Oscilloscope calibration of Dynacomp

[miketbass]

Hello, i have recently brought a 1979 MXR dynacomp back to life and have been probing through the circuit to learn more about it. The recommendation for setting the 2k trim pot is to center it and call it good. This is well enough and sounds just fine, but curiosity has the best of me so i have been injecting a test signal into the circuit and viewing the effects of rotating the trimmer with my scope.

My question is, what exactly am I looking for as far as balancing the inputs on the LM3080? I know this is an academic exercise but I would like to know how this chip is set up ideally. I just am not seeing anything noticeably "special" with the trimmer centered as it isn't a larger gain signal, more compressed or less clipped to my eye. I am measuring from the output pin 6 of the LM3080 and am curious about what exactly I am looking for with ideally balanced inputs.
Any insight is, as always, much appreciated and I have not found any online discussion of this topic before.

[Rob Strand]

My question is, what exactly am I looking for as far as balancing the inputs on the LM3080? I know this is an academic exercise but I would like to know how this chip is set up ideally.

IMHO you would want to minimize feedthrough of the control signal.  When the control signal changes it generates a glitch on the output.  This is caused by two mechanisms:  A DC shift at the audio output and pulses caused by various capacitive couplings.

The idea of the balance is when the control signal changes you null out (or reduce) the amount of DC shift at the output.  For this discussion I'm talking about DC levels on the control and on the output.  This all seems OK in theory but in practice things aren't so nice.   Suppose we set the control signal to minimum then force the control signal to maximum (Say by forcing IABC to a large value)  then we adjust the balance so the DC output voltage in the two states  is the *exactly* same; I mean you need uV levels.   Now suppose you repeat the experiment but this time only use half the previous IABC for yout maximum value.  What you often find is the best balance setting here is a different value.  So that means you cannot really get a perfect balance over all of the DC control settings.   You can old try to find a balance setting that keeps the DC shift for alI IABC settings to a "low level".   If you listen to the resulting click caused by feedthrough you will find that the "optimum" setting doesn't sound much different to the midpoint setting.

[miketbass]

I'm trying to digest this Rob, and my knowledge of electronics simply doesnt allow me to make good meaning of your generous explanation. Thank you for typing up your response. I will have to study more about the terminology used to describe the state of a "balanced" OTA.
 
  Hopefully this is helpful to anyone curious about the subject - for myself, being unable to understand the function of an OTA to the extent that I can make much sense of this means I will not be able to tell "optimal" settings even if I was looking right at it!

  I suspect another read through (or 5) of the 3080 datasheet may help form that understanding. Looking at application circuits presented by the OEM is a great way to form understanding of any device and its functions.

[Rob Strand]

You might find these useful:
http://www.nutsvolts.com/uploads/magazine_downloads/11/April%202003%20Ray%20Marston%20-%20Understanding%20and%20Using%20OTA%20Op-Amps.pdf

http://www.nutsvolts.com/uploads/magazine_downloads/11/May%202003%20Ray%20Marston%20-%20Understanding%20And%20Using%20OTA%20OP-Amps.pdf

This one has balancing examples but it is hard to understand if you don't already know the answer,
https://www.intersil.com/content/dam/Intersil/documents/an66/an6668.pdf

I can make some points which might become clearer after you read those documents:
1) Normal opamps have a voltage high gain.   The input is a voltage and the output is a voltage.
  The circuits use feedback and the gain of the circuit is determined by external resistors.
  You usually don't care about the gain of the opamp itself.

2) OTAs have a voltage input but a current output.   In order to make a voltage output the current
  output drives into a resistor "RL" (the 150k you see on the output of the Dynacomp).

3)  The "gain" of the OTA is the transconductance

                 gm = Iout / Vin

4) But with the load resistor you get a *voltage gain* of,

                Av  = Vout / Vin =  gm * RL             ;(comes from Vout = Iout * RL)

5) OTAs are often connected without feedback (called open-loop as opposed to closed-loop).
  The gain is set by the IABC current, which is a DC current.
   When the IABC current is small, gm is small and this makes the voltage gain small.
   When the IABC current is large, gm is large and this makes the voltage gain large.

   The whole reason OTAs are useful and  different to normal opamps is the fact you can control the gain.

6) The input voltage of an OTA is limited and this is why there is a 15k resistor at the input of the Dynacomp.
   It forms a voltage divider with the trimpot to reduce the input voltage to a level which won't distort.

7) If you look at the Dynacomp you will see a 1n cap in parallel with the 150k output resistor
  and a 10n cap in parallel with the 15k input resistors.   These boost the treble on the input
  and cut the treble on the output.  The whole purpose of these parts is to reduce the noise.

We still haven't got to balancing yet.   Ideally when IABC changes the DC output doesn't change, it should only control the gain.  However parts in the real world vary and this creates imbalances in the chip and the result of that is IABC changes the DC output of the OTA.   Imagine no input signal and the IABC current changing between two values.  The DC out will go up and down effectively  "leaking" the IABC signal to the output.  Typically that produces audible clicks which you don't want.  You only want the gain to be controlled by IABC.