.
>>>
>>> We certainly both agree that trying to cover less range in one go makes life technically simpler for an oscillator designer, almost whatever oscillator design is used. That would be true for a V/Oct oscillator too. If we only have to cover 4 octaves in tune, then our expo convertor has a lot less to do and scale errors are much less of an issue. That's not specific to linear oscillators. Another example where this sort of range switching makes life easier and hides the limitations is the Juno 106's divider-based DCOs. Without range switching, there wouldn't be enough divider resolution to get really low notes, and the stepping at the high end would be even worse than it is already!
>>>
>>> For a linear oscillator run from a DAC, the errors would be worse for the lowest frequencies where you have the fewest DAC steps. As you say, our pitch perception is pretty poor down there, so as long as you've got plenty of resolution by the time the oscillator has got up to the sensitive midrange frequencies, you can probably afford to lose some further down. This definitely helps, I agree!
>>>
>>> Tom
>>>
>>>
>>>> On 16 Mar 2026, at 16:01, Roman Sowa <modular@go2.pl> wrote:
>>>>
>>>> Your point is valid, but rises fundamental question:
>>>> who needs 8 or 10 octave continuous tuning range anyway?
>>>> No analog synth has more than 5 and today most popular is 3-4.
>>>> That means smaller range is enough and you can add octave switch as separate control to cover all frequencies of interesrt. Just like a knob doing that in any synth basicaly.
>>>> And octave switch in DAC controled Hz/V VCO is just switching the resistance on the way from DAMC to VCO core. This is also exactly how it was done in vintage linear VCO synths like MS20 and many more.
>>>> And tuning it further down with modulation or pitchbend doesn't require much of a precicion, especially when it happens at so low frequencies where ear pitch perception is rather limited.
>>>>
>>>> Roman
>>>>
>>>> W dniu 2026-03-16 o 16:13, Tom Wiltshire pisze:
>>>>> That's a thought-provoking point of view, Roman.
>>>>> I thought the fundamental problem with linear VCOs was that it's hard to cover an 8 or even ten octave range in a linear fashion - that's 256 or 1024 times the base value. For a CV, that requires very small offsets at the bottom. Say we're going with 1V/KHz. We have a maximum CV somewhere up around 10 or 12V giving us an impressive 10 or 12KHz maximum frequency. But we have an system of 1mV=1Hz, so errors will need to be in the sub-mV level at the low end to keep things sounding in tune.
>>>>> uPs and DACs do help in some respects, and since DACs are always always linear these days, it's kind-of an obvious way to go. But again, the accuracy requirement bites at the bottom end. Since the width of each octave doubles every time, if I have 1024 steps in a low octave, then I have 262144 steps in an octave eight octaves higher! That's a massive waste of DAC resolution, and eats up DAC accuracy in a way that makes finding a suitable device a lot harder and more expensive.
>>>>> Still, you've given me something to think about, and I shall go and do some calculations to see how well this might work. Thanks.
>>>>> Tom
>>>>>> On 16 Mar 2026, at 11:36, Roman Sowa via Synth-diy <synth-diy@synth-diy.org> wrote:
>>>>>>
>>>>>> IMHO linear response VCOs are simply better and easier to use. No temperature dependency by design and easier to tune - as long as all offsets are small or trimmed, whatever you do, an octave is always an octave, not like with exp VCO where you have to turn 3 to 5 trimmers in several iterations to get them sound remotely acceptable.
>>>>>>
>>>>>> One may say that summing few CV inputs is difficult for linear VCOs, but today the VCOs are controlled by microcontroller anyway, so the problem of the 1970's is no longer any concern.
>>>>>>
>>>>>> Roman
>>>>>>
>>>>>> W dniu 2026-03-16 o 03:49, Thomas Hudson via Synth-diy pisze:
>>>>>>> My first modular synth was a Paia 4700/j synth. Yes I know, cheesy design, but the one thing I miss is how linear response VCOs have a very different detuned chorusing effect with each interval. Just wondering if anyone has experimented with this?
>>>>>>> ________________________________________________________
>>>>>>> This is the Synth-diy mailing list
>>>>>>> Submit email to: Synth-diy@synth-diy.org
>>>>>>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>>>>>>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>>>>>>> Selling or trading? Use marketplace@synth-diy.org
>>>>>> ________________________________________________________
>>>>>> This is the Synth-diy mailing list
>>>>>> Submit email to: Synth-diy@synth-diy.org
>>>>>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>>>>>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>>>>>> Selling or trading? Use marketplace@synth-diy.org
>>> ________________________________________________________
>>> This is the Synth-diy mailing list
>>> Submit email to: Synth-diy@synth-diy.org
>>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>>> Selling or trading? Use marketplace@synth-diy.org
>>>
>> ________________________________________________________
>> This is the Synth-diy mailing list
>> Submit email to: Synth-diy@synth-diy.org
>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>> Selling or trading? Use marketplace@synth-diy.org
>
> ________________________________________________________
> This is the Synth-diy mailing list
> Submit email to: Synth-diy@synth-diy.org
> View archive at: https://synth-diy.org/pipermail/synth-diy/
> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
> Selling or trading? Use marketplace@synth-diy.org

> On Mar 17, 2026, at 18:44, David Manley via Synth-diy <synth-diy@synth-diy.org> wrote:
> 
> It's interesting to see how PAiA's John Simonton solved some these issues in the 1970's by having a custom laser trimmed resistor network built for their 6-bit "Equally Tempered DAC" to be used with linear VCOs.  See the bottom of the schematic on page 18, the resistor values are on the last page.
> 
> https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf
> 
> As is typical for PAiA a very low cost solution: build your own DAC with a few components.
> 
> -Dave
> 
> On 3/17/26 4:12 AM, Tom Wiltshire wrote:
>> While you might not need 8 or 10 octaves at once to play a melody, many V/Oct analog synths *do* have a lot more than five as a continuous range. Arguably that's a problem, since it makes the tuning issues worse and it necessitated the High Frequency Trimmer, but that *is* how it's been done many times. And that is probably thanks to the 3340 and its temperature compensated expo convertor, as you say!
>> The Sequential Pro-One is one example. It has a three octave keyboard and an octave switch, and all the octave switch does is feed nicely calibrated doses of 1V to the oscillators to switch the octave.
>> 
>> We certainly both agree that trying to cover less range in one go makes life technically simpler for an oscillator designer, almost whatever oscillator design is used. That would be true for a V/Oct oscillator too. If we only have to cover 4 octaves in tune, then our expo convertor has a lot less to do and scale errors are much less of an issue. That's not specific to linear oscillators. Another example where this sort of range switching makes life easier and hides the limitations is the Juno 106's divider-based DCOs. Without range switching, there wouldn't be enough divider resolution to get really low notes, and the stepping at the high end would be even worse than it is already!
>> 
>> For a linear oscillator run from a DAC, the errors would be worse for the lowest frequencies where you have the fewest DAC steps. As you say, our pitch perception is pretty poor down there, so as long as you've got plenty of resolution by the time the oscillator has got up to the sensitive midrange frequencies, you can probably afford to lose some further down. This definitely helps, I agree!
>> 
>> Tom
>> 
>> 
>>> On 16 Mar 2026, at 16:01, Roman Sowa <modular@go2.pl> wrote:
>>> 
>>> Your point is valid, but rises fundamental question:
>>> who needs 8 or 10 octave continuous tuning range anyway?
>>> No analog synth has more than 5 and today most popular is 3-4.
>>> That means smaller range is enough and you can add octave switch as separate control to cover all frequencies of interesrt. Just like a knob doing that in any synth basicaly.
>>> And octave switch in DAC controled Hz/V VCO is just switching the resistance on the way from DAMC to VCO core. This is also exactly how it was done in vintage linear VCO synths like MS20 and many more.
>>> And tuning it further down with modulation or pitchbend doesn't require much of a precicion, especially when it happens at so low frequencies where ear pitch perception is rather limited.
>>> 
>>> Roman
>>> 
>>> W dniu 2026-03-16 o 16:13, Tom Wiltshire pisze:
>>>> That's a thought-provoking point of view, Roman.
>>>> I thought the fundamental problem with linear VCOs was that it's hard to cover an 8 or even ten octave range in a linear fashion - that's 256 or 1024 times the base value. For a CV, that requires very small offsets at the bottom. Say we're going with 1V/KHz. We have a maximum CV somewhere up around 10 or 12V giving us an impressive 10 or 12KHz maximum frequency. But we have an system of 1mV=1Hz, so errors will need to be in the sub-mV level at the low end to keep things sounding in tune.
>>>> uPs and DACs do help in some respects, and since DACs are always always linear these days, it's kind-of an obvious way to go. But again, the accuracy requirement bites at the bottom end. Since the width of each octave doubles every time, if I have 1024 steps in a low octave, then I have 262144 steps in an octave eight octaves higher! That's a massive waste of DAC resolution, and eats up DAC accuracy in a way that makes finding a suitable device a lot harder and more expensive.
>>>> Still, you've given me something to think about, and I shall go and do some calculations to see how well this might work. Thanks.
>>>> Tom
>>>>> On 16 Mar 2026, at 11:36, Roman Sowa via Synth-diy <synth-diy@synth-diy.org> wrote:
>>>>> 
>>>>> IMHO linear response VCOs are simply better and easier to use. No temperature dependency by design and easier to tune - as long as all offsets are small or trimmed, whatever you do, an octave is always an octave, not like with exp VCO where you have to turn 3 to 5 trimmers in several iterations to get them sound remotely acceptable.
>>>>> 
>>>>> One may say that summing few CV inputs is difficult for linear VCOs, but today the VCOs are controlled by microcontroller anyway, so the problem of the 1970's is no longer any concern.
>>>>> 
>>>>> Roman
>>>>> 
>>>>> W dniu 2026-03-16 o 03:49, Thomas Hudson via Synth-diy pisze:
>>>>>> My first modular synth was a Paia 4700/j synth. Yes I know, cheesy design, but the one thing I miss is how linear response VCOs have a very different detuned chorusing effect with each interval. Just wondering if anyone has experimented with this?
>>>>>> ________________________________________________________
>>>>>> This is the Synth-diy mailing list
>>>>>> Submit email to: Synth-diy@synth-diy.org
>>>>>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>>>>>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>>>>>> Selling or trading? Use marketplace@synth-diy.org
>>>>> ________________________________________________________
>>>>> This is the Synth-diy mailing list
>>>>> Submit email to: Synth-diy@synth-diy.org
>>>>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>>>>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>>>>> Selling or trading? Use marketplace@synth-diy.org
>> 
>> ________________________________________________________
>> This is the Synth-diy mailing list
>> Submit email to: Synth-diy@synth-diy.org
>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>> Selling or trading? Use marketplace@synth-diy.org
>> 
> ________________________________________________________
> This is the Synth-diy mailing list
> Submit email to: Synth-diy@synth-diy.org
> View archive at: https://synth-diy.org/pipermail/synth-diy/
> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
> Selling or trading? Use marketplace@synth-diy.org

