Simple but good dual vca

I found this schematic on Niklas Ronnberg site
i’ve just add 2 indicator led
the + : in pot but also initial level pot !


I’ve never seen a layout quite like this-are the 1k resistors on the outs and the cap on the in indicating to solder them off-board in line to the hot pins?

1 Like

Is this just using 2 Op amp chips? not seeing anything out of the ordinary. might give this guy a build. :slight_smile:

CA3080 is an old OTA that hasn’t been manufactured in ages. You can use half a 13700 (it’s literally a dual 3080) but you’d have to tweak the layout. No idea where to source 3080s.

AliExpress seems to have plenty of them. Pretty cheap too if you want to risk it.
Even some surface mount ones ?!?

The circuit is even stranger than that, the negative power supply for the CA3080 OTA seems to be coming from the output of the TL074 opamp!

1 Like

Not sure I’d trust the recycled ones, but turns out you can still get the real ones from an end-of-life supplier: (they have a stash of wafers and package them on demand), and you can get their DIPs in small volumes from Thonk: (well, at least if you’re in a location they ship to at the moment).

But since this is a dual VCA anyway, redesigning for a 13700 makes a lot of sense. That negative supply is weird, though.

Also found this:


i bougnt mine on Ebay

I built one and I am currently doing another, that’s why I made a topic: for me a good module and very easy to do


Seriously? They’ll build you an end-of-life IC on demand? IS AMERICA GREAT OR WHAT don’t answer that


That’s a very interesting, albeit badly written, article.
You do have a knack for finding them, thanks.
It also point to an old article on OTAs in Radio Electronics, the magazine where I learned electronics when I was a teenager.


I suspect you have to order more than a handful, though, but you could always give them a call :crazy_face: Looks like they have enough wafers for ~400,000 ICs.


Can one use 1/2 LM3900 as a replacement for an LM13700 and v.v. or are they fundamentally different in their working?

1 Like

The LM3900 is a quad Norton op amp, not an OTA. They’re very different.

1 Like

Well, I knew the names of the thingies and those do not explain a lot.
What is it that makes them different?

1 Like

As I understand it (very minimally) an OTA takes a differential voltage and outputs a current; a Norton amplifier takes a differential current and outputs a voltage.


Which sounds like they are 1 resistor apart. @All: Maybe someone with some expert knowledge could comment on this?

1 Like

Well to stay simple CFA/Norton amplifiers/transimpedance amplifiers are like normal opamp but faster because they directly work with current. You can even find some fast normal opamps that have a CFA inside, with the correct input stage wrapped around.
The main advantage of transconductance amplifiers (OTA), is that you can bias the current output stage with an external current, hence easily creating a VCA or even a full multiplier.

I never heard of current operational amplifier (COA), though… But they seems to exist.

Edit: I just realized that actually COA are basic building blocks for normal opamp (VOA), with the corrects input and output stages added on it.


@Dud le Bricoleur built another copy of this, which reminded me of this thread and that strange power supply connection, and after skimming through some 70s opamp literature I think I’ve figured it out.

Here’s the Iabc input on the CA3080:

The Iabc input (pin 5) controls the current through the differential input pair that handles the input signals (pins 2 & 3), via a transistor connected to the negative supply rail (pin 4). Note that this means that Iabc is held one diode drop above the negative rail, so if you feed it through a passive resistor you have to dimension things for CV + 10.6 V 11.3 V to get the right current (for a ±12 V supply, that is, adjust accordingly for other supplies).

Most designs work around this issue by using an active current source to drive Iabc; either a transistor configured as a common base amplifier (see e.g. the Henry BD++) or better an opamp + transistor (see e.g. Electric Druid’s vintage VCA). This lets you ignore that 11.3 V difference and think only in terms of current, but it adds some external components, and if you’re sneaky you can save one whole transistor by using one that’s already there:

By putting Iabc in the feedback loop, the incoming Icv current (CV / the input resistor) has nowhere to go other than through pin 5 (Iabc), and for that to happen the opamp needs to pull pin 4 down roughly two diode drops (D2 + internal circuitry).

Of course, the output stage uses the same supply rail but since the output is current-driven that’s not a problem as long as you drive into ground or higher (e.g. the current to voltage converter with a virtual ground used here).

You can use the same trick with the LM13700 but it has two diode drops to the rail so you need to add one more diode in series with D2 to have some margin, and of course both halves use the same supply, so you’re probably better off using an external transistor so you can use the second amplifier for something else…

(if anyone wonders about the D1 in the original schematics, it’s there to protect against reverse CV polarity).


Something like this should work:

The transistor can be any small signal PNP, e.g. 2N3906, BC556 through 560, etc. The 10k resistor is optional but will make sure the Iabc current stays well below the 2 mA max rating no matter what the CV circuity does.

(with Tayda prices, this change will make each VCA around 2 cents more expensive)

1 Like

(Now you’ve tricked me into ordering a bunch of 3080s off eBay, from a few usually reliable suppliers I’ve used before. Will report back if they seem legit, or if they’re repainted 358s or something :slight_smile: