Sandaysoft

I thought it might be useful to start a thread where people could report the problem where the station stops responding to requests for data and requires a reset. Perhaps we can see a pattern. It's quite possible that it's an issue with Cumulus, but the same problem does occur with other software. The problem does seem to have become more common recently, but I haven't changed the code which interfaces with the station for a very long time. It appears that Fine Offset have done some kind of change to their stations in the last year or two which causes this problem.

So, if you find you have to reset your station to get it talking again, please post here. Give all relevant information - the 'brand' and model of your Fine Offset station, which build of Cumulus you're using, how long it was since the last time you had to reset, your station's logger interval, whether or not you were seeing 'sensor contact lost' shortly before the lockup, and whether you run Cumulus 24 hours a day or stop and start it from time to time. And any other information you think might be useful.

Note that this is NOT the same problem as the 'Station not initialised' problem. Click on the Cumulus error light, and if the error message says 'Station not initialised', that is a different problem and you need to read this FAQ: http://wiki.sandaysoft.com/a/FAQ#When_I_start_Cumulus.2C_I_get_the_error_message_.22The_station_is_not_initialised.22

Please don't post about other problems in this thread.

_________________
Steve
Sanday Weather
----------------------------------------------------------------------------------------------------------------------------------
Like Cumulus and want to support it? Please donate!

From 'other' software anecdotally the problem seems to have started with more recent examples of the stations, older stations do not seem to showing the problem. So maybe if people could say when the station was bought it may be interesting.

It would also be interesting to hear from anyone who is using Easy Weather - are you getting lockups? Apparently EW does use an unusual memory read sequence.

_________________
Mark
Wilmslow Astro Weather

That's the first point I was going to make. I have two stations. One locks up, the other doesn't. The one that locks up was bought in August 2011.
That would be someone using EasyWeather who leaves it running continuously, 24 hours a day, for weeks at a time.

The 'unusual' pattern I've observed is that EasyWeather finishes its read sequence (15 second intervals) by getting the data from 32 bytes lower than the current pointer. This appears to be pointless, unless it's to leave the USB 'access address' (if there is such a thing) pointing at 'inactive' memory. I changed pywws to follow this example, but still had a lock up. Another theory bites the dust.

I'm in contact with someone at Fine Offset, but he doesn't (yet) accept there's a problem. I'm hoping I can get some more info from him about the USB interface, but I'm not sure if there is any. (I'll PM Steve with the useful info I've got so far, although it's irrelevant to this problem.)

_________________
Jim

Without checking other unit that have stopped working I could not say if the problem is software or hardware related. My unit has stopped working many times. Until last year most of the tricks to getting it going worked, unplug it, resetting and even letting it just sit for a few days(unplugged) and trying again worked. But last year just nothing worked, its a Fine Offset (Branded National Geographic NG265) which ever model that works out to be purchased in 2009.
I got it going 2 weeks ago and would have to say it is a hardware problem with mine at least, see attached.
Hopefully some that have this problem would be able to take a voltage reading as discribe in the attached file.
One thing that I found that would caused it to stop communicating was restarting the laptop with the unit attached.

I will throw this one into the mix just in case its relevant. I worked with USB a lot when working as a software engineer on the UK/US chipped passports. The devices that read the chips on the production lines were USB attached and we had problems with the readers loosing connections to the PC running the software checking the chips.
Two things we found were as follows:
1) Quite a few machines (laptops especially) were incapable of providing the correct power to keep the device running and the only way to guarantee a smooth, non fluctuating power feed was to use powered hubs. That in itself caused a problem as you had to make sure the powered hubs were fully 2.0 USB compliant, a lot of them are 1.0 hubs cheaply modified to support 2.0 that did not fully match the spec nor the required power levels.
2) Windows has settings in the Power Management functions to turn off what it considers idle hubs. If you look in Device Manager and right click on Generic USB Hub or USB Root Hub (anything that has Hub on the end, there are some strange ones!!), and select Properties, you will see a tab labelled Power Management. Click on this and you may find that "Allow the computer to turn of this device to save power" has a tick against it. Untick it and Windows will not arbitarily turn of your USB device when it enters the Power Management functions.
I made sure I did both of these when I attached my Weather Station last April, and to date (crossed fingers, touch wood) I have not had one single disconnection or lockup.

