Well, it was a pretty long, hot afternoon. I pulled into the backyard under a mighty oak, set a fan, grabbed a buncha instruments, and got after it. I am really thankful that @jeepgrandcherokeesal asked for additional measurements on the CDMA antenna, because the 50-ohm termination Did Not Work
. Really stunned, because I was expecting a nominal comm. antenna feed. But armed with that extra info, I was able to fend-off the nag screen as well as keep the radio from registering with Sprint ~99% of the time.
I have pictures to sort and post, so it will take a bit for me to organize, but here's the short answer: Terminate with 10k, and cage it.
First thing I found was the radio is putting 2.5vdc (center positive) on its antenna connector. I attached the 50-ohm load and was crushed to quickly get the nag screen and watch the radio register with the carrier. I was able to record some RF sweeps and got predictable activity to use later. Also observed the load for temperature change, just in case there might be an antenna current issue. I think not, although the load temp did rise about 4 degrees above ambient but stayed well below what the rest of the radio was dissipating thru the case.
So I turned to the antenna and measured:
(with TDR) Termination is 17 feet away, and higher
impedance than the 50-ohm cables.
(with LCR bridge) L: 55mH C: 670pF R: 10k <--- ….Oh, Really?
Sooooo, it was back to the bench, replace the 50-ohm load with 10k, and the next trial went much better. There was no nag screen, but the radio still registered. The RF measurements were way lower but the load was radiating.
Sooooo, it was back to the bench, solder a piece of tubing over the load,
and the results were ideal: No nag; No registration; No RF activity
I left it all running for an hour with no change, and then closed it up.
I drove around later and did notice the radio indicate service for a period when I was within the shadow of a known Sprint tower, but I'm also thinking I need to close the end of the shield over the load - the tubing is open-ended. Just maybe?
I'll gather up the rest of the show and put it up soon, but I think we're very, very close.
This is the current state of the device: