Opamp driving capacitive load

[robzy]

I'm working on a project where I'm using an op-amp to drive a capacitive load, and I'm not certain the best way to do this.

The waveform the op-amp needs to drive consists of very fast transitions, and the capacitance it drives varies significantly – sometimes as low as 10nF and sometimes as high as 1000nF. Typically, the load sits around 500nF.

It's fine if the system is limited to running at 100kHz, which I think should mitigate some of the issues caused by transient waveforms and the capacitive load. But I'm not sure on the right topology to achieve this.

I currently have a simple Microchip opamp running at unity gain. Is there a way to use another op-amp to effectively limit the bandwidth and make the system more stable and usable? Or would I be better off doing some passive filtering with capacitors?

[youngda9]

https://www.ti.com/document-viewer/lit/html/SSZT999

[Conrad Hoffman]

Be sure to figure out the current needed during the transient; you might need a buffer on your amplifier!

[robzy]

https://www.ti.com/document-viewer/lit/html/SSZT999

Thanks for that. I've found a few similar documents online.

It's surprisingly theory heavy. I might have to watch the TI videos that explain bandwidth and bode plots etc.

[edit]: Wait a second, would active compensation of opamps do the trick? I've found this paper from 1979 which seems to imply a solution: https://dflund.se/~getz/Uploads/Synths/ARP_2500/Misc/Active%20Compensation%20of%20Op-Amps%201979.pdf

[Zoli]

TI's Lm8261 is able to drive unlimited capacitive loads; for convenience, I've attached the datasheet.

[ArdWar]

would active compensation of opamps do the trick? I've found this paper from 1979 which seems to imply a solution

I, uh, doubt that actually works, or at least not really worth the hassle. You're more likely to end up with an oscillator with circuit like that

[HalFoster]

Thanks for that. I've found a few similar documents online.

It's surprisingly theory heavy. I might have to watch the TI videos that explain bandwidth and bode plots etc.

[/quote]

Well, when designing op-amp circuits it does help to know how...

[robzy]

TI's Lm8261 is able to drive unlimited capacitive loads; for convenience, I've attached the datasheet.

Wow, that's awesome! Rail-to-rail too, and wide supply voltage. The only thing that gives me pause is "what if they discontinue this opamp?"

I, uh, doubt that actually works, or at least not really worth the hassle. You're more likely to end up with an oscillator with circuit like that

My amateur intuition tells me that something like it ought to work. But the lack of any "prior art" suggests otherwise. This feels like one of those things that starts with "how has no one done it like this before?!" and ends with "oh now I see why no one has done it like this before!"

Well, when designing op-amp circuits it does help to know how...

Heh, that is one way to do it, but the other way to do it is to use a topology that allows you to assume ideal opamps

I've been watching the TI Op Amp Bandwidth videos and they are excellent. I'm going to need to watch them one or two times more, and brush up on my understanding of transfer functions, but I like them.

[tooki]

TI's Lm8261 is able to drive unlimited capacitive loads; for convenience, I've attached the datasheet.

Wow, that's awesome! Rail-to-rail too, and wide supply voltage. The only thing that gives me pause is "what if they discontinue this opamp?"

TI already has a suggested replacement. Look at the product page: https://www.ti.com/product/LM8261

[Avelino Sampaio]

TI has done interesting tests with the OPA192, using low value resistors, with excellent phase margin results.

https://www.ti.com/lit/ug/tidu032c/tidu032c.pdf?ts=1683545366387&ref_url=https%253A%252F%252Fwww.google.com%252F