As I was reading through your symptoms, I was thinking:
... diseqc switch, diseqc switch, coax connector...

It wasn't clear where your diseqc was, but if it's out at the dishes, then I still like it.
But "bad connector on the coax" really works for me, too.
That "no current" is a give away.
The 5 volts is probably just meaningless leakage either through the switch or a bad solder joint, or a crimp.
Without seeing a schematic of how you're set up, I wouldn't want to offer a serious opinion,...
... but the switch or connector sure seem like glaring candidates to me. -

Let us know! - :up
Well, there was no switch in the ciruit at all. LNBs connected straight to my receiver, except it did go through a power pass splitter which I bypassed when I started testing. However my testing has really hit a confusing symptom, which I'll talk about below....
I have wasted time with the voltage test here. The trick is testing with current draw also. Otherwise one strand of conductor might be passing the voltage under no load conditions with a digital high impedance meter.
I agree with this completely, and I generally test voltage with my channel master meter, which measures voltage and current while powering the LNBs, so it was under load. Saturday I got 18V. Today, even before I could connect the LNB to the meter, the meter was saying only 5V.
Well, I tried to determine whether the center conductor or the shield was faulty, so I did an experiment. And the results are a bit confusing.
I shorted the center to the shield of the "faulty" coax back at the receiver, and also connected both to the shield of the "good" coax. I then checked the resistance out at the dish.
I got ~ 1500 ohms between the center and shield of the bad coax.
I got ~ 300 ohms between center of the bad coax and shield of GOOD coax.
I got ~300 ohms between shield of bad coax and shield of GOOD coax.
I then disconnected the good coax, shorted shield and center at receiver end. Checked the resistance, and it was about 11 ohms.
The 11 ohms sounds logical for a good coax, and the 340 ohms makes sense for either the center or the shield for a bad coax, but I can't believe that BOTH would be 340, and I was expecting the 1500 to be twice 300, but instead it was more. This sort of suggested that it's a semiconductor type conductivity of some sort, ie different dependent on the direction of the current, so I reversed the polarity of the tests that gave 300 ohms, and sure enough, one gave something close to 600, but the other polarity gave something like 1500 ohms, PLUS, you could see a capacitive charging/discharging occurring, ie the resistance changed over a period of about 10 seconds while it built up to the steady value. So I'm thinking that I broke the conductor somewhere, but I'm getting conductivity through corrosion or some dielectric or something which is polarity dependent. Anyway, it's pretty clear that the coax is bad. I tried pulling out the center conductor, in case it was broken near the end, but it was solid. Also cut off a couple inches of coax at the dish end, and still got the same readings.
So the coax is clearly bad. I think it's only hope is for me to cut it in the middle, hoping to find the place where I damaged it, but I think that will have to be put off while I'm fixing other things.