> On 18 Mar 2026, at 11:31, mark verbos via Synth-diy <synth-diy@synth-diy.org> wrote:
> 
> Like a TR-909.
> But, surely it is cheaper to use 1% resistors rather than a custom resistor array made.
> 
> 
> Mark
> 
> 
> 
>> On Mar 17, 2026, at 18:44, David Manley via Synth-diy <synth-diy@synth-diy.org> wrote:
>> 
>> It's interesting to see how PAiA's John Simonton solved some these issues in the 1970's by having a custom laser trimmed resistor network built for their 6-bit "Equally Tempered DAC" to be used with linear VCOs.  See the bottom of the schematic on page 18, the resistor values are on the last page.
>> 
>> https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf
>> 
>> As is typical for PAiA a very low cost solution: build your own DAC with a few components.
>> 
>> -Dave
>> 
>> On 3/17/26 4:12 AM, Tom Wiltshire wrote:
>>> While you might not need 8 or 10 octaves at once to play a melody, many V/Oct analog synths *do* have a lot more than five as a continuous range. Arguably that's a problem, since it makes the tuning issues worse and it necessitated the High Frequency Trimmer, but that *is* how it's been done many times. And that is probably thanks to the 3340 and its temperature compensated expo convertor, as you say!
>>> The Sequential Pro-One is one example. It has a three octave keyboard and an octave switch, and all the octave switch does is feed nicely calibrated doses of 1V to the oscillators to switch the octave.
>>> 
>>> We certainly both agree that trying to cover less range in one go makes life technically simpler for an oscillator designer, almost whatever oscillator design is used. That would be true for a V/Oct oscillator too. If we only have to cover 4 octaves in tune, then our expo convertor has a lot less to do and scale errors are much less of an issue. That's not specific to linear oscillators. Another example where this sort of range switching makes life easier and hides the limitations is the Juno 106's divider-based DCOs. Without range switching, there wouldn't be enough divider resolution to get really low notes, and the stepping at the high end would be even worse than it is already!
>>> 
>>> For a linear oscillator run from a DAC, the errors would be worse for the lowest frequencies where you have the fewest DAC steps. As you say, our pitch perception is pretty poor down there, so as long as you've got plenty of resolution by the time the oscillator has got up to the sensitive midrange frequencies, you can probably afford to lose some further down. This definitely helps, I agree!
>>> 
>>> Tom
>>> 
>>> 
>>>> On 16 Mar 2026, at 16:01, Roman Sowa <modular@go2.pl> wrote:
>>>> 
>>>> Your point is valid, but rises fundamental question:
>>>> who needs 8 or 10 octave continuous tuning range anyway?
>>>> No analog synth has more than 5 and today most popular is 3-4.
>>>> That means smaller range is enough and you can add octave switch as separate control to cover all frequencies of interesrt. Just like a knob doing that in any synth basicaly.
>>>> And octave switch in DAC controled Hz/V VCO is just switching the resistance on the way from DAMC to VCO core. This is also exactly how it was done in vintage linear VCO synths like MS20 and many more.
>>>> And tuning it further down with modulation or pitchbend doesn't require much of a precicion, especially when it happens at so low frequencies where ear pitch perception is rather limited.
>>>> 
>>>> Roman
>>>> 
>>>> W dniu 2026-03-16 o 16:13, Tom Wiltshire pisze:
>>>>> That's a thought-provoking point of view, Roman.
>>>>> I thought the fundamental problem with linear VCOs was that it's hard to cover an 8 or even ten octave range in a linear fashion - that's 256 or 1024 times the base value. For a CV, that requires very small offsets at the bottom. Say we're going with 1V/KHz. We have a maximum CV somewhere up around 10 or 12V giving us an impressive 10 or 12KHz maximum frequency. But we have an system of 1mV=1Hz, so errors will need to be in the sub-mV level at the low end to keep things sounding in tune.
>>>>> uPs and DACs do help in some respects, and since DACs are always always linear these days, it's kind-of an obvious way to go. But again, the accuracy requirement bites at the bottom end. Since the width of each octave doubles every time, if I have 1024 steps in a low octave, then I have 262144 steps in an octave eight octaves higher! That's a massive waste of DAC resolution, and eats up DAC accuracy in a way that makes finding a suitable device a lot harder and more expensive.
>>>>> Still, you've given me something to think about, and I shall go and do some calculations to see how well this might work. Thanks.
>>>>> Tom
>>>>>> On 16 Mar 2026, at 11:36, Roman Sowa via Synth-diy <synth-diy@synth-diy.org> wrote:
>>>>>> 
>>>>>> IMHO linear response VCOs are simply better and easier to use. No temperature dependency by design and easier to tune - as long as all offsets are small or trimmed, whatever you do, an octave is always an octave, not like with exp VCO where you have to turn 3 to 5 trimmers in several iterations to get them sound remotely acceptable.
>>>>>> 
>>>>>> One may say that summing few CV inputs is difficult for linear VCOs, but today the VCOs are controlled by microcontroller anyway, so the problem of the 1970's is no longer any concern.
>>>>>> 
>>>>>> Roman
>>>>>> 
>>>>>> W dniu 2026-03-16 o 03:49, Thomas Hudson via Synth-diy pisze:
>>>>>>> My first modular synth was a Paia 4700/j synth. Yes I know, cheesy design, but the one thing I miss is how linear response VCOs have a very different detuned chorusing effect with each interval. Just wondering if anyone has experimented with this?
>>>>>>> ________________________________________________________
>>>>>>> This is the Synth-diy mailing list
>>>>>>> Submit email to: Synth-diy@synth-diy.org
>>>>>>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>>>>>>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>>>>>>> Selling or trading? Use marketplace@synth-diy.org
>>>>>> ________________________________________________________
>>>>>> This is the Synth-diy mailing list
>>>>>> Submit email to: Synth-diy@synth-diy.org
>>>>>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>>>>>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>>>>>> Selling or trading? Use marketplace@synth-diy.org
>>> 
>>> ________________________________________________________
>>> This is the Synth-diy mailing list
>>> Submit email to: Synth-diy@synth-diy.org
>>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>>> Selling or trading? Use marketplace@synth-diy.org
>>> 
>> ________________________________________________________
>> This is the Synth-diy mailing list
>> Submit email to: Synth-diy@synth-diy.org
>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>> Selling or trading? Use marketplace@synth-diy.org
> 
> 
> ________________________________________________________
> This is the Synth-diy mailing list
> Submit email to: Synth-diy@synth-diy.org
> View archive at: https://synth-diy.org/pipermail/synth-diy/
> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
> Selling or trading? Use marketplace@synth-diy.org

> On Mar 18, 2026, at 2:54 AM, Roman Sowa via Synth-diy <synth-diy@synth-diy.org> wrote:
> 
> Today I see no point in building PAiA DAC or any other similar solutions, except for nostalgic reasons.
> Dirt cheap 16-bit commercially available DACs offer much better linearity than any trimmed or selected resistor discrete DAC could ever do.
> 16-bit DAC in MIDImplant offered no more than +/-2 cents errors max in Hz/V scale at lowest notes and 5 octave range, often less.
> 
> Roman
> 
> W dniu 2026-03-18 o 01:30, Thomas Hudson via Synth-diy pisze:
>> Unfortunately this solution didn't work. I built this synth and he had to add three to four trimmers, all of them interacting to tune the thing. It was doable and I was able to get tracking pretty good. Perhaps things have gotten better, but from experience it took some tweaking.
>> 
>>> On Mar 17, 2026, at 1:44 PM, David Manley via Synth-diy <synth-diy@synth-diy.org> wrote:
>>> 
>>> It's interesting to see how PAiA's John Simonton solved some these issues in the 1970's by having a custom laser trimmed resistor network built for their 6-bit "Equally Tempered DAC" to be used with linear VCOs.  See the bottom of the schematic on page 18, the resistor values are on the last page.
>>> 
>>> https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf
>>> 
>>> As is typical for PAiA a very low cost solution: build your own DAC with a few components.
>>> 
>>> -Dave
>>> 
>>> On 3/17/26 4:12 AM, Tom Wiltshire wrote:
>>>> While you might not need 8 or 10 octaves at once to play a melody, many V/Oct analog synths *do* have a lot more than five as a continuous range. Arguably that's a problem, since it makes the tuning issues worse and it necessitated the High Frequency Trimmer, but that *is* how it's been done many times. And that is probably thanks to the 3340 and its temperature compensated expo convertor, as you say!
>>>> The Sequential Pro-One is one example. It has a three octave keyboard and an octave switch, and all the octave switch does is feed nicely calibrated doses of 1V to the oscillators to switch the octave.
>>>> 
>>>> We certainly both agree that trying to cover less range in one go makes life technically simpler for an oscillator designer, almost whatever oscillator design is used. That would be true for a V/Oct oscillator too. If we only have to cover 4 octaves in tune, then our expo convertor has a lot less to do and scale errors are much less of an issue. That's not specific to linear oscillators. Another example where this sort of range switching makes life easier and hides the limitations is the Juno 106's divider-based DCOs. Without range switching, there wouldn't be enough divider resolution to get really low notes, and the stepping at the high end would be even worse than it is already!
>>>> 
>>>> For a linear oscillator run from a DAC, the errors would be worse for the lowest frequencies where you have the fewest DAC steps. As you say, our pitch perception is pretty poor down there, so as long as you've got plenty of resolution by the time the oscillator has got up to the sensitive midrange frequencies, you can probably afford to lose some further down. This definitely helps, I agree!
>>>> 
>>>> Tom
>>>> 
>>>> 
>>>>> On 16 Mar 2026, at 16:01, Roman Sowa <modular@go2.pl> wrote:
>>>>> 
>>>>> Your point is valid, but rises fundamental question:
>>>>> who needs 8 or 10 octave continuous tuning range anyway?
>>>>> No analog synth has more than 5 and today most popular is 3-4.
>>>>> That means smaller range is enough and you can add octave switch as separate control to cover all frequencies of interesrt. Just like a knob doing that in any synth basicaly.
>>>>> And octave switch in DAC controled Hz/V VCO is just switching the resistance on the way from DAMC to VCO core. This is also exactly how it was done in vintage linear VCO synths like MS20 and many more.
>>>>> And tuning it further down with modulation or pitchbend doesn't require much of a precicion, especially when it happens at so low frequencies where ear pitch perception is rather limited.
>>>>> 
>>>>> Roman
>>>>> 
>>>>> W dniu 2026-03-16 o 16:13, Tom Wiltshire pisze:
>>>>>> That's a thought-provoking point of view, Roman.
>>>>>> I thought the fundamental problem with linear VCOs was that it's hard to cover an 8 or even ten octave range in a linear fashion - that's 256 or 1024 times the base value. For a CV, that requires very small offsets at the bottom. Say we're going with 1V/KHz. We have a maximum CV somewhere up around 10 or 12V giving us an impressive 10 or 12KHz maximum frequency. But we have an system of 1mV=1Hz, so errors will need to be in the sub-mV level at the low end to keep things sounding in tune.
>>>>>> uPs and DACs do help in some respects, and since DACs are always always linear these days, it's kind-of an obvious way to go. But again, the accuracy requirement bites at the bottom end. Since the width of each octave doubles every time, if I have 1024 steps in a low octave, then I have 262144 steps in an octave eight octaves higher! That's a massive waste of DAC resolution, and eats up DAC accuracy in a way that makes finding a suitable device a lot harder and more expensive.
>>>>>> Still, you've given me something to think about, and I shall go and do some calculations to see how well this might work. Thanks.
>>>>>> Tom
>>>>>>> On 16 Mar 2026, at 11:36, Roman Sowa via Synth-diy <synth-diy@synth-diy.org> wrote:
>>>>>>> 
>>>>>>> IMHO linear response VCOs are simply better and easier to use. No temperature dependency by design and easier to tune - as long as all offsets are small or trimmed, whatever you do, an octave is always an octave, not like with exp VCO where you have to turn 3 to 5 trimmers in several iterations to get them sound remotely acceptable.
>>>>>>> 
>>>>>>> One may say that summing few CV inputs is difficult for linear VCOs, but today the VCOs are controlled by microcontroller anyway, so the problem of the 1970's is no longer any concern.
>>>>>>> 
>>>>>>> Roman
>>>>>>> 
>>>>>>> W dniu 2026-03-16 o 03:49, Thomas Hudson via Synth-diy pisze:
>>>>>>>> My first modular synth was a Paia 4700/j synth. Yes I know, cheesy design, but the one thing I miss is how linear response VCOs have a very different detuned chorusing effect with each interval. Just wondering if anyone has experimented with this?
>>>>>>>> ________________________________________________________
>>>>>>>> This is the Synth-diy mailing list
>>>>>>>> Submit email to: Synth-diy@synth-diy.org
>>>>>>>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>>>>>>>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>>>>>>>> Selling or trading? Use marketplace@synth-diy.org
>>>>>>> ________________________________________________________
>>>>>>> This is the Synth-diy mailing list
>>>>>>> Submit email to: Synth-diy@synth-diy.org
>>>>>>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>>>>>>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>>>>>>> Selling or trading? Use marketplace@synth-diy.org
>>>> ________________________________________________________
>>>> This is the Synth-diy mailing list
>>>> Submit email to: Synth-diy@synth-diy.org
>>>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>>>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>>>> Selling or trading? Use marketplace@synth-diy.org
>>>> 
>>> ________________________________________________________
>>> This is the Synth-diy mailing list
>>> Submit email to: Synth-diy@synth-diy.org
>>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>>> Selling or trading? Use marketplace@synth-diy.org
>> 
>> ________________________________________________________
>> This is the Synth-diy mailing list
>> Submit email to: Synth-diy@synth-diy.org
>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>> Selling or trading? Use marketplace@synth-diy.org
> 
> ________________________________________________________
> This is the Synth-diy mailing list
> Submit email to: Synth-diy@synth-diy.org
> View archive at: https://synth-diy.org/pipermail/synth-diy/
> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
> Selling or trading? Use marketplace@synth-diy.org

On Mar 18, 2026, at 4:34 AM, Tom Wiltshire wrote:
> Roland had form for this. SH-101 uses a simple DAC built from a few resistors too.
> 
> Like Roman said, it doesn't really make sense nowadays when DACs are cheap, but it was worth it then.
> 
> Tom
> 
> On 18 Mar 2026, at 11:31, mark verbos wrote:
>> Like a TR-909.
>> But, surely it is cheaper to use 1% resistors rather than a custom resistor array made.
>> 
>> Mark
>> 
>> On Mar 17, 2026, at 18:44, David Manley wrote:
>>> It's interesting to see how PAiA's John Simonton solved some these issues in the 1970's by having a custom laser trimmed resistor network built for their 6-bit "Equally Tempered DAC" to be used with linear VCOs.  See the bottom of the schematic on page 18, the resistor values are on the last page.
>>> 
>>> https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf
>>> 
>>> As is typical for PAiA a very low cost solution: build your own DAC with a few components.
>>> 
>>> -Dave

From: Synth-diy <synth-diy-bounces@synth-diy.org> on behalf of brianw <brianw@audiobanshee.com>
Sent: 18 March 2026 19:02
To: synth-diy@synth-diy.org <synth-diy@synth-diy.org>
Subject: Re: [sdiy] Linear response VCOs?

