Page 1 of 3

High readings when station loses communication with sensor

Posted: Sat 10 Jan 2009 6:42 pm
by cilldara
Hi all,

I've been using Cumulus (thank you) with my WS2300-11 for the last few weeks, and one thing has me puzzled, occasionally the base station loses (wireless) communication with its outdoor sensors. When this happens the station shows a blank (-) reading as does Heavy Weather but Cumulus shows extremely high readings, 69.8C for temp, 184km/hr for wind and 100% for outdoor humidity.

Heavy Weather just records these outages as a blank reading, and joins the nearest dots when making graphs so the outages are bearable there but the high readings in Cumulus don't blend in as well.

I've done a default website install so you can see what I'm referring to,
http://weather.kilmoneynet.com/weather/trends.htm

I've had a look both here and in the configuration settings in the program itself and cant see any solution to this, any help would be appreciated

John

PS, I "think" that the outages in communication might be caused by interference from a DECT cordless phone but am not quite sure, if anyone has any ideas I'd appreciate some advice.

Re: High readings when station loses communication with sensor

Posted: Sat 10 Jan 2009 7:10 pm
by steve
The author of Heavy Weather has me at an advantage as he (presumably) has full details of the protocol. I only have what I've gleaned from the Open2300 project. I'm going through all the required procedures to make sure the station is listening to me, and checking that the checksums of the returned data are valid, so I think the station must be returning those values. I wonder if there are some values which get returned which indicate that communication has been lost... I'll investigate and get back to you.

Steve

Re: High readings when station loses communication with sensor

Posted: Sat 10 Jan 2009 7:19 pm
by steve
I've found a flag which indicates that the wind sensor has been disconnected, but I don't know if that applies in this case. I can't find anything similar for temperature...

Edit: Actually, I've realised that I check that flag already.

Steve

Re: High readings when station loses communication with sensor

Posted: Sat 10 Jan 2009 7:29 pm
by steve
Could you turn on the debug log, please, so I can see what the data looks like in these circumstances?

Thanks,
Steve

Re: High readings when station loses communication with sensor

Posted: Sat 10 Jan 2009 8:55 pm
by cilldara
steve wrote:Could you turn on the debug log, please, so I can see what the data looks like in these circumstances?

Thanks,
Steve

Thanks for you prompt attention Steve,

I've turned debugging on, though sods law says the error probably won't recur for a while and unfortunately this laptop isn't connected to the station 24/7, if its any help the problem is the same whether the data is read in real time or from the datalogger on the station.

Is there any way to have Cumulus start with debugging mode on?

Thanks

John

Re: High readings when station loses communication with sensor

Posted: Sat 10 Jan 2009 11:53 pm
by steve
cilldara wrote: if its any help the problem is the same whether the data is read in real time or from the datalogger on the station.
Thanks, that's always useful to know.
Is there any way to have Cumulus start with debugging mode on?
Yes - edit cumulus.ini and add a line to the [Station] section:

Logging=1

Steve

Re: High readings when station loses communication with sensor

Posted: Sun 11 Jan 2009 12:13 am
by cilldara
steve wrote:
Yes - edit cumulus.ini and add a line to the [Station] section:

Logging=1

Steve

Done, will come back to you when the error recurs.

John

Re: High readings when station loses communication with sensor

Posted: Sun 11 Jan 2009 2:00 pm
by cilldara
OK,

I had an outage again this morning, I tried uploading the debug.log fie as an attachment but it doesn't seem to have worked properly so I stuck it up here:

http://www.kilmoneynet.com/files/debug.log

Hope its of some help,


John

EDIT: There are some strange readings between 10PM and midnight yesterday, one of my kids accidentally pulled apart the USB to serial convertor while Cumulus was running. It doesn't seem to cope well with that.

Re: High readings when station loses communication with sensor

Posted: Mon 12 Jan 2009 10:49 am
by steve
I think I've worked out what values are returned when contact is lost, I'll have a new version shortly which attempts to allow for it.

Steve

Re: High readings when station loses communication with sensor

Posted: Mon 12 Jan 2009 11:46 am
by TgT
huh, cilldara, quite lots of bad data there, you really have it that much away from the main unit? Or have you tried connecting a station with rs232(comport) instead of USB?

My friend gets just temperature spikes around 83°C i think... steve, what is your "hack" now? Does it ignore those readings or any other way? If i may suggestion a special editbox where people can add their ignored numbers ...

Re: High readings when station loses communication with sensor

Posted: Mon 12 Jan 2009 12:07 pm
by steve
It seems that it gives 51.0 m/s for wind speed (confirmed on the Open2300 site), so I'm ignoring 51 or above, 110% for humidity, (so I ignore everything above 100 - of course) and 69.8 for outside temp (which I believe is its 'normal' maximum value), so I'm ignoring everything above 69C.

I was already doing some filtering during normal running, but not for the archive data.

Steve

Re: High readings when station loses communication with sensor

Posted: Mon 12 Jan 2009 12:15 pm
by cilldara
steve wrote:I think I've worked out what values are returned when contact is lost, I'll have a new version shortly which attempts to allow for it.

Steve

Cheers

Re: High readings when station loses communication with sensor

Posted: Mon 12 Jan 2009 12:27 pm
by cilldara
TgT wrote:huh, cilldara, quite lots of bad data there, you really have it that much away from the main unit? Or have you tried connecting a station with rs232(comport) instead of USB?

My friend gets just temperature spikes around 83°C i think... steve, what is your "hack" now? Does it ignore those readings or any other way? If i may suggestion a special editbox where people can add their ignored numbers ...

The outside sensors are about 30feet away from the base unit which should be OK I think, though I'm hoping to move them farther away from the house to a more exposed location.

I don't think its bad data, I only get problems when the base unit is showing a - in the outside reading slots, and Heavy Weather does ignore the gaps. As Steve has said he's working in the dark, trying various things to get it working, I don't think he even has access to a unit.

My original plan was to use an old PC to run weather software 24/7 and act as a file/backup server but it died beyond economic repair so I'm keeping my eye out for something decent but also economical to run 24/7.

Re: High readings when station loses communication with sensor

Posted: Mon 12 Jan 2009 1:33 pm
by beteljuice
Easy fix for La Crosse 23xx reception.

Open the base unit and you see a length of wire wrapped round loosely inside.

That's your receiving antenna, get a soldering iron and replace it with twice the length.

You can still get drop-outs through external batteries / sub-zero temperatures, but it should help stop your unit being so volatile.

You could run cabled (as I do), if the base is 'mains' powered then the outside sensors are powered from the base BUT if there is a nearby lightning strike it can take out the COM port of your PC (don't know about USB / serial adapters)

It happened to me :x

Re: High readings when station loses communication with sensor

Posted: Mon 12 Jan 2009 5:22 pm
by Doc D
As a newcomer to a WS2350 and to Cumulus (great software BTW - many thanks!), I have been following this thread since I have also been receving similar aberrant readings, again (I think) relating to reception fall-outs. The dubious readings so far are not constant, but include:

Max temp > 50C (multiple values in 50-60 range, but clustered around 58)
Mac gusts > 120 km/h (multiple values)
Min pressure as low as 50 hPa (multiple values)

Not sure if any of this helps, but I thought that I might mention it!