Discussion about the Korg PolySix synthesizer

Hello everyone;

I'm currently trying to fix two Polysix (including fittin a new processor
board) and both show similar behaviour. I am not sure if this is actually
normal (since the oscillators are true VCO's and not DCO's) or if there is
something wrong with those units.
All calibrations went very well and the unit is good in tune. However it
displays a very slight irregular "trembling" or "wavering" in the tuning -
sort of like a very small random modulation. It's a bit as if the pitch
jumps a little bit to the side and is immediately pulled back to the
regular pitch. When playing a chord or using the modulation effect it
becomes practically inaudible; the pitch deviation is perhaps 1 cent (when
it happens).
First I thought that this could come from some instability in the power
supply. However when making the filter self-oscillate the generated tone is
completely stable without the "trembling" phenomenom hence I don't think
this is a power supply issue.

Since it's been some time since I last worked on a Polysix: Is that perhaps
even normal and in the nature of the VCO's? Or is this a known problem?

Thanks,

Malte

I’ve only heard symptoms like this after severe battery leak damage. The VCO pitch modulation control voltage coming from the KLM-367 was picking up some unwanted signals in damaged areas. With an oscilloscope I could see that the pitch voltage looked strangely fuzzy compared to other DC signals. However, this should not happen with a new processor board. It could be a bad ground connection.
Don B.

From: mailto:PolySix@yahoogroups.com 
Sent: Tuesday, June 28, 2016 11:19 AM
To: PolySix@yahoogroups.com 
Subject: [PolySix] tuning instability?

  
Hello everyone;

I'm currently trying to fix two Polysix (including fittin a new processor
board) and both show similar behaviour. I am not sure if this is actually
normal (since the oscillators are true VCO's and not DCO's) or if there is
something wrong with those units.
All calibrations went very well and the unit is good in tune. However it
displays a very slight irregular "trembling" or "wavering" in the tuning -
sort of like a very small random modulation. It's a bit as if the pitch
jumps a little bit to the side and is immediately pulled back to the
regular pitch. When playing a chord or using the modulation effect it
becomes practically inaudible; the pitch deviation is perhaps 1 cent (when
it happens).
First I thought that this could come from some instability in the power
supply. However when making the filter self-oscillate the generated tone is
completely stable without the "trembling" phenomenom hence I don't think
this is a power supply issue.

Since it's been some time since I last worked on a Polysix: Is that perhaps
even normal and in the nature of the VCO's? Or is this a known problem?

Thanks,

Malte

Thank you, that was a valuable tip (even though it hasn't solved my problem
completely).
There certainly is something coming in over the VCO Mod line but it's only
slightly fuzzier than for example the VCF Mod line (and the self
oscillating filter is stable).
Comparing an original board and a clone in this area uncovered another
discrepancy: The parts list for the clone specifies R10 as 10k with 5%
tolerance while the schematic specifies 10k with 1% tolerance (which is
also used for the original KLM367). My R10 was actually measuring 9.88k
which would be slightly outside the 1%. Swapped it for a 10k resistor and I
believe the situation has slightly improved, at least for the sawtooth.
The PW is still somewhat shaky and I suspect some unwanted PW modulation
which of course could also result in such perceived instability.
Here's an interesting thing which naturally completely slipped by me when
doing the calibration: Is it normal that there is a huge AC signal on the
PW/PWM test point when PW is set to zero? It works quite well for non-zero
amounts but when going to zero there is a sine wave with 1.275MHz and
roughly 500mV peak-to-peak. This of course must somehow influence the PW.

Any ideas?

Thanks and best regards,

Malte




I've only heard symptoms like this after severe battery leak damage. The
VCO pitch modulation control voltage coming from the KLM-367 was picking up
some unwanted signals in damaged areas. With an oscilloscope I could see
that the pitch voltage looked strangely fuzzy compared to other DC signals.
However, this should not happen with a new processor board. It could be a
bad ground connection.
Don B.

