dc offset cancellation

[prcken]

Hi, I am simulating this RX. Had a problem with the feedback loop.



I am wondering why the feedback voltage (output of the opamp) cannot be centered about the output voltage of the TIA, so that the final differential eyediagram cannot be centered around 0V, it has offset. I checked the loop gain has 70dB DC gain and the loop is stable.

I noticed that the inputs of the opamp are equal at the very beginning, i think there is something wrong there. It shouldn't be equal at the beginning, but i couldn't figure out how is that.
any comments?
Thanks

 

[erikl]

Did you try without the OT's influence?

 

[prcken]

Did you try without the OT's influence?

HI erikl, I didn't implement OT (and DAC trimming) so far. I replaced the opamp to a VCVS with a gain of 1000(like an ideal opamp). found that it takes long time (>10us) for the feedback voltage to track at the middle level of the TIA output.
so I think it will finally move to the middle level, just takes time, seems the larger gain, the less time it took. of course the gain also will affect the final error (resolution).
I don't want to design a two stage opamp here to complicate the loop, but has to sacrifice the opamp gain.

 

[Dominik Przyborowski]

I think You should try to move low pass filter from opamp input to its output. The LA will not be loaded by it so should responsing faster, while feedback opamp will set proper dc level.

 

[prcken]

I think You should try to move low pass filter from opamp input to its output. The LA will not be loaded by it so should responsing faster, while feedback opamp will set proper dc level.

In the case you said the opamp is like a high speed comparator.
I think the opamp will go crazy or retarded without the LP filter at the opamp input, the signal swing came out of the the LA is not that small and also at very high-speed.