w8681 / Cumulus misbehaving ! - corrupted data ?

[adrian5750]

HI Folks

I know it's almost certainly the fault of my w8681 - and that Cumulus (1.8.4 b 593 - win xp pro) is only doing what it can with the data recieved.... but I've got a couple of problems...

I don;t know if it's significant - but my PC is set to power-save overnight - this usually works OK.....

Anyway - Cumulus's graphs flatlined since 6pm last night, and, when I restarted Cumuls this morning in order to see if that helped, the program suddenly got the idea that it's seen a peak of rainfall of 25.66 inches per hour (round about the time I restarted the program..?)

Is there a general 'how to unravel data problems' document somewhere ? I see mention of backup files - not sure how you'd go about restoring the data...

Does Cumulus download stored data from the w8681 - or, if communication is disturbed, does that create a flatline display until the next successful communication takes place ?

On a positive note - I 'seem' to have solved the problems with the remote sensor failing - washed the pcb with IPA and then anointed it with WD40 - and, so far, so good ! (I just know I'm going to regret typing that !)

Thanks
Adrian

[rexhugin]

Hi All

Re data corruption........I have had a W-8681 for a month or so now and have got both Easy Weather, (vs. 2 at the mo, am about to install vs. 5), and Cumulus 1.8.4. build 552 installed and do run either on an XP P.C. I am now tending to use Cumulus more than E.W. as I do find the former more informative. I soon discovered, as probably others have that one gets very corrupted data if one runs both programs at the same time even if minimised! I of course do not, (intentionally), run both now together. Occasionally I get a huge shock when I suddenly get huge rainfall amounts, or other corruption, only to find that I have accidentally opened both programs at the same time, or have two or more instances of one of the programs. As soon as I have closed the ‘surplus’ program/s all data does then return to normal. I hope this info helps other ‘new’ owners of this Weather Station who may have done the same as myself.

A general point re this W. S.; I am very happy with it considering its price of less than £70, although of course there are the usual two little niggles re the dimmish display and the outside temperature data which suddenly will give a big rise when the sun shines directly on the ‘transmitter’ unit. I have found that using an Anglepoise desk lamp with a 9w energy saving bulb and directed at the W.S. display, in a position to avoid reflections, makes a big improvement. The display is also propped up at an angle of around 45 degrees so that I see it square on to me as I sit at the PC. Re the outside temperature reading in sunshine I intend to make up some sort of ‘protection’ when I get the time.

I hope some of the above may help others.

rexhugin

[steve]

To 'rewind' from the backup files, just stop Cumulus, choose the backup you want to restore from (the folder names are date/time stamps), copy the contents of the backup folder in the data folder, and start Cumulus again. It will then re-read the stored data from the station from the point of the backup.

When going into power saving mode, Cumulus tries to shut down. If this fails, things will definitely go wrong. Best to shut down manually if possible.

Steve

[harrym1byt]

Just an idea for an improvement on this....

Could a little utility be put in each backup folder, to do the data restore automatically perhaps? It only needs to copy all the files from the current folder, over into Cumulus's data folder.

[harrym1byt]

Just an idea for an improvement on this....

Could a little utility be put in each backup folder, to do the data restore automatically perhaps? It only needs to copy all the files from the current folder, over into Cumulus's data folder.

[steve]

Well, I guess it would be a batch file that looked something like:

Code: Select all

copy /Y *.* ..\..\Data

I suppose I could drop a copy of that into each backup folder...

Steve

[adrian5750]

HI Folks

I wrote a reply to the various comments yesterday - but the forum s/w seems to have binned it for me -
not the first time this has happened...

...anyway

..Thanks for the info on what's where in the backup files....

If I'm understanding how Cumulus works, each time it gets a set of data from the weather station it checks the figures against its 'all-time records' - and updates these records as appropriate....

Trouble comes when you get a corrupted reading - like my rainfall rate of 25.66" / hour that's currently 'stuck' in Cumulus. I've deleted the offending values in the logfile - but what would be really good would be some way of telling Cumulus to go and re-calculate the 'all-time records' from the data files.

I guess the only manual alternative would be to pour all the data into excel /oo and sort the records to find the real 'all-time records' and then manually correct the settings in edit | all-time records. Kind of clunky! <g>

Let's hope the forum deigns to display this message !
Adrian

[harrym1byt]

Looking at the contents of the backup directories, Cumulus now saves copies of the alltime.rec file - so restoring these files back into the data directory will fully 'rewind' Cumulus back to before an error occured. That should make it fairly proof against corruption of the FO's USB data, but if the data is corrupted in the displays memory - it will have to be a manual editing sesssion to resolve it.

[steve]

I wrote a reply to the various comments yesterday - but the forum s/w seems to have binned it for me -
not the first time this has happened...

I suspect that what happened is that someone else had posted while you were typing your reply, and the forum asked you if you wanted to revise your post in the light of this. It's not obvious that it's asking you, you may have thought it had posted it. I still fall over this myself. I'll disable it if i can work out how.

but what would be really good would be some way of telling Cumulus to go and re-calculate the 'all-time records' from the data files.

The problem with that approach is that the all-time records may not actually be in the data files (if it's something like a record high gust, for example). As Harry points out, all the files necessary to rewind to before a bad reading are saved in the backup folder at startup.

Steve

[adrian5750]

Hi again all - Happy Christmas, and all the best for 2009...

It was all going so well <g> - W8681 happily logging data and passing it to Cumulus.....

