It seems the firmware upgrade fixed all the issues. Will continue to monitor and follow up. Thanks.
Hi, I had the solid green issue after an afternoon of playing with soft-ap-setup. Prior to solid green, I had a number of times gotten to the stage of successfully configuring a SSID and attaching to the network.
The problem is 0.4.7 appears to have taken away the wipe SSID credentials stage - ie holding reset + setup, then setup for 10 seconds. All you get is DFU yellow.
This is a massive step backwards in functionality. How are we supposed to clear creds?
Edit: Ignore me, found I can do the same with just pressing setup for 10 seconds.
This worked for me, thanks @rohit1101190!
Upgraded to 0.4.7 using dfu-util method - goes right back to steady green light. Canât get into any other mode than DFU. Any other suggestions?
Could you try the CLI with particle update and then particle flash --usb tinker?
Weâve seen a couple of instances in which Photonâs went offline and were discovered to have a solid green light. The WiFi network is WPA and the software was compiled and flashed about four months ago, so Iâm guessing the device is running something like 0.4.9. Any thoughts on what might be causing the solid green light?
I may recompile and refresh with the latest system firmware.
This is a test installation with the potential of becoming a larger opportunity, so weâd like the system to be rock solid. Thanks for any thoughts.
@ctmorrison you might be seeing this issue: https://github.com/spark/firmware/issues/904
Would you please upgrade to 0.5.2 and see if the problem is resolved for you. It may be hard to tell though, unless you are seeing it frequently already.
Will do soâŚas soon as it comes back online! Itâs a fair distance from here. 
sorry about that!
If it has the problem I mentioned, it's likely not going to come back online until it's reset or power is cycled. Have you found that it will eventually come back online all by itself?
This also might be a good case for the Application Watchdog to protect against issues like this: https://docs.particle.io/reference/firmware/photon/#application-watchdog
Iâve not seen it personally. They guy Iâm helping tells me he has found it âsolid greenâ on at least a few occasions.
If I want to employ the watchdog, can I just use âApplicationWatchdog wd(60000, System.reset);â in setup()? I have no long running process. My app just loops waiting for a certain amount of time to elapse before it publishes data to a cloud service via a webhook.
The other 5 (well 4 of the 5) devices are stable and run the same code base. Yesterday, I recompiled and reflashed the app and new system firmware (0.5.2) on all devices (except the subject device, obviously) to see if it helps with stability. One of the other 5 devices had been problematic, too.
(update) I donât have these problems anywhere else, which makes me think itâs something about this installation. We are using an âoldâ Linksys WRT54 router. Perhaps itâs the culprit and perhaps 0.5.2 will help with stability. It may simply be wise to update the router to something from this century! 
I'd have to see all of your firmware to really know for sure, but in general you want to make sure this timeout value is well beyond any possible nuisance-tripping, and if there are any known/potential long blocking loops in your code you'll want to add wd.checkin(); in them.
Shouldn't be a problem. I retired my WRT54G just a few years ago when streaming services overran it's performance capabilities.
I confirm ( like rohit1101190) that I fixed a âtimeoutâ problem (Impossible to flash , always fix green, and impossible to enter the Listen mode ) on a 0.5.3 photon and Windows 8.1 that way :
see Cannot get photon back to listening mode for better explanation: (thanks to them)
- put the Photon on DFU mode
- check that you have the correct driver for DFU mode . You should see here a libusbK USB devices in the device manager (only when you are in DFU mode!) . if Not, use ZAdig_2.2 to install a libusbk driver
- in CMD mode write : particle flash --usb tinker
Now you should have again possibility to enter in listen mode - do still a : particle indentify
that should be fixedâŚ