2
\$\begingroup\$

In order to double output voltage swing I was going to use two OPA548 in difference configuration, connected to opposite inputs like this:

enter image description here

I had one big concern about this circuit - that ground-connected resistor chains on one OP Amp input would affect the gain of the other OP Amp or introduce a bias.

Then I stumbled upon the following circuit in OPA2544 datasheet:

enter image description here

As you can see, it uses second OP Amp with -1 gain connected to the output of the first one.

I think the resistor matching will be much easier to achieve in the second circuit. And there is no cross-affecting resistance on the inputs. However I have a gut feeling that it could be prone to oscillation, and that capacitor looks very much like their way to deal with it.

Question 1: Are there any potential pitfalls in the second circuit?

The isolated amplifier in previous stage is single-supply and has output centered at half positive voltage. For this reason I'd rather connect first OP Amp in difference configuration, to get rid of the DC bias.

Question 2: If I connect first OP Amp in difference configuration, would the mirroring of the output still work as before?

Also, I am going to add adjustable current limit using built-in functionality of OPA548, as discussed here. My understanding is that I only need this on the first OP Amp, the second should fall in line. But since I was wrong about OP Amp behavior many times before I'd like to get a confirmation.

Question 3: Is current limiting on the first OP Amp sufficient?

\$\endgroup\$
  • \$\begingroup\$ Note that this is NOT high-quality audio. THD will be at about .1% to 1%, mostly due to a poor slew rate of 8V/us. High performance audio has a slew rate of 70V/us, but it would cost you 10 times as much. \$\endgroup\$ – Sparky256 Sep 18 '18 at 0:07
1
\$\begingroup\$

Question 1: Are there any potential pitfalls in the second circuit?

Nothing major is obvious based on the available information--at least nothing specific to this configuration. You will need to deal with offset compensation (or at least matching) on both sides of the bridge if you need an accurate zero current, especially if your gain is large, and you will need to apply appropriate compensation if your load is appreciably reactive, but those are important in any op amp circuit.

The isolated amplifier in previous stage is single-supply and has output centered at half positive voltage. For this reason I'd rather connect first OP Amp in difference configuration, to get rid of the DC bias.

Question 2: If I connect first OP Amp in difference configuration, would the mirroring of the output still work as before?

Yes, neither of the inputs to the op amp on the left depend on the op amp on the right.

Question 3: Is current limiting on the first OP Amp sufficient?

You only need to implement current limiting on one op-amp to limit the current through your load in this configuration, as long as the load is connected properly. If you're exposing both op amps to the outside world to interface with a device under test or something, then either op amp could drive an excessive amount of current through a faulty connection to ground, or to either supply rail. Implementing current limiting on only one op-amp would not be able to protect the other op amp in this case. You could even drive excessive current through your load if one side of it is connected to ground our to a supply rail.

\$\endgroup\$
  • \$\begingroup\$ Q3) I am planning on exposing only two OP Amp outputs to the outside load, no ground connection. \$\endgroup\$ – Maple Sep 17 '18 at 23:16
  • \$\begingroup\$ Q1) my thinking was - since amp on the right follows the one on the left it might lag a little bit behind. So, if current limit circuit is active it is possible that rising/falling voltage on the second amp will cause overshot of current, triggering the voltage output reduction on the first. Then second will follow the reduction (again, with some delay) and over-compensating again. Hence oscillation. \$\endgroup\$ – Maple Sep 17 '18 at 23:23
  • \$\begingroup\$ @Maple: gain on the right OP is one. Think about the bandwidth you have then. \$\endgroup\$ – Janka Sep 18 '18 at 0:01
  • 2
    \$\begingroup\$ Your question was about lag. That's a function of the opamp speed, and that's depending on the gain. At unity gain (inverted or not does not matter), you have the whole gain-bandwidth product of the opamp for speed. The left opamp hasn't, it's three times slower. The right opamp cannot ring the left. \$\endgroup\$ – Janka Sep 18 '18 at 5:21
  • 1
    \$\begingroup\$ You have to play with that low pass. It dampens the overshoot on the right opamp and may be required to reduce ringing on that part. It has to be adjusted so the right opamp is still faster than the left. \$\endgroup\$ – Janka Sep 19 '18 at 13:26

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.