Connection lost on WH 2080
wro...@tirnet.de
Running pywws +- successfully at a TP-Link TL MR3220 running openWRT
Life logger was active for about 3 days now.
While fiddling with mySQL synchronisation, I realised that my data had
stopped this morning.
Restart of life logger did not help - it simply seemed frozen, despite
the -vvv option, albeit 0% CPU usage after startup.
Same picture with TestWeatherStation.py
Unplugging / replugging USB did not help.
dmesg looks like correct unloading/loading of interface
While plugging / unplugging and restarting life logger or
TestWeatherStation.py, I sometimes got messages "ressource already in
use", but this may have only been timing issues.
Waiting several seconds before the next action made "ressource in use"
disappear.
The only thing that helped was to remove batteries from the station.
It now reads again for a hour or so, but the data since the hang were
lost.
In the first instance, it looks like a problem with the station
hardware / firmware.
But can it also blamed to a compatibility issue with software?
That pywws somehow screws up the interface?
Or could it be a power issue?
Many Chips are known to be designed for 3,3 V, but running at two
batteries gives only 3,0 V.
Consider attaching a voltage regulator and getting power from USB.
Did anybody try this?
Would it be possible to apply power cycle to the station by USB
commands that way?
Any idea how to manage that?
Any Experience with this type of station?
How to extend range by attaching an external antenna?
I't like to install the station in a tin sandwich walled building,
Sensors ~ 50 m away. I'm afraid that this will shield the signals.
I desperatly needed a somewhat reliable station for everyday decisions
on my farm.
Should I go for a station which is really professional, not only named
so?
Recommendations?
further details:.
http://www.conrad.de/ce/de/product/672286/FUNK-WETTERSTATION-PROFESSIONAL-USB
life logger still on open console (startup script on todo list)
python pywws-11.10_r429/LiveLog.py -vvv data
WUnderground had worked till just a minute ago but is gone now ("Data
not available").
- back again a time of pressing send button....
http://www.wunderground.com/weatherstation/WXDailyHistory.asp?ID=IBAYERNM13
although in my test setup, wind and rainfall sensors are not yet
connected.
SetWeatherStation.py -r 5
Jim Easterbrook
>
> While fiddling with mySQL synchronisation, I realised that my data had
> stopped this morning.
> Restart of life logger did not help - it simply seemed frozen, despite
> the -vvv option, albeit 0% CPU usage after startup.
>
> Same picture with TestWeatherStation.py
> Unplugging / replugging USB did not help.
> dmesg looks like correct unloading/loading of interface
>
> While plugging / unplugging and restarting life logger or
> TestWeatherStation.py, I sometimes got messages "ressource already in
> use", but this may have only been timing issues.
> Waiting several seconds before the next action made "ressource in use"
> disappear.
This suggests that live logging was still running in the background.
Only one program at a time can access the weather station USB device.
> The only thing that helped was to remove batteries from the station.
> It now reads again for a hour or so, but the data since the hang were
> lost.
This sounds like the problem I have with one of my weather stations.
(Some users of Cumulus (another weather station software) have reported
the same problem.) I've yet to discover a cure - the intermittent nature
of the problem makes it very hard to test any changes. Sometimes it
happens just two days after resetting the base station, sometimes after
3 weeks. Another symptom is that the console 'memory' function no longer
works - it bleeps but won't show any past data.
I'm now running Hourly.py instead (hourly cron job) as that will mean
less USB activity, which I hope will be more reliable.
> In the first instance, it looks like a problem with the station
> hardware / firmware.
> But can it also blamed to a compatibility issue with software?
> That pywws somehow screws up the interface?
My old station didn't show the problem in over 3 years use with pywws,
so I think there has been a firmware change in the weather stations. But
I can't prove that.
--
Jim Easterbrook <http://www.jim-easterbrook.me.uk/>
wro...@tirnet.de
> > The only thing that helped was to remove batteries from the station.
> > It now reads again for a hour or so, but the data since the hang were
> > lost.
>
> This sounds like the problem I have with one of my weather stations.
> (Some users of Cumulus (another weather station software) have reported
> the same problem.)
carry just different labels? Or do they only share some chipset,
firmware or even only standards?
Any idea on the inner schematics?
> I've yet to discover a cure - the intermittent nature
> of the problem makes it very hard to test any changes. Sometimes it
> happens just two days after resetting the base station, sometimes after
> 3 weeks.
Do you know of reports whether this problems also occurs on the
original windows setup?
Hm. Intermittend error smells like power, thermal or EMC issue.
I just discovered a external power jack on the station, asking for 6V
(sic!).
It's now running on AC. Let's watch and hope...
