Feedback resistor size - opamps

[dry joint]

The gain is set by a ratio of the 2 resistors +1. Does it matter if you're dealing with large numbers(K) or smaller(00s).

[DUG]

Yes, it does.

Some of the reasons:

The values can be too low for the amplifier to drive (op-amps)

If the values are selected low and there is a capacitor to ground (for AC gain but DC gain =1 for reduced output offset) then the capacitor can become rather large.

If the values are too high the offset voltage due to input currents can increase if not balanced...(Offset drift may still be an issue)

If the values are too high then the input capacitance can cause a phase shift in feedback, causing limits with frequency response, instability or oscillation.

If you look at the datasheets for very high frequency/video amplifiers the feedback resistors are low compared to audio applications...The voltage swing is also small...1-2V

I think there are also noise issues with higher resistance...Someone more experienced should address this aspect of the question.

Hope that helps.

[DF96]

Resistors generate thermal/Johnson noise, which in voltage terms is proportional to the square root of the resistance. Bigger value resistor means more thermal noise. However, there may be little point in reducing noise much below the noise already coming from the source.

[dirkwright]

Originally Posted by DF96

Resistors generate thermal/Johnson noise, which in voltage terms is proportional to the square root of the resistance. Bigger value resistor means more thermal noise. However, there may be little point in reducing noise much below the noise already coming from the source.

Yes, the noise difference between 10k feedback resistors and 1k resistors is dramatic, at least in simulation. I prefer to use the lowest I can get away with.

What I do is figure out the expected load for the opamp, and then put that in parallel with any feedback resistor values, and the result has to be greater than the minimum recommended load for the opamp. In order to provide greater drive power, I add a buffer inside the feedback loop, like the BUF634. The over all result is lower noise.

[dry joint]

In an opamp-buffer-headphones setup, is the load to the opamp largely the resistor between the opamp and buffer?

[dirkwright]

Originally Posted by dry joint

In an opamp-buffer-headphones setup, is the load to the opamp largely the resistor between the opamp and buffer?

I don't put a resistor between them. But, if there was one, then that'd be it.

[jcx]

if is in in series with the (non-inverting) buffer input then the load is the buffer input Z, typically MegOhm || pF

a "Class A bias" R from the op amp output to one of the PS rails does count as load

[dirkwright]

Originally Posted by jcx

if is in in series with the (non-inverting) buffer input then the load is the buffer input Z, typically MegOhm || pF

a "Class A bias" R from the op amp output to one of the PS rails does count as load

Yes, of course, should have clarified, sorry!