Need help with OTA based VCA

[acidblue]

I have the following VCA schem on bread board.
But it's not working as expected, I'm getting a constant 13V from the output, but no sound.
I changed the divider on the input to R1 - 47k and R2 -100R in order to get below 20mV, the default values gave me the same result, thought that might have been the problem, my original input was around 59mV.

I have a slow LFO, roughly swinging between +5 and -4 volts on the CV input.
Using a 2n3904 for the transisor
Everything is running on +15,-15 from same supply.
Voltages:
input from 40106VCO 6.3V
Pin 3 CA3080: 15mV
Pin 6(741) Output steady 13V
3904 collector(3) 0.7mV
3904 Base(2) 640mV
3904 Emitter(1) swings -50 to -75mV

Seems to me that the NPN transistor should be a PNP, but what do I know.
Any insight would be appreciated.

---

OTA based VCA.jpg
Description:
Filesize:	79.88 KB
Viewed:	606 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

[gdavis]

[acidblue]

Thank you good sir for your reply.
I have torn up this circuit and moved on to a LM13700 based VCA which works.

Some time in the foreseeable future I will have to come back to this schem ,just because I have a few CA3080 in my arsenal, and try the things you mentioned.
I'm sure it was something I had wired wrong , but staring at a circuit for an hour trying to find a mistake just wasn't working, time to move on.

[gabbagabi]

a pic from the storage room, it is a pnp there,

i've checked also the formant vca (which also use 3080) there is also a pnp in use

---

1.png
Description:
Filesize:	225.28 KB
Viewed:	475 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

[gdavis]

The circuit can be designed to be either PNP or NPN based, but the transistor type is not interchangeable within a given circuit. Note that the PNP based circuit needs a negative control voltage vs. the positive voltage of the OP circuit. Not sure what (if any) advantage there would be for one over the other.

Actually, I'm struggling to see how the PNP circuit above would work with the negative control voltage...

[gabbagabi]

lol never saw that
i also guess the CV should be positive

anyway some more pnp examples

https://lh3.googleusercontent.com/-n4x88ApBp9A/VcneXrT39KI/AAAAAAAABMg/V_HaZVuPB2s/w623-h538-no/VCA_old.jpg

https://www.elektormagazine.com/assets/upload/img/partial/landscape/3080-vca.PNG

[gdavis]

Ya, probably a typo, it happens.

So ya, either one can be used. PNP does seem to be more common. With NPN, I think the opamp output would have to get to within about 1.4V (or closer?) of V- which might be tough depending on the opamp. With PNP the output range can be controlled with the feedback resistor. Maybe that's why PNP is preferred.

[elektrouwe]

uh, lot of confusion here, both pictures have errors.
The circuit that feeds the OTA is a voltage controlled current source. I'll explain it for the 1st pic. A positive (!) CV creates an input current Iin = CV/R7.
Because the OTA has a control range Iabc =0..1mA, R7 must be around 10k
for a max. CV of +10V. IC3 invertes and pulls down Q1 base. So it must be a PNP (!) with emitter connected to -in of IC3. Because the base current is about 0.5% of the collector current, 99.5% of the CV input current exit Q1 collector and drive the OTA. D1 and R6 protect the OTA in case of negative input voltage, because an OTA can easily be killed by too much Iabc.

[gdavis]

You're right, NPN won't work in the OP circuit, but I'm still trying to wrap my head around exactly why.

The opamp will produce whatever voltage it needs to maintain Vin- at 0 volts. I was thinking this would be negative, yes, but that would be fine since the collector would be at 0V and the emitter would be negative and it would still produce the desired result.

After thinking about it some more, I figure the problem is that the NPN reacts in the opposite direction than we need it to. When the control voltage goes up, the opamp output wants to go down which would let less current through the NPN when we want more. In effect, the NPN connected to Vin- actually creates positive feedback rather than the desired negative feedback (2 negatives = positive). A PNP gives the opposite behavior - more negative opamp output gives more current, so connecting it to the Vin- input of the opamp gives negative feedback.

I tried a spice simulation and was a little surprised by the results. It looks like it finds a stable state which turns out to be putting a positive voltage to the NPN base, forward biasing the base-collector junction as well as the base-emitter juntion. This voltage varies with the control input voltage to maintain a constant 1.5mA emitter current output, which obviously isn't what we want.

Swapping the opamp inputs seems to work, I believe since it now produces the desired negative feedback.

[acidblue]

Just to clarify.
The CA3080 turned out to be fake.
Circuit should work fine, just use a PNP, 2n3906 should be ok.