The Prophet 5, Rev 1 and Rev 2, use a 7-bit DAC made from hand-picked resistors. There is a note in the Service Manual that you should *not* replace these resistors because of the challenge of matching a new one to the network. 1 LSB is calibrated to 1/12 V (0.0833 V) for easy use in 1V/8va scaling. CV ranges from 0 V to 10.583 V (127/12), but the Prophet 5 only uses the lower 6 bits for pitch, limiting the range to 5.333 V maximum and thus 5 octaves. All CV were 7-bit, but the pitch combined coarse and fine with the scale of the DAC changed so that there were 64 steps in the coarse range plus another 128 steps in the fine range. This wasn't quite as accurate as a 13-bit DAC, but still quite accurate for the time.

The Prophet 5 Rev 3 simply used a 16-bit DAC, but maintained the firmware design with 7 bits per CV, so the pitch did not enjoy a full 16-bit precision. The 13-bit pitch values still have 16-bit accuracy, though, just not 65536 steps of precision.

One thing to note, Mark, is that a 6-bit DAC has an LSB that's 1.56% of the total range, so 1% resistors would be quite awful. Then there's the fact that a 1% error in the MSB could throw the whole binary scale off enough that the values are not monotonic (i.e. an increase in the code could actually cause a decrease in voltage!). A 7-bit DAC has the LSB at 0.78% so you definitely need better than 1% precision. These manufacturers were not making a custom resistor array so much as hand-selecting individual resistors that were matched well across the whole group.

Today, not only are 1% resistors more readily available than they were in the seventies, but you can even get 0.1% tolerance resistors at a reasonable. Still, that doesn't even get you to a full 9-bit DAC. This illustrates how impressive DAC chip technology is. One of the fasted DAC chips I've designed with can run at a sample rate of 125 MHz (yeah, MHz, not kHz) based on current switching rather than voltage, but it stops at 14-bit precision because the smallest current is only 0.0061% of the largest, and it's difficult to be precise enough at such a large scale factor. Larger DAC precision requires a different technique than binary-weighted digits. Fortunately, there are many ways to implement a DAC.

Brian


On Mar 18, 2026, at 4:34 AM, Tom Wiltshire wrote:
> Roland had form for this. SH-101 uses a simple DAC built from a few resistors too.
>
> Like Roman said, it doesn't really make sense nowadays when DACs are cheap, but it was worth it then.
>
> Tom
>
> On 18 Mar 2026, at 11:31, mark verbos wrote:
>> Like a TR-909.
>> But, surely it is cheaper to use 1% resistors rather than a custom resistor array made.
>>
>> Mark
>>
>> On Mar 17, 2026, at 18:44, David Manley wrote:
>>> It's interesting to see how PAiA's John Simonton solved some these issues in the 1970's by having a custom laser trimmed resistor network built for their 6-bit "Equally Tempered DAC" to be used with linear VCOs.  See the bottom of the schematic on page 18, the resistor values are on the last page.
>>>
>>> https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf
>>>
>>> As is typical for PAiA a very low cost solution: build your own DAC with a few components.
>>>
>>> -Dave


________________________________________________________
This is the Synth-diy mailing list
Submit email to: Synth-diy@synth-diy.org
View archive at: https://synth-diy.org/pipermail/synth-diy/
Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
Selling or trading? Use marketplace@synth-diy.org

> Does anybody know why these old synths didn't use PWM/PDM techniques ?
> 
> LS-TTL or CMOS feeding a comparator into an analogue integrator gave 12 
> bits performance at audio frequencies even in the 70s so CVs good enough 
> for tuning would have been easy.
> ------------------------------------------------------------------------
> *From:* Synth-diy <synth-diy-bounces@synth-diy.org> on behalf of brianw 
> <brianw@audiobanshee.com>
> *Sent:* 18 March 2026 19:02
> *To:* synth-diy@synth-diy.org <synth-diy@synth-diy.org>
> *Subject:* Re: [sdiy] Linear response VCOs?
> The Prophet 5, Rev 1 and Rev 2, use a 7-bit DAC made from hand-picked 
> resistors. There is a note in the Service Manual that you should *not* 
> replace these resistors because of the challenge of matching a new one 
> to the network. 1 LSB is calibrated to 1/12 V (0.0833 V) for easy use in 
> 1V/8va scaling. CV ranges from 0 V to 10.583 V (127/12), but the Prophet 
> 5 only uses the lower 6 bits for pitch, limiting the range to 5.333 V 
> maximum and thus 5 octaves. All CV were 7-bit, but the pitch combined 
> coarse and fine with the scale of the DAC changed so that there were 64 
> steps in the coarse range plus another 128 steps in the fine range. This 
> wasn't quite as accurate as a 13-bit DAC, but still quite accurate for 
> the time.
> 
> The Prophet 5 Rev 3 simply used a 16-bit DAC, but maintained the 
> firmware design with 7 bits per CV, so the pitch did not enjoy a full 
> 16-bit precision. The 13-bit pitch values still have 16-bit accuracy, 
> though, just not 65536 steps of precision.
> 
> One thing to note, Mark, is that a 6-bit DAC has an LSB that's 1.56% of 
> the total range, so 1% resistors would be quite awful. Then there's the 
> fact that a 1% error in the MSB could throw the whole binary scale off 
> enough that the values are not monotonic (i.e. an increase in the code 
> could actually cause a decrease in voltage!). A 7-bit DAC has the LSB at 
> 0.78% so you definitely need better than 1% precision. These 
> manufacturers were not making a custom resistor array so much as 
> hand-selecting individual resistors that were matched well across the 
> whole group.
> 
> Today, not only are 1% resistors more readily available than they were 
> in the seventies, but you can even get 0.1% tolerance resistors at a 
> reasonable. Still, that doesn't even get you to a full 9-bit DAC. This 
> illustrates how impressive DAC chip technology is. One of the fasted DAC 
> chips I've designed with can run at a sample rate of 125 MHz (yeah, MHz, 
> not kHz) based on current switching rather than voltage, but it stops at 
> 14-bit precision because the smallest current is only 0.0061% of the 
> largest, and it's difficult to be precise enough at such a large scale 
> factor. Larger DAC precision requires a different technique than 
> binary-weighted digits. Fortunately, there are many ways to implement a DAC.
> 
> Brian
> 
> 
> On Mar 18, 2026, at 4:34 AM, Tom Wiltshire wrote:
>> Roland had form for this. SH-101 uses a simple DAC built from a few resistors too.
>> 
>> Like Roman said, it doesn't really make sense nowadays when DACs are cheap, but it was worth it then.
>> 
>> Tom
>> 
>> On 18 Mar 2026, at 11:31, mark verbos wrote:
>>> Like a TR-909.
>>> But, surely it is cheaper to use 1% resistors rather than a custom resistor array made.
>>> 
>>> Mark
>>> 
>>> On Mar 17, 2026, at 18:44, David Manley wrote:
>>>> It's interesting to see how PAiA's John Simonton solved some these issues in the 1970's by having a custom laser trimmed resistor network built for their 6-bit "Equally Tempered DAC" to be used with linear VCOs.  See the bottom of the schematic on page  18, the resistor values are on the last page.
>>>> 
>>>> https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf 
> <https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf>
>>>> 
>>>> As is typical for PAiA a very low cost solution: build your own DAC with a few components.
>>>> 
>>>> -Dave
> 
> 
> ________________________________________________________
> This is the Synth-diy mailing list
> Submit email to: Synth-diy@synth-diy.org
> View archive at: https://synth-diy.org/pipermail/synth-diy/ 
> <https://synth-diy.org/pipermail/synth-diy/>
> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy 
> <https://synth-diy.org/mailman/listinfo/synth-diy>
> Selling or trading? Use marketplace@synth-diy.org
> 
> ________________________________________________________
> This is the Synth-diy mailing list
> Submit email to: Synth-diy@synth-diy.org
> View archive at: https://synth-diy.org/pipermail/synth-diy/
> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
> Selling or trading? Use marketplace@synth-diy.org

> Does anybody know why these old synths didn't use PWM/PDM techniques ?
>
> LS-TTL or CMOS feeding a comparator into an analogue integrator gave 12
> bits performance at audio frequencies even in the 70s so CVs good enough
> for tuning would have been easy.
> ------------------------------------------------------------------------
> *From:* Synth-diy <synth-diy-bounces@synth-diy.org> on behalf of brianw
> <brianw@audiobanshee.com>
> *Sent:* 18 March 2026 19:02
> *To:* synth-diy@synth-diy.org <synth-diy@synth-diy.org>
> *Subject:* Re: [sdiy] Linear response VCOs?
> The Prophet 5, Rev 1 and Rev 2, use a 7-bit DAC made from hand-picked
> resistors. There is a note in the Service Manual that you should *not*
> replace these resistors because of the challenge of matching a new one
> to the network. 1 LSB is calibrated to 1/12 V (0.0833 V) for easy use in
> 1V/8va scaling. CV ranges from 0 V to 10.583 V (127/12), but the Prophet
> 5 only uses the lower 6 bits for pitch, limiting the range to 5.333 V
> maximum and thus 5 octaves. All CV were 7-bit, but the pitch combined
> coarse and fine with the scale of the DAC changed so that there were 64
> steps in the coarse range plus another 128 steps in the fine range. This
> wasn't quite as accurate as a 13-bit DAC, but still quite accurate for
> the time.
>
> The Prophet 5 Rev 3 simply used a 16-bit DAC, but maintained the
> firmware design with 7 bits per CV, so the pitch did not enjoy a full
> 16-bit precision. The 13-bit pitch values still have 16-bit accuracy,
> though, just not 65536 steps of precision.
>
> One thing to note, Mark, is that a 6-bit DAC has an LSB that's 1.56% of
> the total range, so 1% resistors would be quite awful. Then there's the
> fact that a 1% error in the MSB could throw the whole binary scale off
> enough that the values are not monotonic (i.e. an increase in the code
> could actually cause a decrease in voltage!). A 7-bit DAC has the LSB at
> 0.78% so you definitely need better than 1% precision. These
> manufacturers were not making a custom resistor array so much as
> hand-selecting individual resistors that were matched well across the
> whole group.
>
> Today, not only are 1% resistors more readily available than they were
> in the seventies, but you can even get 0.1% tolerance resistors at a
> reasonable. Still, that doesn't even get you to a full 9-bit DAC. This
> illustrates how impressive DAC chip technology is. One of the fasted DAC
> chips I've designed with can run at a sample rate of 125 MHz (yeah, MHz,
> not kHz) based on current switching rather than voltage, but it stops at
> 14-bit precision because the smallest current is only 0.0061% of the
> largest, and it's difficult to be precise enough at such a large scale
> factor. Larger DAC precision requires a different technique than
> binary-weighted digits. Fortunately, there are many ways to implement a DAC.
>
> Brian
>
>
> On Mar 18, 2026, at 4:34 AM, Tom Wiltshire wrote:
>> Roland had form for this. SH-101 uses a simple DAC built from a few resistors too.
>>
>> Like Roman said, it doesn't really make sense nowadays when DACs are cheap, but it was worth it then.
>>
>> Tom
>>
>> On 18 Mar 2026, at 11:31, mark verbos wrote:
>>> Like a TR-909.
>>> But, surely it is cheaper to use 1% resistors rather than a custom resistor array made.
>>>
>>> Mark
>>>
>>> On Mar 17, 2026, at 18:44, David Manley wrote:
>>>> It's interesting to see how PAiA's John Simonton solved some these issues in the 1970's by having a custom laser trimmed resistor network built for their 6-bit "Equally Tempered DAC" to be used with linear VCOs.  See the bottom of the schematic on page  18, the resistor values are on the last page.
>>>>
>>>> https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf
> <https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf>
>>>>
>>>> As is typical for PAiA a very low cost solution: build your own DAC with a few components.
>>>>
>>>> -Dave
>
>
> ________________________________________________________
> This is the Synth-diy mailing list
> Submit email to: Synth-diy@synth-diy.org
> View archive at: https://synth-diy.org/pipermail/synth-diy/
> <https://synth-diy.org/pipermail/synth-diy/>
> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
> <https://synth-diy.org/mailman/listinfo/synth-diy>
> Selling or trading? Use marketplace@synth-diy.org
>
> ________________________________________________________
> This is the Synth-diy mailing list
> Submit email to: Synth-diy@synth-diy.org
> View archive at: https://synth-diy.org/pipermail/synth-diy/
> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
> Selling or trading? Use marketplace@synth-diy.org

