OTAs: What, why and what for?

[iainpunk]

I sure am willing to try a OTA distortion/overdrive But I'd feel a bit violated - those 13700s sure don't come as cheap as your average op-amp.

in that case, go discrete! most OTA's don't have that much transistors anyways, or get yourself a LTP chip like the CA3028AM, which is actually a differential amplefier, but with some added circuitry can be (ab)used as an OTA, and they are like 80ct in my local shop, i haven't checked internet prices yet.

they can also be used as a VCA, there is a chematic in the aplication note (AN5337, figure 2C)

cheers

[aion]

I just have to say, this thread is an example of what makes this forum one-of-a-kind. Someone asks a basic question to which the response might have been some flavor of "go read X book or Y article and learn it yourself, the info is out there" - but instead, several seasoned travelers take the time to share their knowledge and write out thoughtful and thorough responses that cover theory, application and history without a trace of condescension.

And there are 500 other threads just like this one about different concepts. You guys have no idea how much I've learned throughout the past ten years just by typing "site:diystompboxes.com [search term]" into Google when trying to understand something.

[garcho]

The old Maplin R.A. Penfold books have a bunch of OTA circuits. There's an overdrive and tremolo that both use the OTA as something like a VCA. Very musical effects, I've been meaning to revisit the trem; it kicks in as your input decreases.
They can be a bit noisy and it's easy to fry Iabc. Otherwise they're DIY solid gold!

[idiot savant]

Anyone looking for a deeper dive on OTAs should check out Ray Martsen's nice articles from Nuts & Volts.

https://www.nutsvolts.com/uploads/magazine_downloads/11/April%202003%20Ray%20Marston%20-%20Understanding%20and%20Using%20OTA%20Op-Amps.pdf

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

edit:

Looks like these were posted already. Electronotes also has some good OTA info if you search it out...

[bluebunny]

The E&MM Valve Driver (December 1983) uses an OTA.  I have one half-built in my backlog.

[KarenColumbo]

The E&MM Valve Driver (December 1983) uses an OTA.  I have one half-built in my backlog.

That's one interesting circuit!

[Frank_NH]

Speaking of the prolific R.A. Penfold, here is a link to book of his where, on p. 52, he describes an OTA-based limiter circuit using the LM13700 chip.  I'm going to try this out on my breadboard as it looks like it has some potential.

https://worldradiohistory.com/UK/Bernards-And-Babani/Babani/309-Penfold-Preamplifier-and-filter-circuits.pdf

[Frank_NH]

The E&MM Valve Driver (December 1983) uses an OTA.  I have one half-built in my backlog.

Wow - yet another one for my breadboard.  Thanks!    You could probably use a +/- 9V supply from a charge pump to power the circuit.

[bluebunny]

You could probably use a +/- 9V supply from a charge pump to power the circuit.

It's just an ordinary 9V circuit.  But they've confused us a little by labelling what we know as V_REF with an earth symbol, and calling the negative end of the power supply "-V".

[Frank_NH]

You could probably use a +/- 9V supply from a charge pump to power the circuit.

It's just an ordinary 9V circuit.  But they've confused us a little by labelling what we know as V_REF with an earth symbol, and calling the negative end of the power supply "-V".

I noticed that but thought a +/- 9V split supply would provide some increased headroom.  Plus, I have a charge pump supply available for my breadboard.

[KarenColumbo]

Speaking of the prolific R.A. Penfold, here is a link to book of his where, on p. 52, he describes an OTA-based limiter circuit using the LM13700 chip.  I'm going to try this out on my breadboard as it looks like it has some potential.

https://worldradiohistory.com/UK/Bernards-And-Babani/Babani/309-Penfold-Preamplifier-and-filter-circuits.pdf

Whoa, cool!

[KarenColumbo]

I posted this in the "Causality 4 Phaser" thread, too (at the very end), but I think, it's also well placed here:

I see that probably the biggest difference between Causality and Tri-Vibe is that the former doesn't utilize the LM13700's "Diode Bias" pins (2 and 15). "Instead" there's 1k8 resistance 'tween the internal diff amps positive and negative inputs. If I now go and connect the diode pins to V+ with some resistance in series AND rip out the resistor between pos. and neg. op-amp inputs and put them to Vref via small resistors (470R @ Tri-Vibe), what will happen from a theoretical point of view? Since the LFO's output connects to pins 1 and 16, which are "Bias Inputs", AND the pos. and neg. amp's input are connected to V+, Diode Bias connected to Vref halfway between V+ and GND, what does that mean for effectivity? (I feel like I'm just about somehow forming a mental pictureof how OTAs work)

[anotherjim]

Here's the Wasp synth OTA filter...