From: <mailto:PolySix@yahoogroups.com>mailto:PolySix@yahoogroups.com
Sent: Tuesday, June 28, 2016 11:19 AM
To: <mailto:PolySix@yahoogroups.com>PolySix@yahoogroups.com
Subject: [PolySix] tuning instability?



Hello everyone;

I'm currently trying to fix two Polysix (including fittin a new processor
board) and both show similar behaviour. I am not sure if this is actually
normal (since the oscillators are true VCO's and not DCO's) or if there is
something wrong with those units.
All calibrations went very well and the unit is good in tune. However it
displays a very slight irregular "trembling" or "wavering" in the tuning -
sort of like a very small random modulation. It's a bit as if the pitch
jumps a little bit to the side and is immediately pulled back to the
regular pitch. When playing a chord or using the modulation effect it
becomes practically inaudible; the pitch deviation is perhaps 1 cent (when
it happens).
First I thought that this could come from some instability in the power
supply. However when making the filter self-oscillate the generated tone is
completely stable without the "trembling" phenomenom hence I don't think
this is a power supply issue.

Since it's been some time since I last worked on a Polysix: Is that perhaps
even normal and in the nature of the VCO's? Or is this a known problem?

Thanks,

Malte

Hi Malte

On 28.06.16 23:59 , Malte Rogacki gacki@gacki.sax.de [PolySix] wrote:
Is it normal that there is a huge AC signal on the
> PW/PWM test point when PW is set to zero? It works quite well for non-zero
> amounts but when going to zero there is a sine wave with 1.275MHz and
> roughly 500mV peak-to-peak. This of course must somehow influence the PW.
This is not normal, but it might be caused by getting the ground for the 
scope from a wrong / too far away point.

Florian

Update:

I've checked with my second clone board - same behaviour for PW zero.
Fortunately I still have one somewhat battered original board - there is a
bit fuzz on the PW signal but not this very clear sine signal as with my
two clone boards. So either it's something I did wrong when assembling the
clone boards - or possibly another clone board error.
I've also tried both old and new boards in a number of different Polysix
chassis with different voice boards - all showed some kind of "fluttering"
to some degree. The "fluttering" is not related to the LFO or PWM speed.

Next update:

After practically everything else seemed to check out I've moved one of the
original opamps to the clone board.
And lo and behold - the oscillation went away.

Originally I had been using an MC4558L; I have also swapped this particular
one against another MC4558L with the same result (self oscillation).

Did the swap to the original opamp on the second board as well - same
result, self oscillation gone.

This was IC15; I'll check later or tomorrow what happens if I also swap IC12.

Another update:

After swapping out all new MC4458L's for the old opamps from the damaged
board I did not notice a strong improvement but perhaps a bit.
However, very careful listening to the signal showed that there actually
was a pattern; it wasn't random (at least not fully). Long story short:
It's neither the LFO nor the PWM LFO, it's the arpeggiator!
Right now it appears as if the instability is caused possibly by a
combination of the LFO and the arpeggiator; or more precisely: the two LEDs
(with the arpeggiator clearly being the main culprit).
What I'm not sure of is: Is this the fault of the power supply; or are the
power-hungry LEDs to blame (or are possibly the caps near the LEDs dried
out); or is the whole power distribution problematic? I've removed the
arpeggiator LED for now on one machine and have a much more stable tuning;
the slight pulse is gone.
This behaviour was also verified on a second Polysix with both old board a
clone board.

If you want to try this for yourself here's the setup: After the Polysix is
sufficiently warmed up connect it to a precision tuner. Turn arpeggio speed
way down. Play a single note and use the master tune knob to get a precise
reading. Now turn arpeggio speed to max (so the LED is always on). On my
machine this results in a very slight pitch increase, roughly 1 cent. With
a strobe tuner it is clearly visible since the strobe starts to move; and
it stops again when turning the arpeggio rate down.
I find it easier to hear with a square wave.

I'd be interested to hear if other Polysix owners can reproduce this.
You'll need a very precise tuner for this, though.

Hi Man,

So if the arpeggiator pulses influences the voice tuning, usuall the best path this goes are supply / ground lines.

