Hi,

I am working on a high-gain transimpedance amplifier that has a closed-loop bandwidth of about 70MHz. I am wondering what would be the overall best way to measure (on the bench) the phase margin in this system. So far I have 2 methods that I've tried in the past (with much lower frequency systems)

1) Apply small square-wave input and estimate phase margin from overshoot
2) Insert small resistor in FB loop, apply AC signal to this via transformer. Sweep input signal
   up in frequency until voltage on one side of resistor is the same as voltae on the other (Av=1)
   Phase difference between input and output signals is phase margin.

Are there any oher methods I am missing? These are common tests for control systems, but this is my first analog design with BW >  3MHz .

Thanks for any help offered.

 -darin

Note that the phase margin will be sensitive to the termination at the input i.e; whether you're driving with a voltage source (low impedance) or a current source (high impedance). You can easily figure out that the current source case is the one that will cause you pain.

Power supply bypass is another point often overlooked by TIA designers. Make sure you model your off-chip caps correctly. Another recommendation is to use a small amount of local feedback on the 'open loop' amplifier - like current series.

Whatever method you use, ensure you run your sims with layout pex (parasitic extraction). TIA's are particularly sensitive to the feedback capacitance Cf.

M.G.Rajan
www.eecalc.com