High readings when station loses communication with sensor

[steve]

I don't think any of the changes I've done are going to help with those values

Which version/build are you using?

Steve

[Doc D]

1.8.4 build 605.

BTW, I am a complete novice in this malarky, and the set-up is very recent (last week) and still being finalised - hence please accept my comments with due caution. Nonetheless, I thought that the errant readings might be germane to the discussion. FWIW, equivalent errant readings did not seem to occur in Heavy Weather, but I have not used that as much.

Apologies for any confusion stemming from my contribution!

[steve]

No, that's fine, no confusion at all. I am somewhat relieved to see you say you're using 1.8.4, as I've done some changes in 1.8.5 which appear to have solved a lot of the false data problems (other than this latest lost communication business).

I think you will find that installing 1.8.5 will improve things for you - but I think you might as well hold on until the next build before installing it - unless you're really keen

Steve

[Doc D]

Very keen, but even more incompetent - I shall wait for the next biuld, and try to optimise existing siting and set-up in preparation!

Thanks again for your time and effort!

[helresa]

Just to confirm that I am seeing the high readings on my system as I am awaiting a replacement for the external transmitter to arrive from the UK as mine died after 2 years of service

[steve]

Latest build attempts to fix the 'lost communication' data issues:

http://www.nybbles.co.uk/downloads/beta ... sSetup.exe

Steve

[cilldara]

Thanks Steve,

I've downloaded and installed the new version, the station has been behaving itself for the last day or two, so I can't tell if its made an difference, I'll let you know how it goes if and when the problem recurs.


John

[steve]

OK - thanks, John.

Steve

[Doc D]

Hi,

I have installed the new build but am still obtaining aberrant readings. Since these include data measured at the console, I am not sure that reception breakdown is the sole explanation. What I notice (e.g. for pressure readings) is that the wacky data contain elements of the preceding and following readings, with the problem seeminly affecting the digits preceding the final two: e.g.

1007.2
77.2
1007.2

1008.6
88.6
1008.6

Dunno if this is of any use, but in case it is...

Many thanks!

[steve]

The pressure data is returned in 3 bytes of BCD, so this suggests that the least significant byte, representing the 7.2 in your first example, is OK, but the other two, representing x100 (the x is part of the pressure in inHg) are not. But the data is checksummed, so I don't understand what could be going wrong. The checksum isn't very sophisticated - add the bytes and take the least significant byte.

I'm not actually logging the entire packet to the debug log, just the data, but could you turn on the debug log and let me have it when you see the problem again? And I'll look into tracing the entire packet including the checksum.

Steve

[Doc D]

Thanks! Will do!

[cilldara]

Hi,

I have installed the new build but am still obtaining aberrant readings. Since these include data measured at the console, I am not sure that reception breakdown is the sole explanation. What I notice (e.g. for pressure readings) is that the wacky data contain elements of the preceding and following readings, with the problem seeminly affecting the digits preceding the final two: e.g.

1007.2
77.2
1007.2

1008.6
88.6
1008.6

Dunno if this is of any use, but in case it is...

Many thanks!

I'm seeing something similar, strange single readings which don't recur, if you look at yesterdays high pressure reading you can see what I mean, at about 8.50pm

http://weather.kilmoneynet.com/weather/yesterday.htm

I've uploaded my debug.log to http://www.kilmoneynet.com/files/debug.log, it has got quite big 7.5MB so hopefully you will be able to see what's going on.


John

[steve]

You seem to have had a temperature glitch just now also

I've found that high pressure reading in the log and the values agree, and there is no checksum error. I noticed there was a checksum error for another pressure reading, so the checksum code does appear to be working.

Could you please install this: http://www.nybbles.co.uk/downloads/test ... sSetup.exe - it logs the checksums received and calculated.

You could delete the debug.log as it's getting so big.

Steve

[steve]

I have a theory about what may be happening. The checksums are correct so the data is correct. However, it's not the data I've asked for. The requests from Cumulus are (occasionally) getting garbled and the station is sending data from an address other than the one I've asked for. Each byte is requested separately, so this explains why some of the data bytes making up the pressure (etc) are correct, and some are not.

I'll see if there's something I can do about this.

Steve

[steve]

Having looked at the code, I see that I am already verifying that the station has understood what I've asked it. So I'm baffled, basically.

I've added some more tracing to the debug log, and some more filtering of bad data to the latest beta. I also spotted a few problems if you have 'calculate dewpoint' or 'calculate windchill' set, where I was confusing units, and hopefully fixed these. There's a chance that in fixing that I have introduced other bugs, so keep an eye on your readings for anything odd.

http://www.nybbles.co.uk/downloads/beta ... sSetup.exe

Steve