In the circuit below I am trying to achieve 15.56 gain on IC2.

enter image description here

The following combinations (in kOhms) are fairly close:

20/1.3 or 200/13 = 15.38
47/3 or 470/30 = 15.67
51/3.3 or 510/33 = 15.45
56/3.6 or 560/36 = 15.56

The pair 56k/3.6k is actually spot-on and seems to be OK for OPA548 and required bandwidth (under 200kHz) per this answer. Although it is rather on low side, just as 560k/36k seems a bit too high.

Now, to consider SI8920B requirements I have to look at input impedance of difference amplifier, or maybe at all three impedances according to this answer.

The problem is that I don't understand which one is specified in the datasheet. It says "Output Resistive Load" minimum is 5k. If this is differential impedance then all combinations but 20/1.3 are OK. However an application example in the datasheet has 5k resistor on both lines. If this means datasheet specification is for individual outputs then I am left with high resistance combinations only.

So, my gut feeling tells me to go with 200k/13k.

Question: is this a right choice?

  • \$\begingroup\$ Thanks! I take it, you interpret datasheet as differential impedance then. I've found some suitable 0.01% arrays with 0.005% matching, although quite expensive. Will probably go with discrete to save money. \$\endgroup\$ – Maple Sep 28 '18 at 20:37
  • \$\begingroup\$ I posted my comment as an answer, per EE.SE rules. \$\endgroup\$ – user105652 Sep 28 '18 at 20:40

56K/3.6K 0.1% are valid values. Avoid trim pots as they have a high thermal drift and can affect common mode DC errors.

For several $USD each you can get 0.025% to 0.001% tolerance resistors that are ultra-stable, made by Caddock. Matched pairs can be had for lots of money. It's your call.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.