Help: Overheating op-amp in state-variable filter!

[Jacob Brashears]

I am trying to build an audio-frequency state-variable resonant filter.

The schematic is attached. Power supply is ±15V. The circuit contains three TL072 op-amps and three LM13700 OTAs. It receives audio input from -2.5V to +2.5V and outputs a filtered signal depending on its cutoff frequency control voltage. This part works perfectly.

Decreasing the voltage on REZ from 5V to 0V should introduce resonance.

Here's the problem:

When I try to decrease the resonance control voltage (REZ), it works for a while, then stops outputting sound, starts drawing way too much current, and IC13 heats up.


I would sincerely appreciate if someone could help me debug this or offer any suggestions.

Thanks in advance.

 

[FvM]

As a first guess, I would expect that 1C13 overheating is either caused by the IC13B output load, or the load of IC13A.

Did you observe the DC operation point?

 

[Jacob Brashears]

As a first guess, I would expect that 1C13 overheating is either caused by the IC13B output load, or the load of IC13A.

Did you observe the DC operation point?

Which operation point?

In addition, the circuit's full schematic is located here: **broken link removed**
I have removed R91, R105, R113, and R133 from the PCB in order to isolate this filter from the lower filter. To clarify, the lower filter has been isolated from the signal path, and the upper filter is exhibiting the problem.

 

[godfreyl]

The overheating could be caused by instability leading to ultrasonic oscillation. Have you checked the output of IC13 with a scope while it's misbehaving?

 

[FvM]

The overheating could be caused by instability leading to ultrasonic oscillation. Have you checked the output of IC13 with a scope while it's misbehaving?

TL072 won't overhead with high frequency large signal as long as it's not loaded with a large capacitance.

The full schematic reveals a serious circuit problem, IC13B output is clamped by the 4053 analog switch with +/- 5V supply.

 

[Jacob Brashears]

TL072 won't overhead with high frequency large signal as long as it's not loaded with a large capacitance.

The full schematic reveals a serious circuit problem, IC13B output is clamped by the 4053 analog switch with +/- 5V supply.

Can you elaborate on exactly what this means?

 

[FvM]

It's not O.K. to connect the output of a +/- 15V supplied OP to a +/-5V analog switch without at least a current limiting means. Otherwise the OP will drive short circuit current into the CD4053 substrate diodes when its output voltage exceeds +/- 5V.

The other question is why filter output exceeds 5V, but this can easily happen if the filter starts self-oscillation or the conditions for stable DC operation point are violated. I'm sure this can be well explained too, but I primarly wanted to answer your original question.