....until yesterday afternoon @ 1600 hours when everything flatlined again - the display on the weather dstation itself is showing no data from the external sensors..... grrr !

The last time this happened I brought the external electronics inside, gave them a wash down with IPA and then anointed them with WD40 - which seemed to keep things running for a month or so....

I suppose there could be an intermittent problem with the electronics on the 'inside' unit (USB-powered at the moment) - but it's all kind of frustrating...

The fellow who supplied the unit would be happy to take it back - but would return it to the distributors for exchange - and (apparently) if they can't see the fault then they'll simply return it marked 'no fault found'...

So - I guess it's 'AVO and prayer mat' time......

Isn't it annoying when a good design idea is let down by poor Quality Assurance in production ??

Any diagnostic suggestions ? - don;t have anything that rx's at the sort of frequencies the radio link operates at.....

Bah - humbug !! <g>

Adrian - West Cork, Ireland

[goldrush]

Just a thought. How far away from your reciever/display unit is your "transmitter".
When I had an 8681 I found that it would regularly "flat line" if the transmitter was more than about 20 metres from the reciever, or indeed if there were more than 1 wall in between!
Although less susceptable than the LaCrosse "400Mhz band" recievers, the receiver does need to be "swamped" with signal to mask any interference caused by car security transmitters and other fantastic "lifestyle" equipment using the same "860mhz free band" etc.

For some reason the 1081 I now have does have somewhat better range than the 8681.
In fact I have my anemometer sensors just some 60 metres from the transmitter unit (extended the cable with cat5) and the transmitter unit around 15 metres from the display/reciever.
Since doing this it has "never missed a beat" (guess that has tempted fate!!).

[adrian5750]

HI
Thanks for the reply.....

Progress so far today.....
took a test meter to the batteries on the external unit - just a shade over 3v - so no problem there...
Checked the plug-in cable under the wind vane - all ok

Flipped out one of the batteries and reinserted it - got a red led on the remote unit.

Came back inside - no life on the display.....

...after a while - disconnected the internal unit from the USB and then dropped out the batteries.
Put them back in - and (lo and behold!) - the base station started talking to the remote unit and then it looks as if Cumulus is also talking to the base station....

I suppose the units are about 30m apart with a fair chunk of house between them.
There's not a lot 'other stuff' going on - rf-wise - we've only got one other near neighbour, and there's no wi-fi going on.....

..crazy idea - apparently the units are a bit sensitive to misbehaving if the time signal rx is dodgy - and last night (apparently) there were strange things going on with ham radio topband (160m) propagation - so I wonder if there's anything happening there ??

Seems like the Watson system isn't all that 'resilient' to error conditions etc - which is a shame, as it's a good system for the price.... when it's working !

Regards
Adrian (EI5JV - West Cork, Ireland)

[goldrush]

It does sound to me somewhat like the problems I had initially with interference.. in fact it was 1000% worse with the Lacrosse than the Fine Offset.
As a matter of interest, our nearest neighbours are some half a mile away!

(it does not necessarily need to be transmitted interference, for example if the interconnecting leads are free to "flap about" under some circumstances, that can generate sufficient "signal" to mess things up.. again particularly with the Lacrosse)

As I mentioned, now my transmitter is much closer to the reciever it has behaved ok.

Not convinced your problem relates to the time signal as my time signal goes on and off like a whores underwear without effecting it!

If you wish, you can easily extend the cables between the units to put the transmitter closer by using ready made cat5 screened twisted pair cables (normally 8 wires arrange as 4 twisted pairs plus screen and sold as Ethernet cables from computer shops).
If bought ready made, (you can usually buy up to 50meters "off the shelf") they will probably have RJ45 plugs on each end

If you do not wish to cut the existing cables, then a twin RJ45 female connector will accept the smaller RJ11 plug from the sensors without modifcation. Ensure that it is a "crossover" connector.

Similarly, you can use a few inches of RJ11/12 to RJ11/12 male cable ("modem" lead?) to connect the cat5 cable to the transmitter unit.
Doing it this way means you do not invalidate any warranty!


and RJ

[adrian5750]

HI

Thanks for the comments....

I must measure the distance between the remote unit and the base station....
I suppose I could stick the base station out in the Studio (which is a bit closer
to the sensors, and made from timber - so it should be a bit more 'rf-transparent')
The PC network extends out there - and I could see the current data via the website...
(kind of a convoluted way to do things, though!)

The interconnect cables are ty-wrapped securely to the little mast that the sensors are sitting on -
so they shouldn't be flapping about....

I _could_ start extending the cables - in fact I'd toyed with the idea of simply junking the existing electronics and using a little PIC micro to sense temperature, wind speed and wind direction using the existing hardware - and then running 'something' at the PC end to grab the data..... but I'm right out of 'round tuits' at the moment - so it would be good if I could get the existing kit working <g>

Perhaps moving the base station the the Studio might be the best plan....
been having my annual tody-up in the office and there's a spare PC about 5 feet away that could also find its way out there as part of the tidy-up - thus gaining double 'Brownie-points'.... <g>

Thanks
Adrian

[harrym1byt]

Mine works fine most of the time - only a rare bit of data is missed. It is around 40 yds away and receives through a single then a double brick wall, followed by two garage walls. It has wifi (mine plus another six local ones), various wireless remotes, plus amateur transmissions to cope with and we are in the middle of an estate. I really did not expect it to work with all this, but it does.