Data Error - Check Connections....

[6719jason]

Good Evening Steve,

Thank you for adding the estimated Wet Bulb Temperature to the Latest Beta Posted a couple more tickets since then

I have a little bit of a weird error... See Screenshot Below....

DataErrorCumulus.PNG

Only thing is.... It's connected via a Real Serial Port with No Extensions in 'Wired Mode' and Time/Date/Year all set.

I'm sorry the attached log file is a bit cluttered with FTP Errors and File Permission, I've been having real issues with my ADSL Connection this last week or so, little too no upload and as a result I manually FTP'ed the dayfile.txt to the Webspace, forgetting it would take with it the Windows File Permissions which cause an issue when Cumulus tries to overwrite the file - My fault on that one, fixed now.

Anyhoot, Bt OpenReach Engineer visited today and reports back with a very recently congested Exchange and a few other niggles so while it is okay at the moment, the fault isn't fixed - just a temporary relief thank goodness! Was driving me

[steve]

This new code checks once a minute whether any data was read during the previous minute. It looks like with your station it sometimes takes more than a minute to read the data! Just after the hour, for some reason. Presumably the station does something on the hour which causes it to stop responding for a while. But only in the early hours of the morning? Or is this getting logged at other times?

I'll have to make the interval between checks configurable; that's a pain. But then La Crosse stations only exist to make things difficult for me

Could you please turn on the debug log a few minutes before the hour, and let it run until a few minutes after the hour. Each time it reads the data, it will print out a timestamped line ' ================= Starting data read ===================='. Then we can see how long it is taking to read the data.

[6719jason]

This new code checks once a minute whether any data was read during the previous minute. It looks like with your station it sometimes takes more than a minute to read the data! Just after the hour, for some reason. Presumably the station does something on the hour which causes it to stop responding for a while. But only in the early hours of the morning? Or is this getting logged at other times?

I see - Data is transmitted every 8seconds and presumably the Console is read every 10seconds by Cumulus? I wonder if it was trying to acquire the DCF Signal at those times? I haven't seen it logged at any other period of time or day.

The DCF Signal Icon is currently not shown on the Console, which me is both good and bad as the DCF Signal tends to play havoc with the Time/Date Settings thus Logger issues then ensume until I correct the Time/Date...

I've just checked the Error Screen and can see that it experienced the same difficualties as the previous night starting at 03:01:00 and continued on the hour until 06:01:00... Hmmmm

I'll have to make the interval between checks configurable; that's a pain.
But then La Crosse stations only exist to make things difficult for me

I wish they didn't make things difficult, I like my La-Crosse Station

Could you please turn on the debug log a few minutes before the hour, and let it run until a few minutes after the hour. Each time it reads the data, it will print out a timestamped line ' ================= Starting data read ===================='. Then we can see how long it is taking to read the data.

I shall do that tonight when I expect the Error will again reappear.

[steve]

presumably the Console is read every 10seconds by Cumulus?

It reads all of the data, then waits 5 seconds, and then starts reading it all again. I don't know how long it takes to read the data; possibly quite a while as it's very hard to get data out of a La Crosse station, it tends to ignore the PC most of the time.

I wonder if it was trying to acquire the DCF Signal at those times? I haven't seen it logged at any other period of time or day.

I think you've probably got it. Is there no way to disable the DCF?

I shall do that tonight when I expect the Error will again reappear.

Thanks.

[6719jason]

It reads all of the data, then waits 5 seconds, and then starts reading it all again. I don't know how long it takes to read the data; possibly quite a while as it's very hard to get data out of a La Crosse station, it tends to ignore the PC most of the time.

I see, I wasn't really aware how it was done but I now know. Thank you for explaining it to me

I think you've probably got it. Is there no way to disable the DCF?

Afraid not, or at least not that I am aware of Thats mostly why I put in a Request for a similar Watson/Fine Offset Logger system for Cumulus.

Thanks.

Your welcome Steve

[steve]

Afraid not, or at least not that I am aware of Thats mostly why I put in a Request for a similar Watson/Fine Offset Logger system for Cumulus.

OK - I'll see what I can do. I think that may be easier than I initially thought.

[6719jason]

OK - I'll see what I can do. I think that may be easier than I initially thought.

Thanks Steve Installed and running the latest Beta with the 'Ignore La-Crosse Clock' ticked - I then restarted Cumulus and all is well so far - Once again, Many Many thanks to you Steve

[steve]

Thanks Steve Installed and running the latest Beta with the 'Ignore La-Crosse Clock' ticked - I then restarted Cumulus and all is well so far - Once again, Many Many thanks to you Steve

Did it get chance to download any logger entries? You might want to try stopping it for a few minutes just so it downloads a couple of logger entries, and check the timestamps in the July log file. They won't be exactly the same as they would have been with the station clock as it doesn't know exactly when the last entry was logged (or, rather, it's ignoring that information), but they should be close enough.