I recommend you measure voltages on supply and ground. The reference point should be star connected GND of the power supply. That may help you find the true weak spot in your circuit.

Cheers

And one more update:

If a note is held down and the "hold" button pressed the pitch raises also
by roughly 1 cent.

I'm starting to suspect this has something to do wih the fact that the
Polysix has no separate power supplies for digital and analog 5V. Most
instruments under computer control have separate regulators and rails for
digital and analog supplies.

I also wonder if the additional +5V supply line going direct from the power
supply to the LFO LED was an attempt by Korg to deal with a similar problem.

Next update:

Recapped most of the power supply (except the two big ones) - no audible
improvement.

Removed a few LEDs from the programmer board (for example, the "Hold" LED).
Consequently when pressing "Hold" now there is no change in pitch. Touching
the solder pads with the LED while "Hold" is active results again in a very
slight pitch increase.

So I tried to use low-power LEDs. This seems to have very little effect.

Power and ground distribution: Between the power supply and the arpeggiator
board (part of KLM369) I have a voltage drop of about 12mV. Between the
power supply and the connection on KLM369 I have a voltage drop of about
7mV. Which means that on KLM369 alone I have a 5mV voltage drop. This drop
seems to appear along only one of the distribution traces.
It's for me somewhat interesting to see that KLM369 is the only board not
getting a direct power feed from the power supply; KLM369 gets its power
via KLM367.
I've removed the +5V supply from KLM369 completely; this still didn't get
rid of the random "trembling".

I also wonder what the additional ground connection from KLM367 to KLM370
is good for. Attaching it to KLM367 lowers the pitch by roughly 2 cents,
attaching it to the "ground" test point of KLM366 lowers the pitch by
roughly 1 cent. Attaching it to the ground test point on the power supply
has no result, similarly attaching it to the ground connections point of
KLM369 (from which KLM370 theoretically should be getting its ground
anyway).
Continuity from power supply to KLM370 is there, also without the
additional connection. Is it just me, or is that ground connection
theoretically forming a (very small) loop?

Of course all measurements may be off a bit since we're dealing with such
small values here.

> Am 03.07.2016 um 14:40 schrieb Malte Rogacki gacki@gacki.sax.de [PolySix] <PolySix@yahoogroups.com>:
> 
> So I tried to use low-power LEDs. This seems to have very little effect.

Did you just replace the LEDs or did you also change the series resistors?
Just replacing the LEDs and leaving the series resistor as they were won’t change the current consumption.

Ingo

I changed to current limiting resistor from 470R to 2.2k. If I'm not wrong
this should result in less than 1.5mA current drawn by the LED.
I'm under the impression that this modification has improved the
arpeggiator pulse problem but it still hasn't completely fixed it (and the
other instabilities as well). I can still make the strobe tuner "move" by
turning arpeggio speed fully up and make it "stop" again by turning
arpeggio speed way down. But the deviation seems to have gotten smaller.

So far I've tried four Polysixes with five different mainboards (2
original, 3 clones) and all have shown those symptoms to some degree.

So, before I'm starting to chase a phantom: Can anyone else reproduce this
(the very slight pulsing or pitch increase from the arpeggio speed and the
pitch increase from the "Hold" button LED)? Perhaps also slight irregular
pitch deviations when holding a single note?

Shall I record a note with those problems?





> Am 03.07.2016 um 14:40 schrieb Malte Rogacki gacki@gacki.sax.de [PolySix]
><PolySix@yahoogroups.com>:
>
> So I tried to use low-power LEDs. This seems to have very little effect.

Did you just replace the LEDs or did you also change the series resistors?
Just replacing the LEDs and leaving the series resistor as they were won't
change the current consumption.

Ingo




-- 
Malte Rogacki gacki@gacki.sax.de 
----------------------------------------------------------------------------
"Don't forget to TURN ON THE SYNTHESIZER. Often this is the reason why you
 get no sound out of it." (ARP 2600 Owner's Manual)
----------------------------------------------------------------------------