OTA hiss

[Keppy]

I've got this voltage-controlled filter on the breadboard straight from the LM13600 datasheet. It works, but I have a ton of hiss. I'm trying to figure out if it's a problem with the IC, if these kinds of circuits are just susceptible to picking up hum from outside sources, or if I'm using it incorrectly.



I have made the following mods:
- powered by +/- 9.6v (Boss PSA/LT1054 inverter)
- 10k bias resistor attached to Vc instead of 15k to adjust for the lower supply voltage
- all 200R resistors changed to 220R (just what I had)
- 200pF cap changed to 220pF (same reason)

There is an IC buffer driving the filter. An audio probe at the buffer output is free of hiss. The input and output of the filter are DC blocked with 1uF non-polar caps. Vc is +9.6v, which is running into a 250k pot before hitting the 10k bias resistor.

I have used these OTAs on the breadboard before, and I know they can be sensitive to excess bias current. So what do you think? Do I just have hissy, abused chips, or do I have another problem? I don't have any more LM13600s to try, and I'd like to know if this circuit's going to work for me before I order more parts.

[R.G.]

They can have a lot of hiss. That's what keeps me from using more of them professionally.

The 100K:200 divider at the input gives a huge reduction in signal. I suspect that's intended for line or even synth levels of 1-10 Volts instead of 100mV for guitar. And once it's divided down, it has to be amplified back up, and there is where the hiss gets in.

What you have to do with an OTA without input predistortion (the pin on the LM13600 with the diodes going to it) is keep the signal at the inputs below 25mV for low distortion. Using the diode predistorter can get that up to a couple of volts.

If I were hacking that circuit, I'd want to reduce the 100Ks down to maybe 10K, and then diddle the other parts to get the frequency range back if needed. At the very least, metal film resistors for the 100Ks might help if you don't already have them.

[Keppy]

Thanks for the tip! I actually tried subbing a few resistors and ended up with distortion and oscillation. The trouble is, I don't know how to adjust the circuit to maintain unity(ish) gain when changing one element, especially given the similarity of the voltage divider in the feedback loop to the one at the input. Should I change anything in the loop?

[Keppy]

Never mind, I was making it too complicated. I was trying to redesign the filter, but I just needed to control the signal level in and out and not bother with the actual gain of the filter. Reducing the series input resistor 10x and putting a 1/10 voltage divider at the end sorted it out. The voltage divider at the end even gave me an opportunity for a much better gain control that I was using before. I'd still like to know what controls the gain WITHIN the filter, but I have my circuit taken care of I think.