Signal Offset with Capacitor before OpAmp

[LaloFP]

Hi guys!

  I know that a High pass filter will remove DC offset from a signal.

  But, if I only put a 100n cap before a buffer, in LTSpice I get like I set the offset to V/2 (4.5v)... (the same is present in the opamp output)
  It only happens when there is an Opamp after the CAP. Without the opamp It behaves like I expected (same 0v offset as before the cap)



  I dont get it 

[GibsonGM]

"Universal opamp" model limitations?  When I sim the same thing w/TL071 model, I get 900mV of offset at the non inverting input and twice that at the output, which is due to 'how the opamp works'.  Offset voltage, because internal components aren't perfectly balanced.

As the graph shows, your input is a sine wave w/no DC, so the DC is coming 'the other way'...

[ElectricDruid]

The sim is wrong because the circuit is wrong, and the circuit is wrong because the op-amp has no DC operating point set, so it's no surprise that the results are weird.

My guess is that the sim is assuming a sensible DC operating point (half the supply) and using that, since it isn't given any information by the schematic.

Add a 1M to a 4.5V ref to the +ve input like you'd do in reality. Then what you see in the sim will match reality.

[Rob Strand]

I agree with what everyone has said.
The circuit is basically wrong.

"Universal opamp" model limitations? 

I'm not in a habit of using those models.

The reason the biasing works is because the subcircuit for the Universal opamp *adds* resistors 2*Rin from the inputs to the +V rail and to the -V rail and that biases the opamp.   It will basically hide circuit problems exactly like we are are seeing here!

IIRC, there's a bug in those Universal opamps.  They produce outputs which exceed the supply rail.
Firstly the model has no default Rout.  But even if you define an Rout the clipping to the rails doesn't work
since it's not done correctly.  Something weird like that anyway.

The opamp model without the supply rail is fine, at least then you don't expect any clipping.

The non-unversal opamp with the supply rails (opamp2) lets you enter/link your own opamp model, for example a LM741 model you got from the web.

[LaloFP]

Thank you friends! Yeah, I supposed it was a bug and I knew the circuit was "wrong". But I wanted to add the offset and forgot to put the resistors before the simulation. And when I saw the correct result I was intrigued.

I wanted to confirm that it was a simulation/model artifact and not a concept I was missing.

Thank you!