Curious about a failure

MMDVM_HS Hat hardware
Post Reply
kd4iz
Posts: 8
Joined: Fri Apr 27, 2018 3:16 am

Curious about a failure

Post by kd4iz »

Of several hotspots, I have one mmdvm_hs hat - acquired second hand. Worked fine for quite a while (year or so). Pwr was always from a UPS and the unit was usually on unless exercising lightning precautions and I shut the entire station down.

Fail occurred last week - it went deaf and dumb and the OLED display went blank. The Pi-Star dashboard stopped showing any info in the Radio Info Trx block. Did not have a schematic or measurement parms list... but attempted several things:
1. Pi-Star Update - performed as expected but nada
2. Pi-Star Upgrade - performed as expected but nada
3. Firmware Update - performed as expected but nada - messages in the SSH window did not suggest a problem to me.
4. Started with a new image on new SD - never could get it to settle down during configuration - kept forgetting modem type and losing the WiFi configuration...
5. Just for S&G, swapped in another known Pi Zero WH board and used working SD card - same results...
6. Put the original configured SD back in and it continues to "look like" everything "else" is OK.

*Have not taken the logic probe to it* -too much trouble and it is not my primary hotspot, so really would do this out of curiosity.
Even though sequestering and pretty bored, I gave up playing at that point and ordered another US$60 hat from our friends at Amazon.
So if any hardware tech wizard(s) would like to help educate me a little here - where is the most likely point(s) of failure to probe on the hat? It might be an interesting exploration.

Thanks in Advance - KD4IZ
User avatar
G8SEZ
Posts: 553
Joined: Fri Apr 13, 2018 8:26 pm

Re: Curious about a failure

Post by G8SEZ »

If the TCXO for the ADF7021 clock input fails then the symptoms you describe are possible, you can flash the board because the STM32 clock is running but the STM32 cannot talk to the RF chip.

Sometimes cheap TCXOs fail, sometimes they have defective soldering because the PCB footprint is not right for the package pads.

I repaired one of my Hats that never initialised properly, it was cheaper than trying to get a refund on eBay.
--

Brian G8SEZ
User avatar
G8SEZ
Posts: 553
Joined: Fri Apr 13, 2018 8:26 pm

Re: Curious about a failure

Post by G8SEZ »

KE7FNS wrote: Thu Jun 11, 2020 9:30 pm
G8SEZ wrote: Thu Jun 11, 2020 8:47 pm If the TCXO for the ADF7021 clock input fails then the symptoms you describe are possible, you can flash the board because the STM32 clock is running but the STM32 cannot talk to the RF chip.
If that were the case the "running" STM32 would still be communicating with the OLED and it wouldn't be blank.
Maybe, but if you also have the logo screensaver set to 0 due to fairly recent changes then I think it's possible. I was convinced by the lack of Rx and Tx too.
--

Brian G8SEZ
User avatar
G8SEZ
Posts: 553
Joined: Fri Apr 13, 2018 8:26 pm

Re: Curious about a failure

Post by G8SEZ »

KE7FNS wrote: Fri Jun 12, 2020 7:12 pm
G8SEZ wrote: Fri Jun 12, 2020 1:54 pm Maybe, but if you also have the logo screensaver set to 0 due to fairly recent changes then I think it's possible. I was convinced by the lack of Rx and Tx too.
The LogoScreensaver configuration setting only affects the "mode" logo in the statusbar, it does not hide the rest of the information displayed on the screen.

https://github.com/g4klx/MMDVMHost/blob ... D.cpp#L619
No, but if it can't transmit or receive then it's unlikely that there is anything to display. I had exactly this myself, replacing the TCXO fixed it. I did have messages in the log indicating problems, but they were not of the "can't do anything, can't find firmware ID string" type. Thinking very hard (this was in 2018), I recall messages containing "RF queue" or something similar. This happened after the Phoenix IPSC2 server connection was made and the PiStar was trying to transmit the "connected to reflector" audio message.
--

Brian G8SEZ
User avatar
G8SEZ
Posts: 553
Joined: Fri Apr 13, 2018 8:26 pm

Re: Curious about a failure

Post by G8SEZ »

KE7FNS wrote: Fri Jun 12, 2020 8:38 pm
G8SEZ wrote: Fri Jun 12, 2020 8:01 pm No, but if it can't transmit or receive then it's unlikely that there is anything to display.
The OLED would of still displayed information after booting and when it entered the first idle mode since that has no direct relation to TX or RX. I also believe it would still be displaying information every idle mode even without a proper TCXO.
Interesting, this isn't what my Pi0W plus Jumbospot PiStars do, in particular I have one that is on the Brandmeister UK server, reflector 4400 which never shows anything in the Trx field on the dashboard until it has seen activity from at least one local or network station. Pretty sure it doesn't show anything on the OLED either, but maybe that appears when something has communicated with the radio successfully.

When I get a clear hour I am going to have a look at this in more detail and check my settings carefully.
--

Brian G8SEZ
kd4iz
Posts: 8
Joined: Fri Apr 27, 2018 3:16 am

Re: Curious about a failure

Post by kd4iz »

Thanks for all the theorizing and the potential solutions. Turned out the problem was most likely a corrupted configuration and a too small SD card. I am not sure if that is what actually drove the failure, but all is in balance now.

Interesting observation: I could not get the latest image (5/28) all the way through configuration so used the March image and updated/upgraded my way to present day. The May image would repeatedly forget settings and give me a ration of crap - particularly by resetting the hat to naught and stripping out the country settings. (Come on y'all I know we are "toxic" but don't take it out on us ;) ).

I normally don't have the patience to wait for a Pi Zero to figure out what it is supposed to be doing, so in the past I have used a Pi3/4 to configure, then moved the hat and SD to a Pi zero to run... Out of an abundance of caution and to my frustration I walked through the entire config on a Pi Zero.

Since I acquired a new hat of the same variety, I put it on another known good Pi Zero and configured it from the ground up with a new 32GB SD. Once that was working I swapped hats and the old hat "found its3lf" so I backed up the config to disk. Next put in a new March image on a fresh SD card, booted, restored the image config from disk and then tweaked the configuration to individualize that one... now both hotspots work nominally. I'll burn them in for a day or two then decide whether to keep both or find a new home for one. I have too many anyway.

KD4IZ
User avatar
G8SEZ
Posts: 553
Joined: Fri Apr 13, 2018 8:26 pm

Re: Curious about a failure

Post by G8SEZ »

Glad you found the problem, remember that anything with more than 2 components is a complex system!
--

Brian G8SEZ
Post Reply