Hi everyone,
I am learning op-amps. A question that came to mind is which combination of RF, R1, R2, and R3 provides better and cleaner signal in a summing op-amp circuit?
In the attached picture you see that if RF=R1=R2=R3 (not matter what they are), the Vo=-(V1+V2+V3). Can anyone advice me if higher Rs are better (make less noise in the output) or lower ones? and why? Or maybe they really do not matter?
Many thanks,
Adam
There are two considerations: (a) not to small, and (2) not too large.
The op amp output has to supply all of the current through the feedback resistor. That current is Vout/Rf (since the summing junction is at zero volts). This current MUST be less than the total drive current that the OA can supply. Actually, the sum of the feedback current and the load current must be less than the available drive current. This limits the minimum value of Rf.
The maximum value is harder to give a precise rule. If you make it too large, there will be offset (DC voltage shift) due to input bias current. That offset will be no larger than Ibias * Rf (because some of the bias current will flow through the input resistors, also; in the case of 4 equal resistors, the offset voltage will be Ibias * Rf/4). If Rf is too large, there may also be frequency response "issues" arising from stray capacitances, both from the summing junction to ground and across Rf. This MAY result in peaking of the frequency response or a frequency response that is not as high as expected. These effects depend on circuit construction and components used (including the OA, itself).
Noise is usually dominated by the op-amp, itself, and even that is normally only discernible at gains much higher than -1X. In particularly sensitive circuits, such as transconductance amplifiers used for optical receivers, noise is critical and specific resistor values may be recommended. Ditto, amplifiers used for audio. Likewise, for video amplifiers (need unusually high bandwidth), the required resistors may be fairly small.
Generally, I find that resistors in the range of 1K to 100K are a good starting point.
Hope this helps
Jim
Hi Jim,
Thank you for your help. Actually I am using 1.5k resistors and the circuit is working. I can see the Vo signal and it is the same I expect (it is -V1-V2-V3) however the noise in the Vo signal is ~4x the noise in each input signals. I know that this circuit adds the noise from input signals as well (and not only the actual signals) so it makes sense if the output noise is ~4x input noise (am I right?). I just wonder if there is someway to decrease this output noise or not?
I thought maybe changing the resistors will help (it did not though). Any comments on this?
The noise in each input signal is ~2-3 mV while in output is ~6-8 mV.
Thanks again,
Adam
Douglas Self has a book 'small signal audio design' that covers this and related topics. I'd suggest a read of this will answer your question in detail.
Most opamps can hump out 20ma. Some are stronger. Lets say we have +-15V. 15V/.02 is 750 ohms. Dont use a load smaller than that. In your mixer/adder application, make sure the sources can drive the 1.5K input Rs. The output load is in addition to the 1.5K feedback R, so it shouldnt be more than 1.5K either. What opamp? Volts? Output current?
It would help if you can identify this "noise"! Is it really random (eg. real noise)? Or, is it periodic (or nearly so)? If periodic, is it mains frequency or is it related to something else happening in the circuit?
Jim
Douglas Self has a book 'small signal audio design' that covers this and related topics. I'd suggest a read of this will answer your question in detail.
+1 For sure.
You may also wish to consider Johnson noise.