> On 19 Mar 2026, at 19:30, Mike Bryant <mbryant@futurehorizons.com> wrote:
> 
> Ah okay, the poly requirement makes sense.   But we ran PDM DACs at 8MHz using LS-TTL to give better than 16 bit audio for a single channel so you could still feed that to a MUX and S/H.
> From: Roman Sowa <modular@go2.pl>
> Sent: 19 March 2026 11:10
> To: Mike Bryant <mbryant@futurehorizons.com>; synth-diy@synth-diy.org <synth-diy@synth-diy.org>
> Subject: Re: [sdiy] Linear response VCOs?
>  
> Correct me if I'm wrong but old synths using DAC for CV were all 
> polyhonic. That means a lot of CV sources needed. So they used S&H and 
> muxed DAC. To have PWM with fast enough response to feed MUX and S$H, it 
> would have to run at enormouse frequency, not suitable to affordable 
> technology back then. And putting separate counter as PWM generator for 
> every CV is much more expensive, and takes more space than DAC-MUX-S&H.
> Back then if you wanted a timer, you got 8253 offering 3 timers in one 
> package, and I'm not even sure if it had PWM mode at all.
> 
> Roman
> 
> W dniu 2026-03-18 o 21:39, Mike Bryant pisze:
> > Does anybody know why these old synths didn't use PWM/PDM techniques ?
> > 
> > LS-TTL or CMOS feeding a comparator into an analogue integrator gave 12 
> > bits performance at audio frequencies even in the 70s so CVs good enough 
> > for tuning would have been easy.
> > ------------------------------------------------------------------------
> > *From:* Synth-diy <synth-diy-bounces@synth-diy.org> on behalf of brianw 
> > <brianw@audiobanshee.com>
> > *Sent:* 18 March 2026 19:02
> > *To:* synth-diy@synth-diy.org <synth-diy@synth-diy.org>
> > *Subject:* Re: [sdiy] Linear response VCOs?
> > The Prophet 5, Rev 1 and Rev 2, use a 7-bit DAC made from hand-picked 
> > resistors. There is a note in the Service Manual that you should *not* 
> > replace these resistors because of the challenge of matching a new one 
> > to the network. 1 LSB is calibrated to 1/12 V (0.0833 V) for easy use in 
> > 1V/8va scaling. CV ranges from 0 V to 10.583 V (127/12), but the Prophet 
> > 5 only uses the lower 6 bits for pitch, limiting the range to 5.333 V 
> > maximum and thus 5 octaves. All CV were 7-bit, but the pitch combined 
> > coarse and fine with the scale of the DAC changed so that there were 64 
> > steps in the coarse range plus another 128 steps in the fine range. This 
> > wasn't quite as accurate as a 13-bit DAC, but still quite accurate for 
> > the time.
> > 
> > The Prophet 5 Rev 3 simply used a 16-bit DAC, but maintained the 
> > firmware design with 7 bits per CV, so the pitch did not enjoy a full 
> > 16-bit precision. The 13-bit pitch values still have 16-bit accuracy, 
> > though, just not 65536 steps of precision.
> > 
> > One thing to note, Mark, is that a 6-bit DAC has an LSB that's 1.56% of 
> > the total range, so 1% resistors would be quite awful. Then there's the 
> > fact that a 1% error in the MSB could throw the whole binary scale off 
> > enough that the values are not monotonic (i.e. an increase in the code 
> > could actually cause a decrease in voltage!). A 7-bit DAC has the LSB at 
> > 0.78% so you definitely need better than 1% precision. These 
> > manufacturers were not making a custom resistor array so much as 
> > hand-selecting individual resistors that were matched well across the 
> > whole group.
> > 
> > Today, not only are 1% resistors more readily available than they were 
> > in the seventies, but you can even get 0.1% tolerance resistors at a 
> > reasonable. Still, that doesn't even get you to a full 9-bit DAC. This 
> > illustrates how impressive DAC chip technology is. One of the fasted DAC 
> > chips I've designed with can run at a sample rate of 125 MHz (yeah, MHz, 
> > not kHz) based on current switching rather than voltage, but it stops at 
> > 14-bit precision because the smallest current is only 0.0061% of the 
> > largest, and it's difficult to be precise enough at such a large scale 
> > factor. Larger DAC precision requires a different technique than 
> > binary-weighted digits. Fortunately, there are many ways to implement a DAC.
> > 
> > Brian
> > 
> > 
> > On Mar 18, 2026, at 4:34 AM, Tom Wiltshire wrote:
> >> Roland had form for this. SH-101 uses a simple DAC built from a few resistors too.
> >> 
> >> Like Roman said, it doesn't really make sense nowadays when DACs are cheap, but it was worth it then.
> >> 
> >> Tom
> >> 
> >> On 18 Mar 2026, at 11:31, mark verbos wrote:
> >>> Like a TR-909.
> >>> But, surely it is cheaper to use 1% resistors rather than a custom resistor array made.
> >>> 
> >>> Mark
> >>> 
> >>> On Mar 17, 2026, at 18:44, David Manley wrote:
> >>>> It's interesting to see how PAiA's John Simonton solved some these issues in the 1970's by having a custom laser trimmed resistor network built for their 6-bit "Equally Tempered DAC" to be used with linear VCOs.  See the bottom of the schematic on page  18, the resistor values are on the last page.
> >>>> 
> >>>> https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf <https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf> 
> > <https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf <https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf>>
> >>>> 
> >>>> As is typical for PAiA a very low cost solution: build your own DAC with a few components.
> >>>> 
> >>>> -Dave
> > 
> > 
> > ________________________________________________________
> > This is the Synth-diy mailing list
> > Submit email to: Synth-diy@synth-diy.org
> > View archive at: https://synth-diy.org/pipermail/synth-diy/ <https://synth-diy.org/pipermail/synth-diy/> 
> > <https://synth-diy.org/pipermail/synth-diy/ <https://synth-diy.org/pipermail/synth-diy/>>
> > Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy <https://synth-diy.org/mailman/listinfo/synth-diy> 
> > <https://synth-diy.org/mailman/listinfo/synth-diy <https://synth-diy.org/mailman/listinfo/synth-diy>>
> > Selling or trading? Use marketplace@synth-diy.org
> > 
> > ________________________________________________________
> > This is the Synth-diy mailing list
> > Submit email to: Synth-diy@synth-diy.org
> > View archive at: https://synth-diy.org/pipermail/synth-diy/ <https://synth-diy.org/pipermail/synth-diy/>
> > Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy <https://synth-diy.org/mailman/listinfo/synth-diy>
> > Selling or trading? Use marketplace@synth-diy.org
> ________________________________________________________
> This is the Synth-diy mailing list
> Submit email to: Synth-diy@synth-diy.org
> View archive at: https://synth-diy.org/pipermail/synth-diy/
> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
> Selling or trading? Use marketplace@synth-diy.org

> On Mar 19, 2026, at 5:11 PM, Tom Wiltshire <tom@electricdruid.net> wrote:
> 
> The whole "Mux and S&H" thing was only because decent quality DACs were damn expensive, so it was better to buy one good one and re-use it a lot of times. If there'd been a sixteen channel DAC for the same price, you can bet they've have used it.
> 
> Incidentally, I once worked on a PWM-DAC based CV transposition module. We used one 3-bit PWM to generate octave voltages, and one 4-bit PWM to generate 12 note voltages. Since each PWM was very low resolution, they were able to be very high frequency and able to respond fast. And since the two channels were entirely separate, trimming was a simple matter of tweaking one to give octaves and the other to give semitones and done. Once added together, the whole thing could produce good quality Note CVs over a wide range without using any expensive DAC. Of course, the secret here is that it was semitones only, so the full MIDI range suddenly boils down to seven bit accuracy.
> 
> 
>> On 19 Mar 2026, at 19:30, Mike Bryant <mbryant@futurehorizons.com <mailto:mbryant@futurehorizons.com>> wrote:
>> 
>> Ah okay, the poly requirement makes sense.   But we ran PDM DACs at 8MHz using LS-TTL to give better than 16 bit audio for a single channel so you could still feed that to a MUX and S/H.
>> From: Roman Sowa <modular@go2.pl <mailto:modular@go2.pl>>
>> Sent: 19 March 2026 11:10
>> To: Mike Bryant <mbryant@futurehorizons.com <mailto:mbryant@futurehorizons.com>>; synth-diy@synth-diy.org <mailto:synth-diy@synth-diy.org> <synth-diy@synth-diy.org <mailto:synth-diy@synth-diy.org>>
>> Subject: Re: [sdiy] Linear response VCOs?
>>  
>> Correct me if I'm wrong but old synths using DAC for CV were all 
>> polyhonic. That means a lot of CV sources needed. So they used S&H and 
>> muxed DAC. To have PWM with fast enough response to feed MUX and S$H, it 
>> would have to run at enormouse frequency, not suitable to affordable 
>> technology back then. And putting separate counter as PWM generator for 
>> every CV is much more expensive, and takes more space than DAC-MUX-S&H.
>> Back then if you wanted a timer, you got 8253 offering 3 timers in one 
>> package, and I'm not even sure if it had PWM mode at all.
>> 
>> Roman
>> 
>> W dniu 2026-03-18 o 21:39, Mike Bryant pisze:
>> > Does anybody know why these old synths didn't use PWM/PDM techniques ?
>> > 
>> > LS-TTL or CMOS feeding a comparator into an analogue integrator gave 12 
>> > bits performance at audio frequencies even in the 70s so CVs good enough 
>> > for tuning would have been easy.
>> > ------------------------------------------------------------------------
>> > *From:* Synth-diy <synth-diy-bounces@synth-diy.org <mailto:synth-diy-bounces@synth-diy.org>> on behalf of brianw 
>> > <brianw@audiobanshee.com <mailto:brianw@audiobanshee.com>>
>> > *Sent:* 18 March 2026 19:02
>> > *To:* synth-diy@synth-diy.org <mailto:synth-diy@synth-diy.org> <synth-diy@synth-diy.org <mailto:synth-diy@synth-diy.org>>
>> > *Subject:* Re: [sdiy] Linear response VCOs?
>> > The Prophet 5, Rev 1 and Rev 2, use a 7-bit DAC made from hand-picked 
>> > resistors. There is a note in the Service Manual that you should *not* 
>> > replace these resistors because of the challenge of matching a new one 
>> > to the network. 1 LSB is calibrated to 1/12 V (0.0833 V) for easy use in 
>> > 1V/8va scaling. CV ranges from 0 V to 10.583 V (127/12), but the Prophet 
>> > 5 only uses the lower 6 bits for pitch, limiting the range to 5.333 V 
>> > maximum and thus 5 octaves. All CV were 7-bit, but the pitch combined 
>> > coarse and fine with the scale of the DAC changed so that there were 64 
>> > steps in the coarse range plus another 128 steps in the fine range. This 
>> > wasn't quite as accurate as a 13-bit DAC, but still quite accurate for 
>> > the time.
>> > 
>> > The Prophet 5 Rev 3 simply used a 16-bit DAC, but maintained the 
>> > firmware design with 7 bits per CV, so the pitch did not enjoy a full 
>> > 16-bit precision. The 13-bit pitch values still have 16-bit accuracy, 
>> > though, just not 65536 steps of precision.
>> > 
>> > One thing to note, Mark, is that a 6-bit DAC has an LSB that's 1.56% of 
>> > the total range, so 1% resistors would be quite awful. Then there's the 
>> > fact that a 1% error in the MSB could throw the whole binary scale off 
>> > enough that the values are not monotonic (i.e. an increase in the code 
>> > could actually cause a decrease in voltage!). A 7-bit DAC has the LSB at 
>> > 0.78% so you definitely need better than 1% precision. These 
>> > manufacturers were not making a custom resistor array so much as 
>> > hand-selecting individual resistors that were matched well across the 
>> > whole group.
>> > 
>> > Today, not only are 1% resistors more readily available than they were 
>> > in the seventies, but you can even get 0.1% tolerance resistors at a 
>> > reasonable. Still, that doesn't even get you to a full 9-bit DAC. This 
>> > illustrates how impressive DAC chip technology is. One of the fasted DAC 
>> > chips I've designed with can run at a sample rate of 125 MHz (yeah, MHz, 
>> > not kHz) based on current switching rather than voltage, but it stops at 
>> > 14-bit precision because the smallest current is only 0.0061% of the 
>> > largest, and it's difficult to be precise enough at such a large scale 
>> > factor. Larger DAC precision requires a different technique than 
>> > binary-weighted digits. Fortunately, there are many ways to implement a DAC.
>> > 
>> > Brian
>> > 
>> > 
>> > On Mar 18, 2026, at 4:34 AM, Tom Wiltshire wrote:
>> >> Roland had form for this. SH-101 uses a simple DAC built from a few resistors too.
>> >> 
>> >> Like Roman said, it doesn't really make sense nowadays when DACs are cheap, but it was worth it then.
>> >> 
>> >> Tom
>> >> 
>> >> On 18 Mar 2026, at 11:31, mark verbos wrote:
>> >>> Like a TR-909.
>> >>> But, surely it is cheaper to use 1% resistors rather than a custom resistor array made.
>> >>> 
>> >>> Mark
>> >>> 
>> >>> On Mar 17, 2026, at 18:44, David Manley wrote:
>> >>>> It's interesting to see how PAiA's John Simonton solved some these issues in the 1970's by having a custom laser trimmed resistor network built for their 6-bit "Equally Tempered DAC" to be used with linear VCOs.  See the bottom of the schematic on page  18, the resistor values are on the last page.
>> >>>> 
>> >>>> https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf 
>> > <https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf>
>> >>>> 
>> >>>> As is typical for PAiA a very low cost solution: build your own DAC with a few components.
>> >>>> 
>> >>>> -Dave
>> > 
>> > 
>> > ________________________________________________________
>> > This is the Synth-diy mailing list
>> > Submit email to: Synth-diy@synth-diy.org <mailto:Synth-diy@synth-diy.org>
>> > View archive at: https://synth-diy.org/pipermail/synth-diy/ 
>> > <https://synth-diy.org/pipermail/synth-diy/>
>> > Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy 
>> > <https://synth-diy.org/mailman/listinfo/synth-diy>
>> > Selling or trading? Use marketplace@synth-diy.org <mailto:marketplace@synth-diy.org>
>> > 
>> > ________________________________________________________
>> > This is the Synth-diy mailing list
>> > Submit email to: Synth-diy@synth-diy.org <mailto:Synth-diy@synth-diy.org>
>> > View archive at: https://synth-diy.org/pipermail/synth-diy/
>> > Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>> > Selling or trading? Use marketplace@synth-diy.org <mailto:marketplace@synth-diy.org>
>> ________________________________________________________
>> This is the Synth-diy mailing list
>> Submit email to: Synth-diy@synth-diy.org <mailto:Synth-diy@synth-diy.org>
>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>> Selling or trading? Use marketplace@synth-diy.org <mailto:marketplace@synth-diy.org>
> ________________________________________________________
> This is the Synth-diy mailing list
> Submit email to: Synth-diy@synth-diy.org
> View archive at: https://synth-diy.org/pipermail/synth-diy/
> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
> Selling or trading? Use marketplace@synth-diy.org

