Hi Max,
Your thinking is same as my thinking.
My talk of 100 car batteries was only a half-joke - guaranteed to require everybody go home in a taxi. But series batteries are good for making a fast high voltage and high power source with clean power. The biggest cost that a hobbyist would have is a clean power source to provide 1 amp at 1kv like the commercial SIP testers. In actual practice, 1200 volts from 100 car batteries would not send 100 amps through ordinary ground. Maybe lucky to send 1 amp, and you would use maybe 2 amp fuse to prevent dangerous current from short circuit. So all that is really needed is smaller batteries in series. 80 1000-mah 12v. nicads could do this job for a short test period maybe up to a half hour. My thinking is to experiment with a smaller scale than the commercial units, and the same time you don't need any ripple filters when using a battery.
The only reason they use 1kv and 1 amp is to put many probes in the ground and you need high voltage to push the current for long distances in the ground. If an experimenter works on a smaller scale and lower voltage, then he can still calibrate his test probes and he can get some experience for how to operate the SIP equipment. By putting the probes closer together, he can achieve the same current density in the ground as a high voltage commercial SIP tester. I believe you are right, that it is best to put several probes for the best resolution over an area. My thinking is it is good to start with small inexpensive testing and find maybe 20-30 ounces of gold first before spending money for expensive power sources.
You are correct about the switching circuitry, and this is needed regardless of what kind of power you use. It must be built along with the test probes and any radio telemetry. The problem with using big high voltage power sources is you need to use more expensive semiconductors. I think it is maybe a good idea to experiment with a small test version before building the pic controller and all the electronics for it. Maybe a simple FET switching circuit that turns on power while sending the signal to the sensor, and electronics to measure the phase angle, and frequency. The reason to keep it simple until you have good knowledge of what voltage and current ranges you will be using, and also about the ranges of signal to noise, and how many controls you will want for filtering and adjusting power levels. Seden says these low test frequencies can be from 0.01 to 1 HZ, so I expect switch timing errors should be small when measuring phase angle. This looks like an excellent project to use with imaging electronics like Tim Williams showed us. Maybe a good idea to include a signal output on the electronics for imaging. I think a short time experimenting would bring good information to know how to configure the pic and other electronic controls. But this is only my thinking. For people who have more knowledge of SIP testing, maybe best to start with the pic.
I think the biggest problem will be noise in the low frequencies from atmospheric and man-made interferences, because the sensor is looking for a very weak voltage and current in the ground mixed with other ground currents that could come from hundreds of unknown sources. Maybe it is better to use extra power probes close together for strong ground current density and better resolution. This also means more work to cover an area, and maybe more research to find a location where you expect to find a target.
Best wishes,
J_P