> I'm now running Hourly.py instead (hourly cron job) as that will mean
> less USB activity, which I hope will be more reliable.
> My old station didn't show the problem in over 3 years use with pywws,
> so I think there has been a firmware change in the weather stations. But
> I can't prove that.
Should not read that many conspiracy sites....
Did the old station run on 2 Batteries, too? Or on AC, or higher
voltage?
I just recoverd the post I had rememberd, referring power cycling via
USB.
They were using it for UMTS-sticks, some of which also show instable
behaviour in long term operation
Here is a German translation for the Russian instruction:
http://www.ip-phone-forum.de/showthread.php?p=1603542#post1603542
Essence:
2. get http://www.gniibe.org/software/hub-ctrl.c
4.2. switch off port 1
./hub-ctrl -b 001 -d 004 -P 1 -p 0
4.3. switch on port 1
./hub-ctrl -b 001 -d 004 -P 1 -p 1
of course, that' a really crude workaround.
And not all hardware is supporting this. Playground ahead...
yours Wolfgang Rosner
Jim Easterbrook
>
> Do you know whether all station come from the same manufacturer and
> carry just different labels? Or do they only share some chipset,
> firmware or even only standards?
As far as I know they (those that use 'EasyWeather' software) are all
made by the same company: http://www.foshk.com/Weather_Professional/
> Any idea on the inner schematics?
No. I've dismantled mine and it looks like those shown here:
http://sandaysoft.com/forum/viewtopic.php?f=13&t=640
>> I've yet to discover a cure - the intermittent nature
>> of the problem makes it very hard to test any changes. Sometimes it
>> happens just two days after resetting the base station, sometimes after
>> 3 weeks.
>
> So there are little chances for reliable use?
> Do you know of reports whether this problems also occurs on the
> original windows setup?
I don't know, but I think it would. Very few people would run the
Windows / EasyWeather setup continuously though, so they may not have
found the problem. I've use a USB snooper program to check how
EasyWeather works, and pywws is doing exactly the commands. It's
possible that pywws is reading data too often, or by chance it's reading
data at the wrong time and upsetting the weather station somehow.
> Hm. Intermittend error smells like power, thermal or EMC issue.
Or timing?
> I just discovered a external power jack on the station, asking for 6V
> (sic!).
> It's now running on AC. Let's watch and hope...
If it is a power problem then we have an easy solution.
>> I'm now running Hourly.py instead (hourly cron job) as that will mean
>> less USB activity, which I hope will be more reliable.
>
> Is this proven to help? Or still in trial stage?
I've only been doing it for about a week so far, so cannot say.
> Did the old station run on 2 Batteries, too? Or on AC, or higher
> voltage?
In both cases I'm running on internal batteries / USB power. The
stations have a relay that should switch over to USB power when it's
available.
Ger
> life logger still on open console (startup script on todo list)
> python pywws-11.10_r429/LiveLog.py -vvv data
>
> WUnderground had worked till just a minute ago but is gone now ("Data
> not available").
Christian Chaney
I've got the same hardware from Conrad and the same problems...
I have checked with batteries, USB power supply and / or 6V power
supply...
Nothing to get it working more than 3-4 days. :-(
I think that is a problem with the USB hardware, maybe the firmware...
But no information about how to change/upgrade it.
Regards,
Christian
> life logger still on open console (startup script on todo list)
> python pywws-11.10_r429/LiveLog.py -vvv data
>
> WUnderground had worked till just a minute ago but is gone now ("Data
> not available").
Wolfgang Rosner
WH 2080 now running for 2 weeks
AC power did not help as expected when still reading in 5 min interval:
station was dead after some days again, when memory indicator was at about
70 %
After repower, I did not set the shorter data collection interval but left it
at the standard 30 min. Still reading data with pywws-11.10_r429/LiveLog.py
Some days ago memory indicator reached 100 % und runs as cyclic buffer now (I
suppose), but reading is still alive.
Barometric pressure is gone, but this may have happened when I tried to power
by AC only, w/o batteries .
Don't want to fix this now, prefer to countiously let it run as long as
possible.
Outdoor sensor connection was gone for some hours, maybe due to worsended
conditions from snow and cold, but is back again.
Of course, this may just be good luck.
I'll keep you informed.
gnx
Wolfgang Rosner
wro...@tirnet.de
http://www.wunderground.com/weatherstation/WXDailyHistory.asp?ID=IBAYERNM13
WH2080 conrad branded, I think it is the same like
http://www.foshk.com/Weather_Professional/WH2080.htm
openwert ATTITUDE ADJUSTMENT (bleeding edge, r29631)
on TP-Link MR3220