Toshiba 50UL2063DB randomly rebooting with HDMI-CEC and/or HDMI-ARC enabled

[spleenharvester]

I have this TV hooked up to only an AppleTV 4K and a Creative soundbar. Originally HDMI-CEC and HDMI-ARC wouldn't work at all, I replaced all the HDMI cables with decent UGREEN cables meeting the latest standards, now it all works perfectly but once or twice a day the TV itself reboots while watching a program, the AppleTV keeps running. I haven't fully pinned down whether it is ARC or CEC causing it yet. All are running latest firmware.

I assume it is just HDMI-CEC being buggy as usual on a shitty Vestel machine, but can anyone think of anything I can try before disabling it?

[spleenharvester]

I need to wait longer to be sure but I think I have actually figured this out - software problem related to the WiFi. It's hard to tell exactly what is happening with the lack of date/timestamps in the log dumped to USB, but it was getting spammed full of this

[1;37;41mcheck [0/1] COM.NVT.PRIVSERVER[0m[30;41mICE_WIFI ERR : [0m[1501]ice_WiFi_MTK_ScanAp():Retry over 20 times, scan time out
    > ecl/wificontroller: Scanning access points.
[0mICE_WIFI_WAR : [0m[2421]ICE_WiFiConnectAp():create monitor task!
[0mICE_WIFI_WAR : [0m[2432]ICE_WiFiConnectAp():[END] ret=1
    > ecl/wificontroller: ret: 1​

Followed by messages like this

[1;37;41mcheck [0/1] COM.NVT.PRIVSERVER[0mcat: write error: No space left on device

It occurred to me one thing I did when I enabled CEC was factory reset the TV and connect to WiFi to check for updates, which I hadn't done previously. When I went to check the WiFi settings in the GUI I noticed it was just hanging repeatedly trying to connect to my router, which was bizarre as the credentials were 100% correct. We run a PiHole on our network, and as soon as we temporarily disabled blocking it connected no problem. My best guess is that it was trying to phone home to an address on my PiHole blocklists after authenticating to check internet was connected, and when it couldn't do that it would just try and do it over and over again until it ran out of memory. It would also explain why the reboots were always spaced about 2 hours apart and never in quick succession. Interestingly I also noted in the log files that it seems this model actually can't authenticate with the software upgrade server properly because of having no date/time setting. It defaults to 01/01/2021 whenever it's powered up, and tries to get the update file from an AWS address. If you enter that address into your browser it tells you access denied because the token expired on 01/01/2021.


Anyway I disabled networking entirely and it's been running for about 8 hours without any problems. I'll update this post if it is definitely fixed.

[spleenharvester]

Update, it is still restarting with CEC and ARC disabled. I am starting to think it happening at the same time as replacing all the cables was just a coincidence, and that the TV is actually kaput. It's now restarting about half a dozen times per day, every couple of hours ish. There's no abnormalities displayed at the time it does it, it's literally just as if someone turned it off and then straight back on again.

Before I replace the PSU board, any ideas on what I can try/check? All components on both the PSU board and the mainboard look physically OK.

[nomoresonys]

Can you just try a simple antenna with over the air, unhook all other hdmi etc, see if it still does it.

[spleenharvester]

Unfortunately our roof antenna hasn't worked in a long time lol

I found last night that somehow the TV turned HDMI-CEC back on by itself so I've turned it back off again, will wait and see if it resolves it. I've also ordered a PSU board as managed to get a working one for £8 delivered. Even if it turns out the board is fine it probably doesn't hurt to have a spare, it's an old set.

[nomoresonys]

A spare is never a bad idea, you can use just an old piece of coax with a stripped end where the wire is exposed a little for an antenna just to get a nearby station, just for checking, will probably work if towers aren't too far away.

[spleenharvester]

Will pick up one of those mini coax antennas to test with I think. I just changed the HDMI cables and all the settings back to what they were before I started messing with HDMI-CEC and it is still rebooting, so I am now fairly sure it is a coincidental hardware problem. I might recap the new board while it's out on the bench too.

Can anyone think of a particular reason why, on UK 240V mains, I would need a 200V cap on the isolated side? There is a 200V 56uF 10x20mm capacitor and I am struggling to source a replacement... I can only find 50V 56uF. If that won't suffice I might just have to leave that one alone. (Edit, it looks like the 65" version of this TV came with two 47uF 200V caps in place of that 56uF 200V cap, so I might be able to do that instead?)

[nomoresonys]

47UF should be ok, you just never want to go with less voltage so 200 or a little more like 250 or so is fine.

[spleenharvester]

Ok thanks I'll get some on order,

I've enabled USB logging through the service menu (home -> 4725) and left it running until the TV reboots. From the first attempt there are a few suspect lines suggesting it is trying unsuccessfully to install an update but it is mostly just verbose noise, and the TV has no date/time setting so there's no clear indication of what is happening when.

I've found the 17MB230 schematic, am I correct in thinking TP254 and TP255 on the Novatek NT72671TBG IC are the UART pads? Wondering if I could get better information with an RS232 connection

[spleenharvester]

I need to wait longer to be sure but I think I have actually figured this out - software problem related to the WiFi. It's hard to tell exactly what is happening with the lack of date/timestamps in the log dumped to USB, but it was getting spammed full of this

[1;37;41mcheck [0/1] COM.NVT.PRIVSERVER[0m[30;41mICE_WIFI ERR : [0m[1501]ice_WiFi_MTK_ScanAp():Retry over 20 times, scan time out
    > ecl/wificontroller: Scanning access points.
[0mICE_WIFI_WAR : [0m[2421]ICE_WiFiConnectAp():create monitor task!
[0mICE_WIFI_WAR : [0m[2432]ICE_WiFiConnectAp():[END] ret=1
    > ecl/wificontroller: ret: 1​

Followed by messages like this

[1;37;41mcheck [0/1] COM.NVT.PRIVSERVER[0mcat: write error: No space left on device

It occurred to me one thing I did when I enabled CEC was factory reset the TV and connect to WiFi to check for updates, which I hadn't done previously. When I went to check the WiFi settings in the GUI I noticed it was just hanging repeatedly trying to connect to my router, which was bizarre as the credentials were 100% correct. We run a PiHole on our network, and as soon as we temporarily disabled blocking it connected no problem. My best guess is that it was trying to phone home to an address on my PiHole blocklists after authenticating to check internet was connected, and when it couldn't do that it would just try and do it over and over again until it ran out of memory. It would also explain why the reboots were always spaced about 2 hours apart and never in quick succession. Interestingly I also noted in the log files that it seems this model actually can't authenticate with the software upgrade server properly because of having no date/time setting. It defaults to 01/01/2021 whenever it's powered up, and tries to get the update file from an AWS address. If you enter that address into your browser it tells you access denied because the token expired on 01/01/2021.


Anyway I disabled networking entirely and it's been running for about 8 hours without any problems. I'll update this post if it is definitely fixed.

[nomoresonys]

Well done, very nice detective work, glad to see you solved it.

[spleenharvester]

Cheers :-)

Three days and no reboots with HDMI-CEC and HDMI-ARC re-enabled too, so confident it is fixed now