[sdiy] Re: diodes/LDR

Tue Jan 14 19:49:26 CET 2003

Hello Fabio

The pitch to voltage conversion has some trade-offs.

One: Is the input frequency continuous (like a theremin, or portamento
circuit... ot a motor). These sources cannot change frequency in a
step (jump to new frequency)

Two: What is the lowest frequency you can have.  If a 1Hz signal must
be measured... you cannot get it in less than 1 second timebase...

Three: How fast do you need the data, and to what precision.

One of the biggest problems is having an input frequency that starts
and stops, like a musical note.  In this case you like to count while the signal is valid, then stop (and hold?) when the signal is no good
or missing.

More comments inline...

>Harry, and the list
>Now that you mentioned the PV-1 I want to share some questions about diferents
>methods of F/V convertion. I searched the web for info and get some text about,
>(found less than I expected to find) explaining diferents methods that seem to
>me are variants (simple or very complex sometimes) of a only one that have a
>charging/discharging capacitor as core.
>This implies that the designer have to decide what he wish to get, a fast system
>with an imprecise response or a precise one but too slow to follow small or 
>large fast changes.

This is correct. You need to say what is acceptable performance. My
experience with guitar is that NO system is fast enough.  Probably the fastest method is used by the AXON 'neural net' converter.  They basically use the impulse of the pick hitting a guitar string to guess
the length of the string (therefore the pitch)...then follow this with
a true conversion of the frequency.

>(please note that I arrive to this technic as a complete ignorant and my
>opinions/deductions could be wrong) 
>I also found documentation about mixed analog/digital methods to obtain
>a voltage (or current) with direct dependence of a given frequency, wich could
>be constant or variable, but these implementations have one of their stages based
>in the temporal storage of a charge in a capacitor (or several capacitors) for
>the conversion process and suffer of the same binary issue giving you again the
>election: ripple or a slow response

I guess the rule is... you can't know what will happen until after
is has happened. 

>I have no found concrete information about pure digital methods of f/v conversion
>and this makes me wonder if there is a sort of "trade secret" about such methods
>or I searched in wrong places.

You have to count the period of a half cycle of the input frequency...
minimum. For guitar (and some other instruments) this will not work
because positive and negative half cycles are not equal length (and vary with time also).  Usually the answer is to divide by two and make a square wave.  The PV-1 effectively does this by measuring two consecutive half cycles.

Even then there might be some variation in the frequency that will cause jitter (ripple) in the output....

>Going off topic I remember when was in school (many years ago) our e-teacher
>left to us to decide what project we wished to develop to school science exhibition.
>I proposed the design and construction of a digital frequency counter based in
>the CMOS 4000 series (slow but cheap). My classmates complain alleging complexity
>and/or dificulties but then I assured that the word "digital" and a lot of
>changing numbers in a led display should be enough to ruin them the party to the
>other grades projects. I won ...and got the attention of some girls (ok, it was
>a tech-school, they were tech-girls, and I was so young... )
>The next year we were running out of money then for the next science exhibition
>I suggested that we could modify the last year frequency counter and add it a
>period measuring option. It could be simple to implement, just putting the clock
>frequency in the signal-in and using the "unknown" signal as sampling rate. (We have
>used 1 or 2 Mhz clock divided to get a 0.5s sampling rate, I remember clearly,
>specially the chain of 4017s).  With this method we only must be sure that the
>"unknown" signal should be lower than the clock, and all our investment could be
>reduced to the cost of a switch. (I lost ...we presented a thyristor lamp dimmer)

>Going on topic again I wonder if exist a f/v (or pitch to cv) converter based in
>this principle, (measuring the period of a wave by digital methods and then
>converting the info to an analog signal)

Most Roland units did this... (they output MIDI rather than analog...but if you can do one you can do the other...

>Is there a company producing such converter ? ..is this method patented or under
>trade secret ? ...or was abandoned for something I miss ?
>I think such method should not be hard to implement for musical purpose, it
>could have good accuracy ...and could be fast.

No faster than a half cycle minimum... more likely one to two cycles
in practice...

>Whats your opinion ?

Try to get the highest possible input frequency. One list member strung
his two lowest guitar strings an octave up, to cut the delay in half.

Yamaha made a guitar with all high-E strings to make the detection faster.

Use only a high pitched input...and one that cannot make step changes
and it would work better. Theremin works very well...its really hard
to move your hand fast.

H^) harry

>(I have the PV-1 schem in hands and know your pitch > cv experience is far away
>from mine)
>Any response welcome, respectfully,
>Fabio Gonzalez
>Posadas, Argentina
>harrybissell wrote:
>> 
>> Hi Dan...
>> 
>> I wasn't really attempting to be 'snippy'... just playful.
>> I might even reverse engineer one just for the fun of it.
>> 
>> The points were (as I saw them) that the MoogerFooger is
>> a current product for Big Briar and if they wanted to
>> post schematics it would be their business.
>> 
>> Most of the effect schematics on the web have been there for
>> many years, are often the result OF reverse-engineering and are
>> not scans of manufacturer's documents (unless they are long
>> out-of-business).  Even then, a scan of a schematic used without
>> permission violates copyright laws.
>> 
>> The comment (not mine) to "Ask Bob" was not flippant at all. Bob
>> Moog HAS been known to offer information when asked.  He gave me
>> permission to make the PV-1 pitch to voltage converter based on his
>> design.  All he asked was that the board have the text "Based on Etherwave
>> Pitch to Voltage converter by Bob Moog, used with permission" and that I
>> send him a unit to play with. So I did.
>> 
>> I think that Bob recognised that the product would not be commercially viable...
>> 
>> based on his knowledge of the Theremin market. He was actually doing everyone a
>> favor by giving permission... maybe some folks might even buy a theremin to go
>> with the converter...
>> 
>> The MoogerFoogers are a different story... posting schematics would probably
>> not enhance the sale of $300 units (expensive for you and me both....).  But I
>> think
>> the advice given (in a light-hearted way) is really valid.  Reverse Engineering
>> is a
>> protected practice (its NOT illegal)... and the community would benefit from the
>> 
>> information.
>> 
>> Its also possible that the information is "Trade Secret" (as opposed to
>> patented)
>> in which case there is no way that you will be given a schematic by Big Briar...
>> 
>> as trade secret protection implies that the only way you could have gotten the
>> information is by theft.
>> 
>> Again, sorry if I came off as snippy. Not my intention at all....
>> 
>> H^) harry
>>