Oscope TDR, using the calibration port?

I have used a signal generator and an oscilloscope to make a "cheap" TDR. To see if trace impedance's could be resolved. It was tough. Used a LeCroy 9210? pulse generator with about a 20ps? rise time and I believe a Tektronix TDS 694C oscilloscope, (I'll verify later in the week). I was able to resolve the impedance change with my finger, but not a via or impedance changes of less then 10 ohms. Anyway, I often wondered if it were possible to use the calibration port of the oscilloscope. I'm sure you would only be able to see gross changes in maybe a cable as the rise-time of the cal port is quite slow. I'll try it with the Tek scope and see what happens. Has anyone tried this?

I have not tried this but, it sounds interesting. Can you put together a quick block diagram of how you plan to hook up the system for the measurement.

Do you know what the rise time is of your cal port? You might want to measure it.

So, here is what I found out. The scope I tried first is the Tektronix 2335 100MHz oscilloscope, what I thought was a calibration port is actually the amplitude cal port. I also looked at the TDS 694c. Both amp cal ports are rated at 1kHz. But, the 2335 amplitude is .2v p-p while the 695C is .5v p-p. The slew rates are: rt = 677ns, ft = 314ns. I have simulated it in numerous configurations to get some kind of response with no success.

yes it sounds like a pretty slow response. What you need to do is just use your own oscillator with a fast rising edge. I like the idea of building a tool to use an oscilloscope as a TDR.

What if I used the Cal port as a trigger and power for a circuit that plugs into the probe port instead? I would have to account for how much current it can provide. Hmm. Does this sound familiar or what?

Sorry for the late reply - I just saw this now.

I have done this before and I believe that you can use the calibration port output - but you will need something to speed up the edge - you can buy these things that speed up pulse edges from PSPL. What I've done is to use a power splitter to connect the sampler input, the pulser output and the DUT together. With some specially written algorithms, I've gotten reasonable TDR and impedance vs. time measurements. Last time I did this was using a 2 GHz LeCroy WavePro 960, so it wasn't a super high speed TDR.

Pete