[6719jason]

Did it get chance to download any logger entries? You might want to try stopping it for a few minutes just so it downloads a couple of logger entries, and check the timestamps in the July log file. They won't be exactly the same as they would have been with the station clock as it doesn't know exactly when the last entry was logged (or, rather, it's ignoring that information), but they should be close enough.

I'll close down Cumulus for 10mins, it should then attempt to download Two Logger Entries (One every 5mins) - I'll check the July Log file and report back shortly!

[6719jason]

Just a quick report back...

I left Cumulus for I think around 15mins? And all seems well!

03/07/11,22:50,14.7,72,9.7,0.0,0.0,0,0.0,0.0,1016.8,141.9,24.1,48,0.0,14.7,14.7,0.0,0,0.00,0.00,14.7,0,0.0
03/07/11,22:55,14.6,72,9.6,0.0,0.0,0,0.0,0.0,1016.9,141.9,24.2,48,0.0,14.6,14.6,0.0,0,0.00,0.00,14.5,0,0.0
03/07/11,22:59,14.6,72,9.5,0.0,0.0,0,0.0,0.0,1017.0,141.9,24.0,48,0.0,14.6,14.6,0.0,0,0.00,0.00,14.5,0,0.0
03/07/11,23:04,14.5,72,9.4,0.0,0.0,0,0.0,0.0,1016.7,141.9,23.9,48,0.0,14.5,14.5,0.0,0,0.00,0.00,14.4,0,0.0
03/07/11,23:09,14.4,72,9.3,0.0,0.0,0,0.0,0.0,1016.8,141.9,23.8,48,0.0,14.4,14.4,0.0,0,0.00,0.00,14.3,0,0.0

Last five lines of Jul11.txt

Thanks Steve - Cracking job!

[steve]

Jolly good. We'll see if we can sort the spurious 'connection error' messages next.

[6719jason]

Jolly good. We'll see if we can sort the spurious 'connection error' messages next.

Once again, Many many thanks to you Steve!

I'll turn on the Debug before I go to bed - will probably be around the time of the Error Messages though

Though we hope.... they don't return!

[6719jason]

Hi Steve,

Here is my long awaited Debug Log...

Jason

http://www.weeleyweather.co.uk/debug-e.txt - File is too big for forum

'05/07/2011 03:01:00 : Data input appears to have stopped - check your station and connections
05/07/2011 03:02:00 : Data input appears to have stopped - check your station and connections
05/07/2011 03:03:00 : Data input appears to have stopped - check your station and connections
05/07/2011 03:04:00 : Data input appears to have stopped - check your station and connections
05/07/2011 03:05:00 : Data input appears to have stopped - check your station and connections
05/07/2011 03:06:00 : Data input appears to have stopped - check your station and connections
05/07/2011 03:07:00 : Data input appears to have stopped - check your station and connections'

In Cumulus Error Window

[steve]

1001.14984 : 02:57:45 ================= Starting data read ====================
1001.23140 : 02:57:53 ================= Starting data read ====================
1001.71031 : 02:58:41 ================= Starting data read ====================
1001.78843 : 02:58:48 ================= Starting data read ====================
1001.87109 : 02:58:57 ================= Starting data read ====================
1002.28156 : 02:59:38 ================= Starting data read ====================
1002.36656 : 02:59:46 ================= Starting data read ====================
1002.45015 : 02:59:55 ================= Starting data read ====================
1002.53000 : 03:00:03 ================= Starting data read ====================
1007.08297 : 03:07:38 ================= Starting data read ====================
1007.16468 : 03:07:46 ================= Starting data read ====================
1007.24453 : 03:07:54 ================= Starting data read ====================
1007.71859 : 03:08:41 ================= Starting data read ====================
1007.79797 : 03:08:49 ================= Starting data read ====================

It goes along reasonably well, most of the time it gets the data in just a few seconds, but at 03:00 the station just stops responding for over five minutes. Cumulus does lots of retries, but doesn't get anything back. So it seems that for around 5 minutes or more at 03:00 the station is too busy doing something else (as you suggested, probably trying to sync to the time signal).

I'm not really sure what to do about that, other than make an option for turning off those error messages. And you probably don't want to use the new 'restart if data stopped' option!

[6719jason]

It goes along reasonably well, most of the time it gets the data in just a few seconds, but at 03:00 the station just stops responding for over five minutes. Cumulus does lots of retries, but doesn't get anything back. So it seems that for around 5 minutes or more at 03:00 the station is too busy doing something else (as you suggested, probably trying to sync to the time signal).

I think you are spot on there Steve! Stupid really as it's not essential where as possibly Weather Data during those five mins could be!

I'm not really sure what to do about that, other than make an option for turning off those error messages. And you probably don't want to use the new 'restart if data stopped' option!

Yup I think thats all the could be done - Possibly only for La-Crosse Users as I don't think other stations have this issue as such!

Oh you bet I don't use that!

Thanks once again for all your efforts on trying to resolve this issue!