My first port of call was to check the amp output, in case some of the abuse it suffered over the weekend had indeed done for the PA devices. So I fired up my R&S CRTU and dialled up 2409.75. Using the provided sampling port on the modified UMTS amp showed a good healthy waveform the went up and down as I varied the drive. So the amp was working then......on to the dish....
Where I first suspected a length of semi-rigid (or is it semi-flexible) that took the RF the last 90cm of its travel to the POTY. It looked a bit corroded, having been outside for 2 years or so. Interesting, that metered OK, so I looked closer at the POTY. It metered open circuit to DC and I could see the SMA centre pin was indeed no longer connected to the patch antenna. At this point I watched the net instead of joining in with video.
This morning I removed the POTY and after careful cleaning, soldered in a new short piece of copper wire to link the SMA pin to the patch. The tip of the pin was removed as it was probably corroded and suffering from arcing damage.
The POTY was replaced on my dish and my PD4 was dragged outside to make use of the excellent 'beacon MER' gauge for alignment. Careful tweaking got the MER up from low 8's to low 9's, so I knew my dish was pointing in the right direction. The 90cm of thin coax was replaced by a 15cm SMA to N, reducing losses further.
Still now show on the satellite???? So NanoVNA was dragged out, the menus were battled with and I was able to measure the system as being in pretty good trim (see the pics). Now what? While I had the NanoVNA running on the desired screen and frequency, I checked the match of my digital VSWR meter and Bird Termaline at this frequency (the meter is not rated for 2.4GHz). As it was OK I tried some power through the meter.....NOTHING??
Now the people at Three Mile Island chose not to believe their instruments, bit lets be honest, R&S make good gear and the NanoVNA, whilst cheap, does a good job. The digital VSWR meter wasn't a piece of junk either....so something HAD to be wrong with the amp. Both Jim and Martin have suffered flash-over and track vaporisation and although I had not smelt any of the magic smoke it is fully cased so maybe that held it in.
Lid off the amp and the problem was revealed - initially I thought the small soldered in link had suffered 'melt-down' but closer inspection showed that it has actually torn off a piece of the tiny PCB trace, presumably from the N-type plug centre pin being rotated in some way.
Repair was executed by using a longer piece of wire soldered to the track surface, hopefully spreading the load around a but. Lid back on and the amp tested fine, I was back on the 'bird
I don't know how long the POTY has been making a poor connection to the feeder - I have not really noticed any change in my system performance. What is does show is how robust the LDMOS devices must be - I know I transmitted into nothing once when out portable last weekend, and during my shack experiments I was running up to 15 amps at 28V into it, also effectively an open circuit. Wow, I'm impressed and more than a bit relived!!
It does perhaps show that the in-amp monitoring circuit is not that helpful (it was probably not just used to monitor power out), you need to know what's going up the cable. But even then, things can go wrong with what's on the end of the cable. Was it thermal cycling (it's in the sun after all), moisture enhanced corrosion (its sheltered, but not watertight) or my soldering (not my weak point...).
At least I didn't emit any radioactivity....
Gareth
- Amp_pin_fail.jpg (241.56 KiB) Viewed 2542 times
- Amp_Pin_repair.jpg (127.55 KiB) Viewed 2542 times
- Nano_VNA_POTY.jpg (83.81 KiB) Viewed 2542 times