_________________
David
http://www.dmjsystems.co.uk/weather/

My unit is pretty much... intermittent. I don't like the very irregular update times so I tried to move it like, what, one centimeter and it immediately stopped working after around 5 minutes. I let it search for signal once, nothing. Disconnected USB cable and the signal came back... Odd.

_________________

MeteoPortalegre

I had mine connected to an Ubuntu box running wview for a while and ran into the same issue on the that machine. I usually end up disconnecting my WS-2080 bought this year every couple of days. It does seem I have a large number of Sensor contact lost errors a 1/2 to full day before hand.

Hi,

No, not Odd, this is well-known behaviour of the FO units, it appears that the USB cable can radiate interference. If the Console displays and logs properly when the USB is disconnected then a better/screened cable may help (note that these are "USB Male A to Male A" which is rather non-standard). Also, at the frequency used by the radio link, multipath "standing waves" (due to signal reflections) can cause the signal strength to change dramatically with a movement of just a few centimetres.

The Console only reports "Lost contact" (and "--" on the display) after about 6 - 8 consecutive transmissions are "lost", so about 5 minutes can elapse if communications fail completely. This makes "testing" the wireless link quite difficult because the comms can be "struggling" (say > 50% of messages lost) without any obvious "errors". You need to see if the data is updating regularly every 48 seconds (on the Console LCD) and thus require continuously changing data - the rather "lively" (undamped) wind vane can be of some help here.

Sorry this is rather Off-Topic for the Console-USB lockups which Steve is trying to identify in this thread (but it is quite possible that poor radio comms could be a cause of these lockups).

Cheers, Alan.

Well I added an antenna to my other receiver with the wind direction fixed. It seemed to start working very well where it is now, consecutive readings and everything, even when connected to USB, but after a few hours it's not constant anymore even though it never loses signal, it's a little irritating losing wind gusts. It pauses for around 2 minutes, then updates, sometimes even more on the average wind value. Happens the most in weak wind but I've been observing it with winds around 20 km/h too. So at the start it was all fine, now I haven't moved it and it updates unconstant again.

_________________

MeteoPortalegre

I have had intermittent "lost contact" problem when using two transmitters/receivers. All will work well for a few days then one will lose contact for anything up to 4 hours then return to normal. Turning off the second system and the one feeding into the PC and Cumulus become rock solid again 24/7.

_________________
Leumeah NSW AU Weather

I experienced lockups and "lost contact" problems from virtually when the unit was new (Dec2011). This only happened when the console unit was connected via USB cable. After replacing the USB cable I have experienced no further problems, so it would appear that that was the cause.

_________________

The pattern I noticed with mine was that the barometer value would no longer change whenever it locked. Everything else seemed to record fine, just the baro.It didn't seem to matter how full the little memory indicator was either.

Good thread, thanks Steve

_________________

Interesting. I have to confess that my USB connection goes via two extension leads, which is not ideal. It may be that my old station was less sensitive to USB signal quality than the new station is.

My 'pywws' software has always read data from the station twice, and only uses data when the two readings agree. I did this to prevent taking readings during 'mid update', but it might also be masking USB problems. I'm now logging any cases where the two readings differ, and am seeing some when reading memory that should not be changing (logged data 32 bytes below the current data).

_________________
Jim

Just changed USB leads on mine - locked up again within a few hours. Same symptoms, no change of pressure reading, everything else appears to be reading ok (discounting the solar which is yet again broken!!)

This time I have done a complete station reset (up button held in for x seconds until the screen shows the reset in progress).

This has cleared the memory so I'll see how long it lasts. I have a theory that the more memory is used the more often these lockups occur Time will tell.

Ordering a Davis 6162 today

Edit: Easyweather also locks up so I don't think it's anything to do with Cumulus - unless you can empty that memory buffer Steve?

_________________

All good for the last 12 hrs or so with the memory empty

_________________