Unity-gain buffer and capacitive load - which OpAmps handle it well?

Tomasz Jastrzebski

Other Parts Discussed in Thread: OPA350, OPA392, TINA-TI

The design I work on requires a simple G=1 DAC (3V max) buffer/voltage follower.
The challenge is the buffer needs to be able to drive capacitive load, ideally up to 0.5 µF (could be much lower) so stability is high concern.
Simple solution with Riso is not a preferred one - increases Zo which must be kept low, introduces voltage drop. Solution using dual feedback loop is not preferred either - it has to be optimized for constant, know capacitance.
The best options I found so far:

LM8261 - unlimited capacitive load but relatively high V_OS 0.7mV(typ), 7mV(max)
OPA202 - 25nF C_L(max), V_OS 0.2mV(max), high output swing from rail ~1V - specified for V_S=36V only, for V_S=5V probably higher (1.5V according to TINA model)

Are there any other known choices I should consider, besides adding additional components? - e.g. like shown below

source: APPLICATION NOTE 5597 by Mark Fortunato

over 3 years ago

0 Clemens Ladisch over 3 years ago

Guru 300430 points

One of the feature checkboxes in the search function is "High Cload Drive":
ti.com/amplifier-circuit/op-amps/products.html#p480=1;1&p2192=High Cload Drive

+1 Marek Lis over 3 years ago

TI__Guru 64870 points

One op amp capable of driving directly in a buffer configuration micro-farad loads is OPA350 - see below.

Also, please see link below:

e2e.ti.com/.../opa350-capacitive-loading-with-intent-to-filter

0 Tomasz Jastrzebski over 3 years ago in reply to Marek Lis

Prodigy 212 points

Thanks Marek, OPA350 would be an excellent choice. The only disadvantage: no popular SOT-23/SC-70 packages available.

0 kai klaas69 over 3 years ago

Guru 140200 points

Hi Tomasz,

Tomasz Jastrzebski said:
Solution using dual feedback loop is not preferred either - it has to be optimized for constant, know capacitance.

This is less problematic than it sounds and might only be an issue, if the variance of load capacitance is huge. But if you add a dummy load capacitance at the ouput of buffer which is bigger than your varying load capacitance, the dual feedback can work very well. The very most buffers work this way.

Kai

0 Marek Lis over 3 years ago in reply to Tomasz Jastrzebski

TI__Guru 64870 points

Tomasz,

If you absolutely need the SOT-23-5 or SC-70-5 package, our new OPA392 may be your best choice - see below.

By optimizing the dual feedback compensation for a dummy 1uF cap, as Kai suggested, you will assure stability for any external capacitance, Cexternal, from zero to 1uF - see below.

Cexternal = 0

Transient small-signal simulation confirms stable operation - see below

Cexternal = 1nF

Cexternal = 10nF

Cexternal = 100nF

Cexternal = 1uF

Below I have attached Tina-TI schematics for your convenience.

Tomasz OPA392 AC Stability.TSC

Tomasz OPA392 Transient Stability.TSC

0 Tomasz Jastrzebski over 3 years ago in reply to Marek Lis

Prodigy 212 points

Marek and Kai, good point, I have not played with dual feedback model before. After all, adding just three components may make more sense at least in one case - a test/fun project I have just published on GitHub (All-purpose Power Micro Controller). This approach would allow for greater flexibility - a factor not to be underestimated at the time when components' availability is limited.
In another project I currently work on, OPA350 will be the way to go since C_L is known.
Thank you for your support.