The following is a common input and feedback structure in audio power amps:


simulate this circuit – Schematic created using CircuitLab

What I often see is that the resistors are chosen such that \$R1 + R2 = R3\$.


Of course, I understand what that does: equal resistances in the + and - circuits balance the voltages due to the input bias currents (typically in the microampere range for BJT input stages).

But thanks to C2, the amplifier has no gain at DC unity. So the DC offset at the output caused by input offsets is tiny.

For instance, if R1 + R2 = R3 = 39K, the voltage across the resistors is only 39 mV due to a 1 uA bias current.

If only about that much offset appears at the output, it seems insignificant in an amplifier that can swing +/- 50V. The speakers won't dissipate much heat due to this tiny voltage. Why bother with it?

Is there some deeper reason for wanting to null the offset much more precisely, or just it just "neatness counts" esthetics?

  • \$\begingroup\$ What's the gain of the amplifier stage (not at DC)? The voltage difference due to bias current could be multiplied by a large gain, meaning a "small" 39mV could become something significant. \$\endgroup\$ Jul 25, 2013 at 2:02
  • \$\begingroup\$ @helloworld922 The overall gain of the amplifier might be say on the order of 40 to 80 or thereabouts. For example, (R3, R4) = (39K, 1K). \$\endgroup\$
    – Kaz
    Jul 25, 2013 at 2:14
  • \$\begingroup\$ I'll just say that I haven't seen this very often! Do you have an example? \$\endgroup\$
    – Andy aka
    Jul 25, 2013 at 8:19

2 Answers 2


You can't guarantee the input bias currents are less than 1µA, especially in a power amp. Also, loads are sometimes not very tolerant of DC offsets (i.e. speakers).

Matched resistances make the design more robust to variations in the power amp.


Is the thing labeled "STAGES" really an op-amp or equivalent? I don't see any reason for R2, if so. Your supposition about matching resistors to minimize offsets due to input bias currents is correct. Just eliminate R2 and make R1=R3. For audio amplifiers, a typical gain for preamps is ~10 (20dB) and for power amps are typ in the 20-40 range (26-32dB). (80 sounds high) As you point out, the DC gain is 1. So why worry about very small offsets? If this is a power amp, an output offset of 10mV will produce a DC current in the woofer of ~2mA, big whup. A DC servo could be added but why bother.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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