Problems with station communication: USB locks up completely

[steve]

You don't appear to have enabled the synchronisation option?

[rwilliam99]

I missed that - I'm assuming you are referring to the "Syncronize Fine Offset reads" option in the settings dialog? I had assumed that it was the default behavior in the new build, but it seems that was a bad assumption. Sorry! I enabled this and restarted Cumulus.

I've played around with the webtags functionality in Cumulus Toolbox, but I'm apparently missing something - I've tried all sorts of variations of DataStopped (the tag I think is really what I want), but I never get any email. I had a rule for DataStopped = 1 and DataStopped = 0 (both conditions, so one of them should fire), and I'm not getting any emails. I do have the system check for Cumulus running and that DOES work, so I know the email function works.

[steve]

I've played around with the webtags functionality in Cumulus Toolbox, but I'm apparently missing something - I've tried all sorts of variations of DataStopped (the tag I think is really what I want), but I never get any email. I had a rule for DataStopped = 1 and DataStopped = 0 (both conditions, so one of them should fire), and I'm not getting any emails. I do have the system check for Cumulus running and that DOES work, so I know the email function works.

It's possible that the DataStopped tag isn't in the version of the PHP webtags you're using - you'll need to check.

[pete_c]

I haven't had any lock up issues with the Fine Offset weather station in Florida since early May, 2012.

That said I didn't notice the newest beta included an update for the Fine Offset.

Installed it today remotely and all looks well right now.

[ejay]

Note that this isn't a fix, it's a work around. And because the process of synchronising the reading of data necessarily involves reading the data - at a higher rate than normal - there is always going to be the risk that the problem will occur during the synchronisation process itself. The risk of the problem occurring overall should be greatly reduced, however.

So I presume that once Cumulus has determined the intervals at which the station is receiving data (by reading more frequently to detect when the memory data has changed), it should not clash thereafter, assuming that nothing else goes wrong?

Does Cumulus remember the 48 and 60 second intervals if it is restarted for whatever reason? e.g. if the "Restart if unplugged" setting is enabled, and Cumulus does restart because of that, will it cause the synchronisation process to run again, thereby heightening the risk of corrupting the memory?

[steve]

So I presume that once Cumulus has determined the intervals at which the station is receiving data (by reading more frequently to detect when the memory data has changed), it should not clash thereafter, assuming that nothing else goes wrong?

That's the theory, yes.

Does Cumulus remember the 48 and 60 second intervals if it is restarted for whatever reason?

Yes.

It re-synchronises every 24 hours, regardless of whether it's been stopped or not in the meanwhile.

[ejay]

If it re-synchronises every 24 hours, doesn't that mean there is the possibility of a clash every 24 hours? As you said before, synchronising results in reading from the station more frequently. It sounds like it defeats the purpose.

[steve]

See Jim's posts earlier in this thread. According to his research, the Fine Offset clock can drift by 2 seconds a day, so resynchronisation is necessary.

My understanding is that none of Jim's users have hit this problem again since he implemented his work around. And, as far as I'm aware, no-one who has enabled the option in Cumulus has hit the problem so far, since I made it available nearly three weeks ago. I'm assuming that significant numbers are actually using it, given the number of people previously hitting the problem, and I'm assuming that if someone had hit the problem while using the option, they would have reported it here. It would be reassuring to see some reports from people who were hitting the problem frequently before, and haven't hit it since. But I guess three weeks isn't really long enough to gain any confidence.

[stormbay]

Hi - I set up my WH3081 on Xmas day and used the EasyWeatherPlus software to set and gather data. I loaded Cumulus V1.9.3 1057 around the 29th Dec and have left the PC running 24 x 7. In that time, I have had a 'lock-up' as described whereby the console works but the Cumulus failed to 'see' it. I ended Cumulus and started EasyWeatherPlus and whilst its icon could see the console (changed to green) all options were greyed out. I swapped and fiddled until finally, I unplugged the USB from the console (I had previously been unplugging at the computer end) and suddenly EasyWeatherplus recognised it again. I then re-started Cumulus and away it went again. Since then, no lockups. I also note that memory on the device was 25% and I have not purged any data since I plugged it in. I might mention that since it has been running (DEC25-JAN3), I've had to climb the roof and press reset on the UV\Solar twice (approx 3 days apart) because the solar displays ---

Philip
Mt Martha, Victoria
AUSTRALIA

[AllyCat]

I've had to climb the roof and press reset on the UV\Solar twice (approx 3 days apart) because the solar displays ---

Hi Philip,

Welcome to the forum. Did you try holding in the <down arrow> button on the Console for about 10 seconds first?

One cause of a full "crash" of the Solar Pod appears to be due to excessively ligh levels. Several people appear to have solved the issue by covering the Pod (or maybe just the UV sensor to the left of the sloping PV panel) with a slightly opaque film. Try searching the the forum for "3081".

Cheers, Alan.

[rpo]

I turned on the fix "Synchronise Fine Offset reads" and my USB link is still failing every couple of days. Did I do something wrong?

[steve]

I turned on the fix "Synchronise Fine Offset reads" and my USB link is still failing every couple of days. Did I do something wrong?

It may not be the same problem. If you zip up the diags folder and attach it, I'll take a look.

[rwilliam99]

Just wanted to let you know that the new beta version (once I enabled the synchronization option) seems to be working. I used to go about 5 days before a lockup - so far, I've been over 10 days and it is working fine thus far. Fingers crossed.....

[steve]

Just wanted to let you know that the new beta version (once I enabled the synchronization option) seems to be working. I used to go about 5 days before a lockup - so far, I've been over 10 days and it is working fine thus far. Fingers crossed.....

Thanks for the feedback. It does seem to be working, or at least increasing the period between lockups. I would think you are bound to get one eventually, let's just hope that's a long way off.

[jim-easterbrook]

See Jim's posts earlier in this thread. According to his research, the Fine Offset clock can drift by 2 seconds a day, so resynchronisation is necessary.

Note that this figure is based on a sample of one weather station. It's possible some stations may drift by more than this - I'm hoping to get some feedback from pywws users who can look at their log files and extract the drift rates.