What is the purpose of C75 on the EVAL-AD4080-FMCZ board?

Hi John,

                C75 can indeed be removed or redesigned, the default filtering is pretty aggressive on EVAL-AD4080-FMCZ at the ADA4945-1 Fully Differntial Amplifier (FDA) stage, mainly to show of the AD4080's performance where the SNR does not get dominated by the FDA. 

There are 2 filter poles at the FDA stage to be aware of, the pole in the feedback (C17/C18 82pF with 300 Ohm feedback resistor and C75 with the 56 Ohm, R102, R103) So I guess this brings in some trade offs:

• We increasing the signal BW is needed, this also allows more of the FDA noise through. The wide bandwidth ADA4945-1's noise is exceptionally low for its power consumption, but with its full bandwidth its noise would be dominant in that signal chain.
• Other amplifiers ADA4927 or ADA4930 have lower noise, but this comes at the expense of more power consumption.

If you are redesigning the filter or removing C75 it would still be wise to keep 5 Ohms or so on the output of ADA4945 (i.e R102, R103), just to isolate it from stray track capacitance to the next stage. Wider bandwidth amplifiers tend to tolerate a lot less capacitance direct at the outputs.
If you use LTSpice we have the model for AD4080 on their, if you just drop down the part on a new schematic, right-click on the device, you will find that the example circuit has the ADA4945 + AD4080 simulation set up and bandwidth and noise etc. can be simulated if you are designing a different filter.

        I hope this helped.

          Rgds,

           ds

Removed C75. Looks good. RMS broadband noise is up from ~8 units to ~9.8, not serious. Who needs simulation when you have hardware? Spectrum is flatter, but I still see the other filter pole's effect. That pole is about where I'd like it, so all is well.

Thank you very much.