This runs single supply.
There are 33k 1/2 supply dividers at one input (the designer really liked 33k!).
The other input comes from an inverter amp which will bias to 1/2 supply so same (nearly) DC on both inputs. A 1k between the inputs. Signal enters via 51k, so attenuates by the 1k. The inputs act on the difference voltage dropped over the 1k when signal happens.
These are inputting to the inverting inputs because of the output current-voltage buffer being an inverter (an integrator too). The inverter feedback cap is your filter cap. Overall, each filter section is non-inverting.
Any voltage feedback is going back in via the input inverter amp.

[anotherjim]

Oh, there's one more wrinkle in the Wasp OTAs. That 33k divider has no AC bypass cap on it, so those resistors will factor in the amount of attenuation and both inputs will have some signal. With an inverting opamp based amplifier that's behaving properly, you should not get any signal on the -input as feedback should be holding it at the DC Vref on the +input.

[ElectricDruid]

I posted this in the "Causality 4 Phaser" thread, too (at the very end), but I think, it's also well placed here:

I see that probably the biggest difference between Causality and Tri-Vibe is that the former doesn't utilize the LM13700's "Diode Bias" pins (2 and 15). "Instead" there's 1k8 resistance 'tween the internal diff amps positive and negative inputs. If I now go and connect the diode pins to V+ with some resistance in series AND rip out the resistor between pos. and neg. op-amp inputs and put them to Vref via small resistors (470R @ Tri-Vibe), what will happen from a theoretical point of view? Since the LFO's output connects to pins 1 and 16, which are "Bias Inputs", AND the pos. and neg. amp's input are connected to V+, Diode Bias connected to Vref halfway between V+ and GND, what does that mean for effectivity? (I feel like I'm just about somehow forming a mental pictureof how OTAs work)

There's two different things going on here. One is the bias voltage for the two inputs. The other is the linearisation diodes. They're not related.

The Causality 4 connects the positive input of the OTA directly to the Vref. The negative input also needs to be biased to the same voltage, but in the the case of the input, we have a voltage divider (required to lower the incoming signal level) so we have a 27K/1K8 divider down to Vref. Since +ve is already connected to Vref, that's a convenient source of Vref to use. It's just the way it's drawn - the +ve and -ve inputs aren't connected "together", they're just connected to the same bias voltage. It helps reduce output offsets (if that's important) to use the same resistance value on both inputs (so the +ve input could be connected to Vref via 1K8 too) but in this circuit, it doesn't matter.

The diode bias is just a resistor to the positive supply to turn the linearisation diodes on, if you're using them. Leave it out if you're not. It's independent of how you do the input biasing.

[ElectricDruid]

Ignoring the missing cap on the reference supply that Jim mentioned, this filter shows the same pattern.

Notice the +ve input is tied directly to the Vref (here made individually for each OTA with 33Ks, rather than once for everything) and the -ve input has a voltage divider (51K/1K) to reduce the input level. Again, the voltage divider is tied to the Vref. Again, it looks like the two inputs are tied together, and in this case because of the missing cap, they actually are to some extent!

Edit: And there's no linearisation diodes on the CA3080, so that question doesn't arise!

[KarenColumbo]

Okay, I need to readjust my thinking. Thanks for the support, guys I'll get there eventually

[noisette]

On a sidenote, put the 13700 in the feedback path of a sallen key/cascade filter (like that causality phaser, not the wasp though, it is state variable and works different) you get wonderful smooth sinewave self oscillation***, voltage controlled.
A very well known filter doing it is MI ripples.

***That has mainly to do with the differential input pair (look up tanh distortion).
Check Arp 4072 schematic for the same effect in discrete version!

[Frank_NH]

More food for thought...

The Penfold Fuzz:

http://moosapotamus.net/ideas/penfold-fuzz-unit/

[anotherjim]

More food for thought...

The Penfold Fuzz:

http://moosapotamus.net/ideas/penfold-fuzz-unit/

I take it that by "gain stage" the author of that means "side chain". The envelope follower normally won't be intended to supply any audible signal except for any signal ripple that C6 fails to remove. Some kind of ring mod character might be present due to the ripple and deliberately making C6 too small could sound interesting.