I'm using the LMH6554 as a transimpedance amplifier. The input current source has a capacitance C_IN in parallel. With this setup a PSPICE AC analysis results in the following bandwidths:
C_IN = 1 pF => BW = 49 MHz
C_IN = 50 pF => BW = 63 MHz
Usually transimpedance amplifier bandwidth decreases with increasing capacitances at input. What is the reason for this behaviour?
This is the frequency response for 1pF input capacitance (magnitude = green, phase = red):
This is the frequency response for 50pF input capacitance (magnitude = green, phase = red):
The 1pF response appears to be dampened, as if using an excessively high feedback capacitance in parallel to the feedback resistance. However, lowering the feedback capacitance has no effect.
Hi Benjamin,
The reason as to why the 1pF response appears to be more damped compared to the 50pF response is because the former exhibits less phase shift (or more phase margin) compared to the latter.
To understand this, you need to plot the loop gain and phase expression for each of the scenarios. The loop gain and phase curves can be plotted for the LMH6554 using the below circuit, where VM3 gives the loop gain expression. As you can see, the phase margin for 1pF is 91.61' indicating an over-damped system compared to the phase margin of 70.93'.
It is important to note that the LMH6554 is a current-feedback topology where the loop gain expression is given by the ratio of open loop transimpedance gain (Zol) and feedback transimpedance (Zfb). For a better understanding of these terms in a current-feedback amplifier, I would recommend you to read this app-note which explains the loop gain expression in detail.
Best Regards,
Rohit
TINA-TI simulation circuit attached. LMH6554_TIA.TSC
