January 2, 2015 at 12:14 am #1091
One thing I noticed the other day was that my uRADMonitor locked up after a thunder storm. The cause is likely brownouts or blackouts of very short duration.
I can recreate the symptoms if I disconnect power to the monitor and connect it within 1 second; It lights up but won’t respond on the network. Waiting longer before applying power fixes the problem. I don’t know if this is due to the circuitry inside the monitor itself or the USB power adapter perhaps supplying lower than 5 volts during the power disruption (capacitors etc…)
For future versions of the product, could something like a watchdog timer be incorporated into the design, so that if some aspect of the firmware or network interface doesn’t respond, then an automatic reboot is initiated. This would be useful for permanent or semi-permanent installations.January 2, 2015 at 12:42 am #1094MikeParticipant
At least for now, maybe an email notification? Had the same thing happen to me a couple of times as well.January 2, 2015 at 6:29 pm #1097
I’m happy to report that there is a watchdog already:
If the unit is unable to connect to the server within 5 minutes, it will self-reboot.
You can try again, just wait 5 minutes.
RaduJanuary 4, 2015 at 5:06 am #1122
Sure, that’s all well and good to reboot if the network doesn’t respond etc… but what if the internal firmware itself isn’t responding (and thus unable to check the network).
The problem that I appear to be having (and this is just my best guess, I haven’t done any measurements) is that when voltage dips beyond a certain point (but still above 0 volts) the entire thing locks up and won’t recover without manual intervention. You can’t even ping the network interface.January 4, 2015 at 11:30 am #1124
If the voltage is too low, the unit will not function until the proper voltage is restored.
The minimum voltage required for the unit to work properly is 5V.
EDIT: If the firmware doesn’t respond there is little we can do from software side as all code control is gone. I believe you might be referring to an external watchdog of some kind – waiting periodic confirmation from uC that all is ok. If such confirmations are missing over a given time, it would flush the unit’s power to reboot.
This would make sense, but making PCB modifications at this stage is not exactly easy. The current design has been verified for quite some time before readying it for production. So it all goes back to the question: is this something necessary? I was unable to replicate the scenario. In all my tests, if something went wrong, after 5 minutes the units rebooted to normal state (quite a different story before this mechanism was implemented – 2years ago).
Yet if this is something that happens and it happens often, than I believe we should look for solutions.
January 9, 2015 at 7:25 am #1172
- This reply was modified 8 years, 11 months ago by uRADMonitor.
Yes, I’m talking about an external watchdog timer (sorry, I’m not an Electronics Engineer). I realise it would be difficult (and rather pointless) to spin a new board right now, but for future revisions if you decide to take this project further, maybe it could be a valuable addition? I’m sure they add very little cost overall.January 23, 2015 at 6:56 pm #1274WattSekundeParticipant
There is no hardware update necessary. You can use the atmega328p build in Programmable Watchdog Timer.January 23, 2015 at 11:05 pm #1277
Starting with firmware v110, the hardware watchdog timer is in use.
- You must be logged in to reply to this topic.