On Mar 19, 2026, at 2:11 PM, Tom Wiltshire wrote:
> The whole "Mux and S&H" thing was only because decent quality DACs were damn expensive, so it was better to buy one good one and re-use it a lot of times. If there'd been a sixteen channel DAC for the same price, you can bet they've have used it.
> 
> Incidentally, I once worked on a PWM-DAC based CV transposition module. We used one 3-bit PWM to generate octave voltages, and one 4-bit PWM to generate 12 note voltages. Since each PWM was very low resolution, they were able to be very high frequency and able to respond fast. And since the two channels were entirely separate, trimming was a simple matter of tweaking one to give octaves and the other to give semitones and done. Once added together, the whole thing could produce good quality Note CVs over a wide range without using any expensive DAC. Of course, the secret here is that it was semitones only, so the full MIDI range suddenly boils down to seven bit accuracy.
> 
> On 19 Mar 2026, at 19:30, Mike Bryant wrote:
>> Ah okay, the poly requirement makes sense.   But we ran PDM DACs at 8MHz using LS-TTL to give better than 16 bit audio for a single channel so you could still feed that to a MUX and S/H.
>> From: Roman Sowa <modular@go2.pl>
>> Sent: 19 March 2026 11:10
>> To: Mike Bryant <mbryant@futurehorizons.com>; synth-diy@synth-diy.org <synth-diy@synth-diy.org>
>> Subject: Re: [sdiy] Linear response VCOs?
>>  
>> Correct me if I'm wrong but old synths using DAC for CV were all 
>> polyhonic. That means a lot of CV sources needed. So they used S&H and 
>> muxed DAC. To have PWM with fast enough response to feed MUX and S$H, it 
>> would have to run at enormouse frequency, not suitable to affordable 
>> technology back then. And putting separate counter as PWM generator for 
>> every CV is much more expensive, and takes more space than DAC-MUX-S&H.
>> Back then if you wanted a timer, you got 8253 offering 3 timers in one 
>> package, and I'm not even sure if it had PWM mode at all.
>> 
>> Roman
>> 
>> W dniu 2026-03-18 o 21:39, Mike Bryant pisze:
>> > Does anybody know why these old synths didn't use PWM/PDM techniques ?
>> > 
>> > LS-TTL or CMOS feeding a comparator into an analogue integrator gave 12 
>> > bits performance at audio frequencies even in the 70s so CVs good enough 
>> > for tuning would have been easy.
>> > ------------------------------------------------------------------------
>> > *From:* brianw 
>> > *Sent:* 18 March 2026 19:02
>> > *To:* synth-diy@synth-diy.org <synth-diy@synth-diy.org>
>> > *Subject:* Re: [sdiy] Linear response VCOs?
>> > The Prophet 5, Rev 1 and Rev 2, use a 7-bit DAC made from hand-picked 
>> > resistors. There is a note in the Service Manual that you should *not* 
>> > replace these resistors because of the challenge of matching a new one 
>> > to the network. 1 LSB is calibrated to 1/12 V (0.0833 V) for easy use in 
>> > 1V/8va scaling. CV ranges from 0 V to 10.583 V (127/12), but the Prophet 
>> > 5 only uses the lower 6 bits for pitch, limiting the range to 5.333 V 
>> > maximum and thus 5 octaves. All CV were 7-bit, but the pitch combined 
>> > coarse and fine with the scale of the DAC changed so that there were 64 
>> > steps in the coarse range plus another 128 steps in the fine range. This 
>> > wasn't quite as accurate as a 13-bit DAC, but still quite accurate for 
>> > the time.
>> > 
>> > The Prophet 5 Rev 3 simply used a 16-bit DAC, but maintained the 
>> > firmware design with 7 bits per CV, so the pitch did not enjoy a full 
>> > 16-bit precision. The 13-bit pitch values still have 16-bit accuracy, 
>> > though, just not 65536 steps of precision.
>> > 
>> > One thing to note, Mark, is that a 6-bit DAC has an LSB that's 1.56% of 
>> > the total range, so 1% resistors would be quite awful. Then there's the 
>> > fact that a 1% error in the MSB could throw the whole binary scale off 
>> > enough that the values are not monotonic (i.e. an increase in the code 
>> > could actually cause a decrease in voltage!). A 7-bit DAC has the LSB at 
>> > 0.78% so you definitely need better than 1% precision. These 
>> > manufacturers were not making a custom resistor array so much as 
>> > hand-selecting individual resistors that were matched well across the 
>> > whole group.
>> > 
>> > Today, not only are 1% resistors more readily available than they were 
>> > in the seventies, but you can even get 0.1% tolerance resistors at a 
>> > reasonable. Still, that doesn't even get you to a full 9-bit DAC. This 
>> > illustrates how impressive DAC chip technology is. One of the fasted DAC 
>> > chips I've designed with can run at a sample rate of 125 MHz (yeah, MHz, 
>> > not kHz) based on current switching rather than voltage, but it stops at 
>> > 14-bit precision because the smallest current is only 0.0061% of the 
>> > largest, and it's difficult to be precise enough at such a large scale 
>> > factor. Larger DAC precision requires a different technique than 
>> > binary-weighted digits. Fortunately, there are many ways to implement a DAC.
>> > 
>> > Brian
>> > 
>> > 
>> > On Mar 18, 2026, at 4:34 AM, Tom Wiltshire wrote:
>> >> Roland had form for this. SH-101 uses a simple DAC built from a few resistors too.
>> >> 
>> >> Like Roman said, it doesn't really make sense nowadays when DACs are cheap, but it was worth it then.
>> >> 
>> >> Tom
>> >> 
>> >> On 18 Mar 2026, at 11:31, mark verbos wrote:
>> >>> Like a TR-909.
>> >>> But, surely it is cheaper to use 1% resistors rather than a custom resistor array made.
>> >>> 
>> >>> Mark
>> >>> 
>> >>> On Mar 17, 2026, at 18:44, David Manley wrote:
>> >>>> It's interesting to see how PAiA's John Simonton solved some these issues in the 1970's by having a custom laser trimmed resistor network built for their 6-bit "Equally Tempered DAC" to be used with linear VCOs.  See the bottom of the schematic on page  18, the resistor values are on the last page.
>> >>>> 
>> >>>> https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf 
>> > <https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf>
>> >>>> 
>> >>>> As is typical for PAiA a very low cost solution: build your own DAC with a few components.
>> >>>> 
>> >>>> -Dave
>> >

> On 20 Mar 2026, at 09:30, Mike Bryant <mbryant@futurehorizons.com> wrote:
> 
> This is why you use PDM, not PWM.  The pulses are at much higher frequency and easier to filter to the correct DC level with less noise.
> 
> From: Synth-diy <synth-diy-bounces@synth-diy.org <mailto:synth-diy-bounces@synth-diy.org>> on behalf of brianw <brianw@audiobanshee.com <mailto:brianw@audiobanshee.com>>
> Sent: 20 March 2026 08:22
> To: synth-diy@synth-diy.org <mailto:synth-diy@synth-diy.org> <synth-diy@synth-diy.org <mailto:synth-diy@synth-diy.org>>
> Subject: Re: [sdiy] Linear response VCOs?
>  
> The challenge with PWM is that changing a rail-to-rail square wave into a steady DC value requires a lot of filtering. That filter must remove the sharp rise and fall of the raw PWM output, and thus the DC output value cannot sharply rise or fall either. The problem gets worse if a single channel needs to feed multiple unrelated CV values through a mux+S&H. The slew rate is horrible.
> 
> Brian
> 
> 
> ________________________________________________________
> This is the Synth-diy mailing list
> Submit email to: Synth-diy@synth-diy.org <mailto:Synth-diy@synth-diy.org>
> View archive at: https://synth-diy.org/pipermail/synth-diy/ <https://synth-diy.org/pipermail/synth-diy/>
> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy <https://synth-diy.org/mailman/listinfo/synth-diy>
> Selling or trading? Use marketplace@synth-diy.org <mailto:marketplace@synth-diy.org>

From: Tom Wiltshire <tom@electricdruid.net>
Sent: 20 March 2026 10:37
To: Mike Bryant <mbryant@futurehorizons.com>
Cc: brianw <brianw@audiobanshee.com>; synth-diy@synth-diy.org <synth-diy@synth-diy.org>
Subject: Re: [sdiy] Linear response VCOs?

There's also differences in *where* the errors are worst between PWM and PDM.

Generally, you worry most about the fundamental and lowest harmonics of the pulse output frequency, since they're the hardest to filter. For PWM, you get the worst fundamental and 3rd harmonic at the midpoint value, when the output is a 50/50 square wave. You get the worst 2nd harmonic at the 25% and 75% values, since that's where that harmonic peaks for pulse waves. This is all fairly "central" and very likely to be values you're using all the time. In short, if your PWM output is at 100KHz, then 100KHz is what you'll get, and your filtering had better be able to deal with that.

For PDM, the central values are the best-case, rather than the worst. With PDM, the worst-case comes as you get to extreme values, so <5% or >95%. This happens because there's so few pulses going out that the effective output rate drops. Imagine we're using a 2MHz PDM output to create a 10-bit DAC. If we output our midpoint value of 512, we get a lovely squarewave at 1MHz - one period on, one period off. If we output a value of 1, we get one period on followed by 1023 periods off - a very narrow pulse wave at 2MHz/1024= 1953Hz. That's terrible! Of course, as you approach these extremes, the amount of fundamental and lower harmonics in such a narrow pulse drops off markedly, but still - you probably weren't thinking of a 2KHz output when you designed your 2MHz PDM DAC.
My view is that the secret with PDM is to discard the extremes and use the good bit in the middle!

Tom


On 20 Mar 2026, at 09:30, Mike Bryant <mbryant@futurehorizons.com<mailto:mbryant@futurehorizons.com>> wrote:

This is why you use PDM, not PWM.  The pulses are at much higher frequency and easier to filter to the correct DC level with less noise.

