In Reply to: Re: SHF Pulse generator posted by tom on 02/01/02 at 11:18 AM:
I had a look at my copy of the T.H. O'Dell book over the weekend. The circuit described in there actually uses a varicap diode. A varicap diode exhibits similar behaviour to a step recovery diode but the transition generated is not as fast; however, the circuit described is intended as a demonstration of the operating principles, and for this the varicap was fast enough. The author did mention a step recovery diode available from Hewlett-Packard, part no. 5082-0180. Might be worth checking Agilent's website.
If you can't find a SRD, you'll probably get faster transitions from the avalanche transistor circuits I mentioned than from varacap diodes.
Altogether, it sounds like a very challenging problem. To measure +/-1mm implies a resolution better than 5ps (@ 50ohms)! The best time interval counters I'm aware of (the HP5370B, no longer manufactured, and the Stanford Research SR620) could resolve 20 or 25 ps single shot. The limitation is noise. These machines used averaging to get better time resolution. Another problem will be the smearing of the reflected pulse due to signal dispersion, but I guess that this can be allowed for in the calibration if the medium you are measuring is always the same.