Instrumentation for high frequency pulse

to be updated...

I have an optical pulsed laser source (has no manual.) Its monitor output has on its certificate for at 3MHz frequency stated as 0.5 ns pulse duration (ON time.)

The output supposed to have 0.1 ns pulse duration also shown in the certificate.

To verify and capture that stated pulse duration, I couple the optical pulse source to detector by a 1 m long BNC coaxial cable and monitor and record the detector output by 5 GHz Scope.

I capture ns voltage pulses from the detector by the Scope as follows :

As you see, the pulse duration is around 3 ns not 0.1 ns as stated by the manufacturer. Rise time is around 700 ps and fall time is around 2 ns. I don't know how they measured it, but in their plot on paper the scope shows mV level pulse output and the pulse duration is really around 0.1 ns.

I read that coax cables can have around 50 to 100 pF/m capacitance. Because such an RC effect usually stretches the falling edge of a pulse, I was suspicious of it but cannot be sure.

In the photodetector manual on page 11 the "Electrical Specifications" state the maximum output is 2V and in the footnote it says "A higher output voltage will decrease the bandwidth." In my case, the pulse is exceeding 2V even though I use the scope with the 50 ohm input settings.

What could be causing this increase in pulse duration from 0.1 to 3 ns for this instrumentation? What could I try to correct it?

• Do you trust that your photodetector and laser source are as fast as it says they are? Because you could always pulse a regular signal down the cable and see if it gets stretched or not to see whether it is the cable or the optical circuitry. Or test a much shorter cable Sep 21 '21 at 13:28
• Carefully read the specification of the monitor output. Is it a 50 ohm output? If yes, did you use a 50 ohm terminated input of the Picoscope? What is the rise time of the input? What is the input capacity? Is the monitor output a push pull circuit with low output resistance for both high and low state?
– Uwe
Sep 21 '21 at 14:50

I believe the waveform you captured was bandwidth limited by your oscilloscope.

The specs for that oscilloscope look like it is capable of 5 billion samples per second (5GS/s) but the analog bandwidth is listed as 500 MHz. The higher frequency components are probably being filtered out by the limits of the analog circuitry within the scope. You can also see that in the photodetector datasheet the rise time is listed as 70 ps but the rise time in your waveform capture is closer to the 0.7 ns rise time listed in the datasheet for the oscilloscope.

The numbers I listed are based on the assumption that you are using the PicoScope 6404C or the 6404D.

• I see makes sense. But what is the point do you think for such high sampling rate when the BW is only 500MHz? Sep 21 '21 at 17:27
• @user1999 At this site is this link to consider, to start. I think higher sampling rates provide sharper triggers (good by itself) and that's also better when applied to a repetitive waveform to mitigate aliasing and wobbling effects.
– jonk
Sep 21 '21 at 18:02
• @user1999 Some scopes are artificially limited in BW so they can sell the high BW models at an inflated price. Keysight, and others, do this in many of their products.
– qrk
Sep 22 '21 at 0:20
• @Luke Sackash: you mean 5 billion ($10^9$) samples per second. A trillion is $10^{12}$.
– Curd
Sep 22 '21 at 7:48