________________________________
From: Synth-diy <synth-diy-bounces@synth-diy.org<mailto:synth-diy-bounces@synth-diy.org>> on behalf of brianw <brianw@audiobanshee.com<mailto:brianw@audiobanshee.com>>
Sent: 20 March 2026 08:22
To: synth-diy@synth-diy.org<mailto:synth-diy@synth-diy.org> <synth-diy@synth-diy.org<mailto:synth-diy@synth-diy.org>>
Subject: Re: [sdiy] Linear response VCOs?

The challenge with PWM is that changing a rail-to-rail square wave into a steady DC value requires a lot of filtering. That filter must remove the sharp rise and fall of the raw PWM output, and thus the DC output value cannot sharply rise or fall either. The problem gets worse if a single channel needs to feed multiple unrelated CV values through a mux+S&H. The slew rate is horrible.

Brian


________________________________________________________
This is the Synth-diy mailing list
Submit email to: Synth-diy@synth-diy.org<mailto:Synth-diy@synth-diy.org>
View archive at: https://synth-diy.org/pipermail/synth-diy/
Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
Selling or trading? Use marketplace@synth-diy.org<mailto:marketplace@synth-diy.org>

> On Mar 20, 2026, at 7:39 AM, Mike Bryant <mbryant@futurehorizons.com> wrote:
> 
> Well yes I was actually doing frequencies like 1004Hz (a telecomms test freq) at 8MHz sampling, but as you say the secret with PDM is to stay well away from the peak positive and peak negative values.  However a little bit of high frequency noise injection (from a look up table) added to the digital sinewave helped as this also gets filtered out.
> From: Tom Wiltshire <tom@electricdruid.net>
> Sent: 20 March 2026 10:37
> To: Mike Bryant <mbryant@futurehorizons.com>
> Cc: brianw <brianw@audiobanshee.com>; synth-diy@synth-diy.org <synth-diy@synth-diy.org>
> Subject: Re: [sdiy] Linear response VCOs?
>  
> There's also differences in *where* the errors are worst between PWM and PDM.
> 
> Generally, you worry most about the fundamental and lowest harmonics of the pulse output frequency, since they're the hardest to filter. For PWM, you get the worst fundamental and 3rd harmonic at the midpoint value, when the output is a 50/50 square wave. You get the worst 2nd harmonic at the 25% and 75% values, since that's where that harmonic peaks for pulse waves. This is all fairly "central" and very likely to be values you're using all the time. In short, if your PWM output is at 100KHz, then 100KHz is what you'll get, and your filtering had better be able to deal with that.
> 
> For PDM, the central values are the best-case, rather than the worst. With PDM, the worst-case comes as you get to extreme values, so <5% or >95%. This happens because there's so few pulses going out that the effective output rate drops. Imagine we're using a 2MHz PDM output to create a 10-bit DAC. If we output our midpoint value of 512, we get a lovely squarewave at 1MHz - one period on, one period off. If we output a value of 1, we get one period on followed by 1023 periods off - a very narrow pulse wave at 2MHz/1024= 1953Hz. That's terrible! Of course, as you approach these extremes, the amount of fundamental and lower harmonics in such a narrow pulse drops off markedly, but still - you probably weren't thinking of a 2KHz output when you designed your 2MHz PDM DAC.
> My view is that the secret with PDM is to discard the extremes and use the good bit in the middle!
> 
> Tom
> 
> 
>> On 20 Mar 2026, at 09:30, Mike Bryant <mbryant@futurehorizons.com <mailto:mbryant@futurehorizons.com>> wrote:
>> 
>> This is why you use PDM, not PWM.  The pulses are at much higher frequency and easier to filter to the correct DC level with less noise.
>> 
>> From: Synth-diy <synth-diy-bounces@synth-diy.org <mailto:synth-diy-bounces@synth-diy.org>> on behalf of brianw <brianw@audiobanshee.com <mailto:brianw@audiobanshee.com>>
>> Sent: 20 March 2026 08:22
>> To: synth-diy@synth-diy.org <mailto:synth-diy@synth-diy.org> <synth-diy@synth-diy.org <mailto:synth-diy@synth-diy.org>>
>> Subject: Re: [sdiy] Linear response VCOs?
>>  
>> The challenge with PWM is that changing a rail-to-rail square wave into a steady DC value requires a lot of filtering. That filter must remove the sharp rise and fall of the raw PWM output, and thus the DC output value cannot sharply rise or fall either. The problem gets worse if a single channel needs to feed multiple unrelated CV values through a mux+S&H. The slew rate is horrible.
>> 
>> Brian
>> 
>> 
>> ________________________________________________________
>> This is the Synth-diy mailing list
>> Submit email to: Synth-diy@synth-diy.org <mailto:Synth-diy@synth-diy.org>
>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>> Selling or trading? Use marketplace@synth-diy.org <mailto:marketplace@synth-diy.org>
> ________________________________________________________
> This is the Synth-diy mailing list
> Submit email to: Synth-diy@synth-diy.org
> View archive at: https://synth-diy.org/pipermail/synth-diy/
> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
> Selling or trading? Use marketplace@synth-diy.org

> On 20 Mar 2026, at 21:18, Thomas Hudson <thomas.hudson7@icloud.com> wrote:
> 
> Thanks to everyone on this thread. I am currently working on a MIDI to Hertz program for the Raspberry Pi (though I can probably move to an ESP32). Two last questions. Can anyone tell me which commercial synths were linear response? I know about the Moog Taurus / Behringer Toro.
> 
> Lastly, can anyone suggest a good design / schematic / kit for a linear VCO? 
> 
> TIA
> 
>> On Mar 20, 2026, at 7:39 AM, Mike Bryant <mbryant@futurehorizons.com> wrote:
>> 
>> Well yes I was actually doing frequencies like 1004Hz (a telecomms test freq) at 8MHz sampling, but as you say the secret with PDM is to stay well away from the peak positive and peak negative values.  However a little bit of high frequency noise injection (from a look up table) added to the digital sinewave helped as this also gets filtered out.
>> From: Tom Wiltshire <tom@electricdruid.net>
>> Sent: 20 March 2026 10:37
>> To: Mike Bryant <mbryant@futurehorizons.com>
>> Cc: brianw <brianw@audiobanshee.com>; synth-diy@synth-diy.org <synth-diy@synth-diy.org>
>> Subject: Re: [sdiy] Linear response VCOs?
>>  
>> There's also differences in *where* the errors are worst between PWM and PDM.
>> 
>> Generally, you worry most about the fundamental and lowest harmonics of the pulse output frequency, since they're the hardest to filter. For PWM, you get the worst fundamental and 3rd harmonic at the midpoint value, when the output is a 50/50 square wave. You get the worst 2nd harmonic at the 25% and 75% values, since that's where that harmonic peaks for pulse waves. This is all fairly "central" and very likely to be values you're using all the time. In short, if your PWM output is at 100KHz, then 100KHz is what you'll get, and your filtering had better be able to deal with that.
>> 
>> For PDM, the central values are the best-case, rather than the worst. With PDM, the worst-case comes as you get to extreme values, so <5% or >95%. This happens because there's so few pulses going out that the effective output rate drops. Imagine we're using a 2MHz PDM output to create a 10-bit DAC. If we output our midpoint value of 512, we get a lovely squarewave at 1MHz - one period on, one period off. If we output a value of 1, we get one period on followed by 1023 periods off - a very narrow pulse wave at 2MHz/1024= 1953Hz. That's terrible! Of course, as you approach these extremes, the amount of fundamental and lower harmonics in such a narrow pulse drops off markedly, but still - you probably weren't thinking of a 2KHz output when you designed your 2MHz PDM DAC.
>> My view is that the secret with PDM is to discard the extremes and use the good bit in the middle!
>> 
>> Tom
>> 
>> 
>>> On 20 Mar 2026, at 09:30, Mike Bryant <mbryant@futurehorizons.com <mailto:mbryant@futurehorizons.com>> wrote:
>>> 
>>> This is why you use PDM, not PWM.  The pulses are at much higher frequency and easier to filter to the correct DC level with less noise.
>>> 
>>> From: Synth-diy <synth-diy-bounces@synth-diy.org <mailto:synth-diy-bounces@synth-diy.org>> on behalf of brianw <brianw@audiobanshee.com <mailto:brianw@audiobanshee.com>>
>>> Sent: 20 March 2026 08:22
>>> To: synth-diy@synth-diy.org <mailto:synth-diy@synth-diy.org> <synth-diy@synth-diy.org <mailto:synth-diy@synth-diy.org>>
>>> Subject: Re: [sdiy] Linear response VCOs?
>>>  
>>> The challenge with PWM is that changing a rail-to-rail square wave into a steady DC value requires a lot of filtering. That filter must remove the sharp rise and fall of the raw PWM output, and thus the DC output value cannot sharply rise or fall either. The problem gets worse if a single channel needs to feed multiple unrelated CV values through a mux+S&H. The slew rate is horrible.
>>> 
>>> Brian
>>> 
>>> 
>>> ________________________________________________________
>>> This is the Synth-diy mailing list
>>> Submit email to: Synth-diy@synth-diy.org <mailto:Synth-diy@synth-diy.org>
>>> View archive at: https://synth-diy.org/pipermail/synth-diy/ <https://synth-diy.org/pipermail/synth-diy/>
>>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy <https://synth-diy.org/mailman/listinfo/synth-diy>
>>> Selling or trading? Use marketplace@synth-diy.org <mailto:marketplace@synth-diy.org>
>> ________________________________________________________
>> This is the Synth-diy mailing list
>> Submit email to: Synth-diy@synth-diy.org
>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>> Selling or trading? Use marketplace@synth-diy.org
>

> Thanks to everyone on this thread. I am currently working on a MIDI to 
> Hertz program for the Raspberry Pi (though I can probably move to an 
> ESP32). Two last questions. Can anyone tell me which commercial synths 
> were linear response? I know about the Moog Taurus / Behringer Toro.
>
> Lastly, can anyone suggest a good design / schematic / kit for a 
> linear VCO?
>
> TIA
>
>> On Mar 20, 2026, at 7:39 AM, Mike Bryant <mbryant@futurehorizons.com> 
>> wrote:
>>
>> Well yes I was actually doing frequencies like 1004Hz (a telecomms 
>> test freq) at 8MHz sampling, but as you say the secret with PDM is to 
>> stay well away from the peak positive and peak negative values.  
>> However a little bit of high frequency noise injection (from a look 
>> up table) added to the digital sinewave helped as this also gets 
>> filtered out.
>> ------------------------------------------------------------------------
>> *From:*Tom Wiltshire <tom@electricdruid.net>
>> *Sent:*20 March 2026 10:37
>> *To:*Mike Bryant <mbryant@futurehorizons.com>
>> *Cc:*brianw <brianw@audiobanshee.com>; synth-diy@synth-diy.org 
>> <synth-diy@synth-diy.org>
>> *Subject:*Re: [sdiy] Linear response VCOs?
>> There's also differences in *where* the errors are worst between PWM 
>> and PDM.
>>
>> Generally, you worry most about the fundamental and lowest harmonics 
>> of the pulse output frequency, since they're the hardest to filter. 
>> For PWM, you get the worst fundamental and 3rd harmonic at the 
>> midpoint value, when the output is a 50/50 square wave. You get the 
>> worst 2nd harmonic at the 25% and 75% values, since that's where that 
>> harmonic peaks for pulse waves. This is all fairly "central" and very 
>> likely to be values you're using all the time. In short, if your PWM 
>> output is at 100KHz, then 100KHz is what you'll get, and your 
>> filtering had better be able to deal with that.
>>
>> For PDM, the central values are the best-case, rather than the worst. 
>> With PDM, the worst-case comes as you get to extreme values, so <5% 
>> or >95%. This happens because there's so few pulses going out that 
>> the effective output rate drops. Imagine we're using a 2MHz PDM 
>> output to create a 10-bit DAC. If we output our midpoint value of 
>> 512, we get a lovely squarewave at 1MHz - one period on, one period 
>> off. If we output a value of 1, we get one period on followed by 1023 
>> periods off - a very narrow pulse wave at 2MHz/1024= 1953Hz. That's 
>> terrible! Of course, as you approach these extremes, the amount of 
>> fundamental and lower harmonics in such a narrow pulse drops off 
>> markedly, but still - you probably weren't thinking of a 2KHz output 
>> when you designed your 2MHz PDM DAC.
>> My view is that the secret with PDM is to discard the extremes and 
>> use the good bit in the middle!
>>
>> Tom
>>
>>
>>> On 20 Mar 2026, at 09:30, Mike Bryant <mbryant@futurehorizons.com> 
>>> wrote:
>>>
>>> This is why you use PDM, not PWM.  The pulses are at much higher 
>>> frequency and easier to filter to the correct DC level with less noise.
>>>
>>> ------------------------------------------------------------------------
>>> *From:* Synth-diy <synth-diy-bounces@synth-diy.org> on behalf of 
>>> brianw <brianw@audiobanshee.com>
>>> *Sent:* 20 March 2026 08:22
>>> *To:* synth-diy@synth-diy.org<synth-diy@synth-diy.org>
>>> *Subject:* Re: [sdiy] Linear response VCOs?
>>> The challenge with PWM is that changing a rail-to-rail square wave 
>>> into a steady DC value requires a lot of filtering. That filter must 
>>> remove the sharp rise and fall of the raw PWM output, and thus the 
>>> DC output value cannot sharply rise or fall either. The problem gets 
>>> worse if a single channel needs to feed multiple unrelated CV values 
>>> through a mux+S&H. The slew rate is horrible.
>>>
>>> Brian
>>>
>>>
>>> ________________________________________________________
>>> This is the Synth-diy mailing list
>>> Submit email to:Synth-diy@synth-diy.org
>>> View archive at:https://synth-diy.org/pipermail/synth-diy/
>>> Check your settings at:https://synth-diy.org/mailman/listinfo/synth-diy
>>> Selling or trading? Usemarketplace@synth-diy.org
>>
>> ________________________________________________________
>> This is the Synth-diy mailing list
>> Submit email to: Synth-diy@synth-diy.org
>> View archive at: https://synth-diy.org/pipermail/synth-diy/
>> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
>> Selling or trading? Use marketplace@synth-diy.org
>
>
> ________________________________________________________
> This is the Synth-diy mailing list
> Submit email to: Synth-diy@synth-diy.org
> View archive at: https://synth-diy.org/pipermail/synth-diy/
> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
> Selling or trading? Use marketplace@synth-diy.org

