Connecting ThorLabs DET10A to AC-Coupled 50Ohm Digitizer

[Pie566942.0]

Hello,

I'm trying to digitize the signal from a ThorLabs DET10A using an NI 5772 AC FlexRIO Adapter Module.  The DET10A is 50 Ohm and the 5772 is 50 Ohm, but a direct connection between the two does not work.  I expect pulse height N and width M, where N is on the order of 1 V and M is on the order of 10s of nanoseconds.  Yet the digitizer shows pulse height N * 0.1 and pulse width M * 25.  If I turn the DET10A off, the reading from the 5772 doesn't change.  If I turn the DET10A back on, the reading from the 5772 is exactly height N and width M, but the signal decays to N * 0.1 and M * 25 over a second or two.  It's as if the 5772 is sinking too much current or grossly filtering the signal.  I tried connecting the DET10A to a Tek O-Scope.  When the scope's DC-coupled, the signal is exactly height N and width M.  When the scope's AC-coupled with 1M Ohm impedance, the signal is perfect as well.  When the scope's AC-coupled with 50 Ohm impedance, I see nearly the same type of decaying behavior as on the 5772-AC.  We tried 50 Ohm terminating the DET10A and measuring in parallel with the terminator.  Both the scope and the 5772 exhibit the attenuation and ringing I'd expect from mis-matched impedance, but the signal does not decay in this configuration.  The attenuation is somewhat tolerable, but unfortunately the ringing will defeat the algorithm on the FlexRIO FPGA module.  Does anyone have suggestions for digitizing the DET10A with an AC-coupled 50 Ohm terminated digitizer?

Thanks,

Steve K

[David-A]

What do you see when you use a DC coupled version of the 5772?

Also can you post some images of the decaying signal as seen by the 5772?

[Pie566942.0 · Accepted solution]

Thanks David,

An EE here figured out the 5772 AC and the AC-coupled 50 Ohm terminated Tek scope were not providing the DET10 bias a path to ground.  All's well with the measurement now.  Please consider a KB along the lines of, "Using AC-coupled Digitizers with Biased Signal Sources", or something like that.

Regards,

-Steve K