Cumulus stopped - Interesting observation

[werribee_au]

I checked my station web pages this morning and noticed that it had not updated since 6:15am, (It was currently about 7:00am). Checked cumulus and there was zero for all the observations. Checked the may10.txt file and entries stopped at 6:15 and the today.ini timestamp was at 6:15am.

So I thought, what happened at 6:15am when I got up? I turned the on double fluorescent light about 2 meters from the receiving console. The only thing I can think of.

I reckon that the RFI (Radio Frequency Interference) generated by the fluorescent light starting up probably just at the instant the transmitter was sending it's data to the receiver affected the reception of the data packet, and as I have read somewhere in here that cumulus reads the data entries as they are stored, corrupted or not, this is why I reckon things may have fallen over.

It might be worth building a 3 element yagi or even a simple λ/2 dipole and pop it outside to see if that improves things.

Does anyone know the impedance of the receiver rf input? 50Ω would be the value I'd reckon. For such a low power transmission I would want to reduce any losses in the receiving antenna circuit. I have a broken GSM modem antenna that I could use the coaxial cable (around 3mm dia) from. I guess it would be the low loss variety.

Your thoughts?

[steve]

Even if the outside transmitter had been interrupted in some way, there should still be data from the console itself (pressure etc). But if Cumulus thought that the outside data was corrupted, it would ignore all of the data. It's odd that the data should go to zero, though, you would expect it to stay at the last good reading. If you zip up the diags folder and attach it, I'll have a look to see what Cumulus did at 06:15.

[werribee_au]

Thanks Steve,

I was thinking about the usb that stopped working.

Here are the diags.

[steve]

Yes, definitely some sort of USB problem:

16/05/2010 06:15:00.406 : Writing todayfile, LastUpdateTime = 16/05/2010 06:15:00 raindaystart = 486.299987792969
16/05/2010 06:15:01.281 : Error!!! Failed to write to USB!
16/05/2010 06:15:05.078 : 06:15:05 EWUSB: HidControl Device Unplug
16/05/2010 06:15:05.187 : EWUSB: HidControl Device Removal
16/05/2010 06:15:05.187 : 06:15:05 EWUSB: HidControl DeviceChange
16/05/2010 06:15:05.203 : EWUSB: HidControl Device Removal
16/05/2010 06:15:05.203 : 06:15:05 EWUSB: HidControl DeviceChange

What did you do to get it working again? It looks like you had one restart where Cumulus couldn't detect the station, and then on the next restart all was OK. Did you have to unplug and reconnect the station?

[werribee_au]

What did you do to get it working again? It looks like you had one restart where Cumulus couldn't detect the station, and then on the next restart all was OK. Did you have to unplug and reconnect the station?

Yes I saw the restart at 6:15, Cumulus must have restarted itself as the system didn't reboot due to the uptime being 4 hrs 13 mins (Scheduled System Reboot 2:00am each morning).
Strange eh?
6:59am is when I shut down Cumulus.
Then,
Disconnected the USB from the console end and then reconnected.
Had a look at the may10.txt & today.ini files
Started Cumulus and it downloaded the history.
All still ok for the moment.

I reckon it is the usb on the console side of things that is stopping rather than the notebook. But how do you determine which it is? Maybe if I get a usb hub and when it crashes look to see if the hub is still there? Don't know.

I might have a listen to the fluorescent lights starting up on the scanner and see of changing the starters and/or tubes make any difference.

[AndrewMoni]

I had something very similar with my WH1080 - look through the forum and you may find my thread. I kep getting zero readings and having to reset the console to get things going again - though in my case the console readings themselves went to zero as well as losing USB communication. This went on for ages - it would fail after a variable timeperiod. I finally had the console replaced - and the new one has not gone wrong yet (touch wood) so there WAS something wrong with the USB port on the console with my original unit. tOUCHING WOOD is essential methinks with the 1080 - am here now in the forum trying to work out how to reset a rain counter that suddenly (tho i suspect my cat) shot up this morning at 0420 and we have no wind or rain at present time. My new 2080 I got as a replacement appears much more reliable than the original 1080 (or had done until this morning anyway)

[burgla]

Was actively looking at the console, as the wife came up the drive & pressed the remote garage door opener. the rain figures went to - - - - , at the same time the garage door started to open. 10-15 seconds later the figures came back on again. I got her to close it with the hard wired switch on the wall & it made no difference to the display. Methinks the 433.9 mhz opener caused interference to the transmission signal (433 mhz) from the station.
- just my 2 cents worth !!!

[Gina]

Was actively looking at the console, as the wife came up the drive & pressed the remote garage door opener. the rain figures went to - - - - , at the same time the garage door started to open. 10-15 seconds later the figures came back on again. I got her to close it with the hard wired switch on the wall & it made no difference to the display. Methinks the 433.9 mhz opener caused interference to the transmission signal (433 mhz) from the station.
- just my 2 cents worth !!!

Yes, it is very likely to interfere - the station frequency is also 433.9MHz according the the printing on the battery cover on mine!

[werribee_au]

In January I started having the above problems and started the following thread.
https://cumulus.hosiene.co.uk/viewtopic.php?f=13&t=2298

I had moved the station console and notebook to a new location in the house which was about 4 meters closer to the transmitter and out of the way of day to day life. It was placed on top of an upright freezer (I know, i was getting warmer indoor readings due to the condenser radiating heat from the rear of the freezer which didn't worry me). On the outside of the wall there is also a refrigerator. In the wall behind the freezer is a vertical mains cable supplying a power outlet for the refrigerator.

When I thought that the fluorescent light may have caused interference to the data reception of the station console I moved it and the notebook back to where it was in January before the relocation (Where I had not had any problems whatsoever for the previous 4 months). There now has not been any problems for the past week, touch wood.

If interference was the culprit I'd guess that the vertical mains cable 400mm away radiated the interference and the station received a corrupted data packet.

I have my fingers crossed

[Gina]

Sounds like you've found the problem, Gerard. Hope you don't get any more bad data

[werribee_au]

Thanks Gina, like I said my fingers (And toes) are crossed.

[werribee_au]

It's been 10 days now, no issues so far

[Gina]

Great

[werribee_au]

Still as solid as a rock!

2 weeks now without any issues whatsoever after moving the console and notebook away from electrical cables and the freezer and fridge. About the same distance from the fluorescent light and about 5 meters further away from the transmitter with 2 extra walls in the way.

[vern7485]

I have had similar events the WH1080 seems to be working but no action on Cumulus data reading. I have a debug txt which shows errors in arithmetic. Times are not predictable. All batteries are less than 1 yr old.