Hi Steve,

I'm fairly confident that I'm not going to spend the (give or take) five digits most manufacturers want for their latest technological achievments. For one thing, I'm shockingly cheap. But for another, I will never have a need for such technology. They're turning out TDR/TDT (Time Domain Reflectometry/Time Domain Transmission) instruments that can spatially resolve impedances within the confines of an integrated circuit. I can not imagine a circumstance where I will ever have need for such resolution.

You're right that I could just "grow my own" with a fast pulse generator. My LeCroy scope is plenty broadband enough to do TDR on a PCB trace, and I could get a pulse generator capable of producing a 70 ps step transition for a few hundred dollars. In FR-4 that will allow me to resolve impedance mismatches of less than 0.2" separation, and that would be more than sufficient for 99% of anything I'll ever do. But then, sometimes you need to put components closer together than 0.2" if you can. So ...

On the other hand, the S-52 pulse generator from Tektronix, part of a TDR system for their old 7000 Series mainframes, produces a 25 ps signal transition which will start to resolve the top layer pad from the bottom layer pad on a 0.065" thick PCB via. And I'm pretty sure I can get such a set up for just a few hundred dollars more than the pulse generator I just mentioned above. Plus, that would be a packaged instrument that I only need to turn on and use (a big plus for someone not looking for a new project).

Of course, I only know about the things I've researched. I've never actually used a TDR/TDT before, and I don't know if there are other, better options out there. For example, I know that some say you can use Fourier analysis to transform time domain (TDR/TDT) measurements into frequency domain (S-parameters) data, and vice versa. And mathematically they're right. But in practical terms the bandwidth of the sampling instrument (e.g. the oscilloscope) will drastically affect how much that theoretical inference represents the reality of your circuit. So would using a VNA (Vector Network Analyzer) to measure S-paramters really be an option? And if so, might it be a better way to go? Unfortunately I've used a VNA once and only once in my lifetime, and that was back in graduate school. It simply isn't a question that I have the experience to answer.

So thanks Steve, and you too Richard, for the input. I really do appreciate it. In the meantime, if anyone else has any experience with such things I'd appreciate hearing from you too. Thanks,

Joe