Dear Analog Devices Technical Support Team,
I am currently using the ADA4938 in a fully differential transimpedance amplifier configuration and I am trying to estimate the output noise using a noise-gain based TIA model.
In the ADA4938 datasheet, the small-signal −3 dB bandwidth is specified as approximately 1 GHz for G = +1, with RF = RG = 200 Ω. However, I understand that this is a closed-loop bandwidth under a specific feedback condition, not necessarily the open-loop unity-gain frequency or GBWP.
For my TIA noise calculation, I need to estimate the effective frequency at which the amplifier open-loop gain intersects the TIA noise gain. In a simplified model, I would use a term such as:
fp ≈ fu / noise gain
or, for the TIA case,
fp ≈ fu × Cf / (Cf + Cin)
Could you please clarify what value I should use for fu in this type of ADA4938 TIA noise-gain model?
Specifically:
-
Can the 1 GHz G = +1 closed-loop bandwidth from the datasheet be used directly as fu?
-
Or should it be converted using the noise gain of the datasheet test condition, for example approximately fu ≈ 2 GHz when RF = RG and noise gain ≈ 2?
-
Is there an official or recommended effective unity-loop frequency, open-loop gain crossover frequency, or GBWP value for ADA4938 that should be used for this kind of TIA noise analysis?
My goal is to calculate the differential output-referred RMS noise of the ADA4938 TIA and compare it with LTspice simulation and measurement results.
Thank you for your help.
Best regards,
Chanhee Moon