Combating capacitances (to correct gain issues) in op amp circuits

Hello everyone.

I have built a potentiostat and I am testing it at different frequences (0.1Hz --> 100kHz). As I approach 100kHz+, I encounter gain issues. The gain of the counter/reference circuit actually increases with frequency. I have a feeling this has to do with intrinsic capacitances within the op amps I have chosen (TC7650 chopper type, **broken link removed**) and potentially stray capacitance around the board. What can I do to offset these capacitances? Is this a capacitance issue in the first place?

The schematic is just like this one:

**broken link removed**

The IV converter bit can be ignored as I am not having problems with that part. I have two 10K resistors connecting the reference and counter op amps together. For testing, the counter and reference electrodes are connected together and the voltage is read from that point. Here's some illustrations to demonstrate my point:

At 20kHz everything is behaving - **broken link removed**

Increasing to 100kHz causes an increase in amplitude at the counter/reference junction - **broken link removed**

Which gets much worse at higher frequencies (280kHz) - **broken link removed**

I'm only interested up to about 100kHz. Any help is greatly appreciated.

You didn't report your exact test circuit. I presume you are injecting the AC voltage at the reference node.

I don't understand why you are interested in potentiostat behavior at 100kHz or above. Which real world problem is covered by it?

Feedback loops involving multiple OPs generally require frequency compensation. You possibly get more serious problems when a real electrochemical cell is connected instead of the short between two electrodes.

A RC series circuit between output and inverting input of the upper OP can be dimensioned to provide any amount of frequency compensation the circuit may need.

Thankyou for your reply. The signal is being injected where the 10V supply is shown in the schematic.

Very small capacitances would need to be measured at higher frequencies... however I do not think I will need a potentiostat that has the ability to operate at frequencies of 100kHz+. The design is in early stages and we are not quite sure of the frequencies we are going to need out of this device yet - my theory is making it as wide-band as possible (within reason) at this early stage of design may save a lot of hastle later on in the development. I did actually connect a frequency limiting capacitor from the -in to output of the upper op-amp a few days a go which seemed to fix the problem quite well (it took quite a few attempts with different capacitances to find one that suited to eliminate this kind of gain-distortion). I didn't want to mention this fix in case there was something else I should be doing, but you have confirmed I was in the right ball park. Would adding a resistance improve the design I already have in place (more reliable response perhaps than just using a 'shunting' capacitor)?

As a side, what exactly is causing the gain to increase as frequency increases? Is it to do with propagation delay and phase shift within the feedback loop?

Thanks again for your help.

Dan

Generally it's a problem of feedback loop phase margin. To achieve a predefined loop gain characteristic up to several 100 kHz, your OPs must have a considerable larger gain-bandwidth product, e.g. 10 MHz.