On 3/21/26 12:18 PM, Gordonjcp wrote:
> On Fri, Mar 20, 2026 at 10:40:57PM +0000, Tom Wiltshire wrote:
>> The Korg Polysix kind-of was. The VCOs were linear, and there was an expo convertor that was multiplexed out to each of them, if I remember. There's a schematic I drew of the VCO design flying around on the web.
>>
> How do they get tuned, and what's the kind of optocoupler thing associated with CV generation for?
>

On 3/22/26 7:12 AM, Michael E Caloroso via Synth-diy wrote:
>  > Correct me if I'm wrong but old synths using DAC for CV were all
>  > polyphonic.
> 
> If it was programmable using solid state memory, it had DAC with MUX/ 
> S&H.  Not limited to just polyphonics.
> 
> Oberheim OB-1 monophonic was programmable and used a DAC with MUX/S&H 
> for CV.  Released in 1977.
> 
> Moog Source was another one, released in 1980.
> 
> MC
> 
> On Thu, Mar 19, 2026 at 7:14 AM Roman Sowa via Synth-diy <synth- 
> diy@synth-diy.org <mailto:synth-diy@synth-diy.org>> wrote:
> 
>     Correct me if I'm wrong but old synths using DAC for CV were all
>     polyhonic. That means a lot of CV sources needed. So they used S&H and
>     muxed DAC. To have PWM with fast enough response to feed MUX and
>     S$H, it
>     would have to run at enormouse frequency, not suitable to affordable
>     technology back then. And putting separate counter as PWM generator for
>     every CV is much more expensive, and takes more space than DAC-MUX-S&H.
>     Back then if you wanted a timer, you got 8253 offering 3 timers in one
>     package, and I'm not even sure if it had PWM mode at all.
> 
>     Roman
> 
>     W dniu 2026-03-18 o 21:39, Mike Bryant pisze:
>      > Does anybody know why these old synths didn't use PWM/PDM
>     techniques ?
>      >
>      > LS-TTL or CMOS feeding a comparator into an analogue integrator
>     gave 12
>      > bits performance at audio frequencies even in the 70s so CVs good
>     enough
>      > for tuning would have been easy.
>      >
>     ------------------------------------------------------------------------
>      > *From:* Synth-diy <synth-diy-bounces@synth-diy.org <mailto:synth-
>     diy-bounces@synth-diy.org>> on behalf of brianw
>      > <brianw@audiobanshee.com <mailto:brianw@audiobanshee.com>>
>      > *Sent:* 18 March 2026 19:02
>      > *To:* synth-diy@synth-diy.org <mailto:synth-diy@synth-diy.org>
>     <synth-diy@synth-diy.org <mailto:synth-diy@synth-diy.org>>
>      > *Subject:* Re: [sdiy] Linear response VCOs?
>      > The Prophet 5, Rev 1 and Rev 2, use a 7-bit DAC made from hand-
>     picked
>      > resistors. There is a note in the Service Manual that you should
>     *not*
>      > replace these resistors because of the challenge of matching a
>     new one
>      > to the network. 1 LSB is calibrated to 1/12 V (0.0833 V) for easy
>     use in
>      > 1V/8va scaling. CV ranges from 0 V to 10.583 V (127/12), but the
>     Prophet
>      > 5 only uses the lower 6 bits for pitch, limiting the range to
>     5.333 V
>      > maximum and thus 5 octaves. All CV were 7-bit, but the pitch
>     combined
>      > coarse and fine with the scale of the DAC changed so that there
>     were 64
>      > steps in the coarse range plus another 128 steps in the fine
>     range. This
>      > wasn't quite as accurate as a 13-bit DAC, but still quite
>     accurate for
>      > the time.
>      >
>      > The Prophet 5 Rev 3 simply used a 16-bit DAC, but maintained the
>      > firmware design with 7 bits per CV, so the pitch did not enjoy a
>     full
>      > 16-bit precision. The 13-bit pitch values still have 16-bit
>     accuracy,
>      > though, just not 65536 steps of precision.
>      >
>      > One thing to note, Mark, is that a 6-bit DAC has an LSB that's
>     1.56% of
>      > the total range, so 1% resistors would be quite awful. Then
>     there's the
>      > fact that a 1% error in the MSB could throw the whole binary
>     scale off
>      > enough that the values are not monotonic (i.e. an increase in the
>     code
>      > could actually cause a decrease in voltage!). A 7-bit DAC has the
>     LSB at
>      > 0.78% so you definitely need better than 1% precision. These
>      > manufacturers were not making a custom resistor array so much as
>      > hand-selecting individual resistors that were matched well across
>     the
>      > whole group.
>      >
>      > Today, not only are 1% resistors more readily available than they
>     were
>      > in the seventies, but you can even get 0.1% tolerance resistors at a
>      > reasonable. Still, that doesn't even get you to a full 9-bit DAC.
>     This
>      > illustrates how impressive DAC chip technology is. One of the
>     fasted DAC
>      > chips I've designed with can run at a sample rate of 125 MHz
>     (yeah, MHz,
>      > not kHz) based on current switching rather than voltage, but it
>     stops at
>      > 14-bit precision because the smallest current is only 0.0061% of the
>      > largest, and it's difficult to be precise enough at such a large
>     scale
>      > factor. Larger DAC precision requires a different technique than
>      > binary-weighted digits. Fortunately, there are many ways to
>     implement a DAC.
>      >
>      > Brian
>      >
>      >
>      > On Mar 18, 2026, at 4:34 AM, Tom Wiltshire wrote:
>      >> Roland had form for this. SH-101 uses a simple DAC built from a
>     few resistors too.
>      >>
>      >> Like Roman said, it doesn't really make sense nowadays when DACs
>     are cheap, but it was worth it then.
>      >>
>      >> Tom
>      >>
>      >> On 18 Mar 2026, at 11:31, mark verbos wrote:
>      >>> Like a TR-909.
>      >>> But, surely it is cheaper to use 1% resistors rather than a
>     custom resistor array made.
>      >>>
>      >>> Mark
>      >>>
>      >>> On Mar 17, 2026, at 18:44, David Manley wrote:
>      >>>> It's interesting to see how PAiA's John Simonton solved some
>     these issues in the 1970's by having a custom laser trimmed resistor
>     network built for their 6-bit "Equally Tempered DAC" to be used with
>     linear VCOs.  See the bottom of the schematic on page  18, the
>     resistor values are on the last page.
>      >>>>
>      >>>> https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf
>     <https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf>
>      > <https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf
>     <https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf>>
>      >>>>
>      >>>> As is typical for PAiA a very low cost solution: build your
>     own DAC with a few components.
>      >>>>
>      >>>> -Dave
>      >
>      >
>      > ________________________________________________________
>      > This is the Synth-diy mailing list
>      > Submit email to: Synth-diy@synth-diy.org <mailto:Synth-diy@synth-
>     diy.org>
>      > View archive at: https://synth-diy.org/pipermail/synth-diy/
>     <https://synth-diy.org/pipermail/synth-diy/>
>      > <https://synth-diy.org/pipermail/synth-diy/ <https://synth-
>     diy.org/pipermail/synth-diy/>>
>      > Check your settings at: https://synth-diy.org/mailman/listinfo/
>     synth-diy <https://synth-diy.org/mailman/listinfo/synth-diy>
>      > <https://synth-diy.org/mailman/listinfo/synth-diy <https://synth-
>     diy.org/mailman/listinfo/synth-diy>>
>      > Selling or trading? Use marketplace@synth-diy.org
>     <mailto:marketplace@synth-diy.org>
>      >
>      > ________________________________________________________
>      > This is the Synth-diy mailing list
>      > Submit email to: Synth-diy@synth-diy.org <mailto:Synth-diy@synth-
>     diy.org>
>      > View archive at: https://synth-diy.org/pipermail/synth-diy/
>     <https://synth-diy.org/pipermail/synth-diy/>
>      > Check your settings at: https://synth-diy.org/mailman/listinfo/
>     synth-diy <https://synth-diy.org/mailman/listinfo/synth-diy>
>      > Selling or trading? Use marketplace@synth-diy.org
>     <mailto:marketplace@synth-diy.org>
>     ________________________________________________________
>     This is the Synth-diy mailing list
>     Submit email to: Synth-diy@synth-diy.org <mailto:Synth-diy@synth-
>     diy.org>
>     View archive at: https://synth-diy.org/pipermail/synth-diy/
>     <https://synth-diy.org/pipermail/synth-diy/>
>     Check your settings at: https://synth-diy.org/mailman/listinfo/
>     synth-diy <https://synth-diy.org/mailman/listinfo/synth-diy>
>     Selling or trading? Use marketplace@synth-diy.org
>     <mailto:marketplace@synth-diy.org>
> 
> 
> ________________________________________________________
> This is the Synth-diy mailing list
> Submit email to: Synth-diy@synth-diy.org
> View archive at: https://synth-diy.org/pipermail/synth-diy/
> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
> Selling or trading? Use marketplace@synth-diy.org

