Op-amp amplification of offset input signal

[the_anomaly]

Hi all,

I'm missing something, probably basic, about using op-amps to amplify offset signals. 



If you are not familiar with LTspice, the op-amp is powered by a 10V source.  The input signal is a 50mV sine wave that is offset by 1V DC.  I have a second source holding the + input at 1VDC as well.  R2/R1 should give 10x multiplication (which it does).  I was expecting the output to center around 1V, going up to 1.5V and down to 0.5V.  Instead, it centers around 0.75V.  What am I missing here?

Thanks!   

** apologies the blue input signal looks very faint.  If you right click and open the image in a new tab, it might show better.  My image got compressed a little on upload. 

[davekni]

What am I missing here?

AD8029 input bias current is causing your puzzle.  For future reference, you can probe currents into circuit elements.  Pretend I didn't give away the answer in first sentence.  Here's what I'd do to debug such an issue:  First probe the two opamp input pins to verify that they are close to the same voltage, 1.00Vdc in your case.  Then probe the currents through R1 and R2.  With an ideal opamp, R1 and R2 currents must match because opamp input current is ideally zero.  When you find that R1 and R2 currents are not equal, then probe opamp - input current.  You'll find it is the difference between R1 and R2 currents, as the sum of all currents into a circuit node must be zero.

[the_anomaly]

Thank you Dave!!  I don't work with op-amps often and I neglected to think about bias current.  I noticed the AD8029 has a relatively large input bias current.  I suppose one option is to pick something with a much lower input bias and I see there are many options on Digikey.   

I also found this resource: https://www.analog.com/media/en/training-seminars/tutorials/MT-038.pdf.  Following the advice in figure 3, I added a 9090 ohm resistor in series with my offset supply (connected to the positive input) and this worked!  Output is nicely centered around 1VDC. 

Thanks again!

[davekni]

I suppose one option is to pick something with a much lower input bias and I see there are many options on Digikey.

Yes, there are many opamp options with infinitesimal bias current (FET input stage).  Especially for rail-to-rail, BJT input stages are rare.  Only very high speed opamps (and some high-power opamps) tend to have BJT input stages these days.

I also found this resource: https://www.analog.com/media/en/training-seminars/tutorials/MT-038.pdf.  Following the advice in figure 3, I added a 9090 ohm resistor in series with my offset supply (connected to the positive input) and this worked!  Output is nicely centered around 1VDC. 

Be careful here.  This was a common solution for older non-rail-to-rail BJT input opamps where bias current was much higher than offset current.  Cancels out bias current, but not offset current.  Newer BJT input opamps such as AD8029 often cancel out nominal bias current internally.  So offset current can be as high as bias current, which is the case I think for this part.  Model happens to have 0 offset current, but real world could end up above or below 1V output.  Better to change opamp, or reduce resistor values so that bias/offset current translates to less voltage.

Thank you Dave!

You're welcome!  Glad to be useful sometimes.