On Mar 22, 2026, at 3:44 PM, Olav Kvern wrote:
> The Sequential Pro-One is another mono synth with a DAC. I still think that the way it's done is clever.
> 
> Thanks,
> 
> Ole
> 
> On 3/22/26 7:12 AM, Michael E Caloroso wrote:
>> > Correct me if I'm wrong but old synths using DAC for CV were all
>> > polyphonic.
>> If it was programmable using solid state memory, it had DAC with MUX/ S&H.  Not limited to just polyphonics.
>> Oberheim OB-1 monophonic was programmable and used a DAC with MUX/S&H for CV.  Released in 1977.
>> Moog Source was another one, released in 1980.
>> MC
>> On Thu, Mar 19, 2026 at 7:14 AM Roman Sowa wrote:
>>    Correct me if I'm wrong but old synths using DAC for CV were all
>>    polyhonic. That means a lot of CV sources needed. So they used S&H and
>>    muxed DAC. To have PWM with fast enough response to feed MUX and
>>    S$H, it
>>    would have to run at enormouse frequency, not suitable to affordable
>>    technology back then. And putting separate counter as PWM generator for
>>    every CV is much more expensive, and takes more space than DAC-MUX-S&H.
>>    Back then if you wanted a timer, you got 8253 offering 3 timers in one
>>    package, and I'm not even sure if it had PWM mode at all.
>>    Roman
>>    W dniu 2026-03-18 o 21:39, Mike Bryant pisze:
>>     > Does anybody know why these old synths didn't use PWM/PDM
>>    techniques ?
>>     >
>>     > LS-TTL or CMOS feeding a comparator into an analogue integrator
>>    gave 12
>>     > bits performance at audio frequencies even in the 70s so CVs good
>>    enough
>>     > for tuning would have been easy.
>>     >
>>    ------------------------------------------------------------------------
>>     > From: brianw
>>     > *Sent:* 18 March 2026 19:02
>>     > The Prophet 5, Rev 1 and Rev 2, use a 7-bit DAC made from hand-picked
>>     > resistors. There is a note in the Service Manual that you should *not*
>>     > replace these resistors because of the challenge of matching a new one
>>     > to the network. 1 LSB is calibrated to 1/12 V (0.0833 V) for easy use in
>>     > 1V/8va scaling. CV ranges from 0 V to 10.583 V (127/12), but the Prophet
>>     > 5 only uses the lower 6 bits for pitch, limiting the range to 5.333 V
>>     > maximum and thus 5 octaves. All CV were 7-bit, but the pitch combined
>>     > coarse and fine with the scale of the DAC changed so that there were 64
>>     > steps in the coarse range plus another 128 steps in the fine range. This
>>     > wasn't quite as accurate as a 13-bit DAC, but still quite accurate for
>>     > the time.
>>     >
>>     > The Prophet 5 Rev 3 simply used a 16-bit DAC, but maintained the
>>     > firmware design with 7 bits per CV, so the pitch did not enjoy a full
>>     > 16-bit precision. The 13-bit pitch values still have 16-bit accuracy,
>>     > though, just not 65536 steps of precision.
>>     >
>>     > One thing to note, Mark, is that a 6-bit DAC has an LSB that's 1.56% of
>>     > the total range, so 1% resistors would be quite awful. Then there's the
>>     > fact that a 1% error in the MSB could throw the whole binary scale off
>>     > enough that the values are not monotonic (i.e. an increase in the code
>>     > could actually cause a decrease in voltage!). A 7-bit DAC has the LSB at
>>     > 0.78% so you definitely need better than 1% precision. These
>>     > manufacturers were not making a custom resistor array so much as
>>     > hand-selecting individual resistors that were matched well across the
>>     > whole group.
>>     >
>>     > Today, not only are 1% resistors more readily available than they were
>>     > in the seventies, but you can even get 0.1% tolerance resistors at a
>>     > reasonable. Still, that doesn't even get you to a full 9-bit DAC. This
>>     > illustrates how impressive DAC chip technology is. One of the fasted DAC
>>     > chips I've designed with can run at a sample rate of 125 MHz (yeah, MHz,
>>     > not kHz) based on current switching rather than voltage, but it stops at
>>     > 14-bit precision because the smallest current is only 0.0061% of the
>>     > largest, and it's difficult to be precise enough at such a large scale
>>     > factor. Larger DAC precision requires a different technique than
>>     > binary-weighted digits. Fortunately, there are many ways to implement a DAC.
>>     >
>>     > Brian
>>     >
>>     >
>>     > On Mar 18, 2026, at 4:34 AM, Tom Wiltshire wrote:
>>     >> Roland had form for this. SH-101 uses a simple DAC built from a few resistors too.
>>     >>
>>     >> Like Roman said, it doesn't really make sense nowadays when DACs are cheap, but it was worth it then.
>>     >>
>>     >> Tom
>>     >>
>>     >> On 18 Mar 2026, at 11:31, mark verbos wrote:
>>     >>> Like a TR-909.
>>     >>> But, surely it is cheaper to use 1% resistors rather than a custom resistor array made.
>>     >>>
>>     >>> Mark
>>     >>>
>>     >>> On Mar 17, 2026, at 18:44, David Manley wrote:
>>     >>>> It's interesting to see how PAiA's John Simonton solved some these issues in the 1970's by having a custom laser trimmed resistor network built for their 6-bit "Equally Tempered DAC" to be used with linear VCOs.  See the bottom of the schematic on page  18, the resistor values are on the last page.
>>     >>>>
>>     >>>> https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf
>>    <https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf>
>>     > <https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf
>>    <https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf>>
>>     >>>>
>>     >>>> As is typical for PAiA a very low cost solution: build your
>>    own DAC with a few components.
>>     >>>>
>>     >>>> -Dave
>>     >

On 3/22/26 6:11 PM, brianw wrote:
> Thanks for mentioning the Pro-One DAC, Olav, because I had not looked at these details before (and I own one!)
> 
> Which aspects of the Pro-One DAC design do you consider to be clever?
> 
> 
> Choosing an 8-bit DAC, but only engaging the upper 6 bits?
> 
> Stealing the 6-column 8021 Port 0 outputs from the keyboard matrix to double as DAC parallel data inputs?
> 
> Calibrating a laser-calibrated DAC, where each step used is normally 40 mV, so that it actually has 41.667 mV (one quarter step)? *
> 
> Placing a CPU between the keyboard scan and the VCO input, so that sequencing and transposing are simple to achieve?
> 
> 
> [Further Ramblings]
> 
> I believe that reading multiple notes from a keyboard for polyphonic voice assignment requires a CPU. At the very least it requires matrix wiring of the keys, some sort of digital scan, and a method to deliver a unique pitch CV to a selected voice. That's probably extremely difficult without a CPU unless the voice assignment algorithm is baked into the logic design. I think that explains why most polyphonic synths use both a CPU and a DAC, even though nothing requires that they (CPU & DAC) can't each be of benefit on their own.
> 
> I'm excluding duo-phonic keyboard wirings that can read both a low-note-priority and a high-note-priority CV from the same set of keys for a two-voice architecture; and I'm excluding full-polyphony keyboards where each key has a dedicated voice. Those designs do not require a CPU, of course.
> 
> The Pro-One is a bit of an exception, here, since it's monophonic but still has a CPU to read the keyboard. One side effect of this is the ease with which a sequencer with transposition can be implemented. I suspect that the fact that the Prophet 5 was designed (in 1978) before the Pro-One (in 1981) meant that they were already familiar with using a CPU to scan a keyboard, so the fact that it wasn't necessary for a monophonic keyboard was moot. They just used the technology that they already knew, and enjoyed the advantages that come with that design. In fact, I recall that Dave Smith came up with the idea to use a CPU at a time (1975 for the Sequential Circuits Model 800) when the synth industry was not doing that.
> 
> Brian
> 
> 
> * Note that the AD558 DAC incorporated laser-trimmed resistors. Normally, the Vout, Vout-sense, and Vout-select pins are all shorted together to produce exactly 10 mV per step in full 8-bit mode. By placing a resistor and trim pot in series between Vout and Vout-sense, I assume that the Pro-One tweaks this to 10.41667 mV. Then, using only the upper 6 bits of the input, this can output in steps of 41.667 mV, up to 2.635 V total. I haven't figured out the gain of the 3280 + TL082 op-amp pair, but they might double that voltage to the expected half step resolution.
> 
> 
> On Mar 22, 2026, at 3:44 PM, Olav Kvern wrote:
>> The Sequential Pro-One is another mono synth with a DAC. I still think that the way it's done is clever.
>>
>> Thanks,
>>
>> Ole
>>
>> On 3/22/26 7:12 AM, Michael E Caloroso wrote:
>>>> Correct me if I'm wrong but old synths using DAC for CV were all
>>>> polyphonic.
>>> If it was programmable using solid state memory, it had DAC with MUX/ S&H.  Not limited to just polyphonics.
>>> Oberheim OB-1 monophonic was programmable and used a DAC with MUX/S&H for CV.  Released in 1977.
>>> Moog Source was another one, released in 1980.
>>> MC
>>> On Thu, Mar 19, 2026 at 7:14 AM Roman Sowa wrote:
>>>     Correct me if I'm wrong but old synths using DAC for CV were all
>>>     polyhonic. That means a lot of CV sources needed. So they used S&H and
>>>     muxed DAC. To have PWM with fast enough response to feed MUX and
>>>     S$H, it
>>>     would have to run at enormouse frequency, not suitable to affordable
>>>     technology back then. And putting separate counter as PWM generator for
>>>     every CV is much more expensive, and takes more space than DAC-MUX-S&H.
>>>     Back then if you wanted a timer, you got 8253 offering 3 timers in one
>>>     package, and I'm not even sure if it had PWM mode at all.
>>>     Roman
>>>     W dniu 2026-03-18 o 21:39, Mike Bryant pisze:
>>>      > Does anybody know why these old synths didn't use PWM/PDM
>>>     techniques ?
>>>      >
>>>      > LS-TTL or CMOS feeding a comparator into an analogue integrator
>>>     gave 12
>>>      > bits performance at audio frequencies even in the 70s so CVs good
>>>     enough
>>>      > for tuning would have been easy.
>>>      >
>>>     ------------------------------------------------------------------------
>>>      > From: brianw
>>>      > *Sent:* 18 March 2026 19:02
>>>      > The Prophet 5, Rev 1 and Rev 2, use a 7-bit DAC made from hand-picked
>>>      > resistors. There is a note in the Service Manual that you should *not*
>>>      > replace these resistors because of the challenge of matching a new one
>>>      > to the network. 1 LSB is calibrated to 1/12 V (0.0833 V) for easy use in
>>>      > 1V/8va scaling. CV ranges from 0 V to 10.583 V (127/12), but the Prophet
>>>      > 5 only uses the lower 6 bits for pitch, limiting the range to 5.333 V
>>>      > maximum and thus 5 octaves. All CV were 7-bit, but the pitch combined
>>>      > coarse and fine with the scale of the DAC changed so that there were 64
>>>      > steps in the coarse range plus another 128 steps in the fine range. This
>>>      > wasn't quite as accurate as a 13-bit DAC, but still quite accurate for
>>>      > the time.
>>>      >
>>>      > The Prophet 5 Rev 3 simply used a 16-bit DAC, but maintained the
>>>      > firmware design with 7 bits per CV, so the pitch did not enjoy a full
>>>      > 16-bit precision. The 13-bit pitch values still have 16-bit accuracy,
>>>      > though, just not 65536 steps of precision.
>>>      >
>>>      > One thing to note, Mark, is that a 6-bit DAC has an LSB that's 1.56% of
>>>      > the total range, so 1% resistors would be quite awful. Then there's the
>>>      > fact that a 1% error in the MSB could throw the whole binary scale off
>>>      > enough that the values are not monotonic (i.e. an increase in the code
>>>      > could actually cause a decrease in voltage!). A 7-bit DAC has the LSB at
>>>      > 0.78% so you definitely need better than 1% precision. These
>>>      > manufacturers were not making a custom resistor array so much as
>>>      > hand-selecting individual resistors that were matched well across the
>>>      > whole group.
>>>      >
>>>      > Today, not only are 1% resistors more readily available than they were
>>>      > in the seventies, but you can even get 0.1% tolerance resistors at a
>>>      > reasonable. Still, that doesn't even get you to a full 9-bit DAC. This
>>>      > illustrates how impressive DAC chip technology is. One of the fasted DAC
>>>      > chips I've designed with can run at a sample rate of 125 MHz (yeah, MHz,
>>>      > not kHz) based on current switching rather than voltage, but it stops at
>>>      > 14-bit precision because the smallest current is only 0.0061% of the
>>>      > largest, and it's difficult to be precise enough at such a large scale
>>>      > factor. Larger DAC precision requires a different technique than
>>>      > binary-weighted digits. Fortunately, there are many ways to implement a DAC.
>>>      >
>>>      > Brian
>>>      >
>>>      >
>>>      > On Mar 18, 2026, at 4:34 AM, Tom Wiltshire wrote:
>>>      >> Roland had form for this. SH-101 uses a simple DAC built from a few resistors too.
>>>      >>
>>>      >> Like Roman said, it doesn't really make sense nowadays when DACs are cheap, but it was worth it then.
>>>      >>
>>>      >> Tom
>>>      >>
>>>      >> On 18 Mar 2026, at 11:31, mark verbos wrote:
>>>      >>> Like a TR-909.
>>>      >>> But, surely it is cheaper to use 1% resistors rather than a custom resistor array made.
>>>      >>>
>>>      >>> Mark
>>>      >>>
>>>      >>> On Mar 17, 2026, at 18:44, David Manley wrote:
>>>      >>>> It's interesting to see how PAiA's John Simonton solved some these issues in the 1970's by having a custom laser trimmed resistor network built for their 6-bit "Equally Tempered DAC" to be used with linear VCOs.  See the bottom of the schematic on page  18, the resistor values are on the last page.
>>>      >>>>
>>>      >>>> https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf
>>>     <https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf>
>>>      > <https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf
>>>     <https://paia.com/wp-content/uploads/2024/05/8780pgs.pdf>>
>>>      >>>>
>>>      >>>> As is typical for PAiA a very low cost solution: build your
>>>     own DAC with a few components.
>>>      >>>>
>>>      >>>> -Dave
>>>      >
> 
> ________________________________________________________
> This is the Synth-diy mailing list
> Submit email to: Synth-diy@synth-diy.org
> View archive at: https://synth-diy.org/pipermail/synth-diy/
> Check your settings at: https://synth-diy.org/mailman/listinfo/synth-diy
> Selling or trading? Use marketplace@synth-diy.org