Cumulus Support

I am getting occasional glitches in the data - meaning extremes are totally inaccurate. One day I had a reading of 853mm rain per hour! Recently I got a temperature reading of 25.1'c, which it hasn't been in the UK for quite a while! I have reconnnected EasyWeather, it downloaded the data for that period and it doesn't seem to show any such weirdness. Is this related to the fact that I get "Lost Sensor Contact" as an error quite regularly (twice a week maybe)?

Ian

iwilliamsd wrote:I am getting occasional glitches in the data - meaning extremes are totally inaccurate. One day I had a reading of 853mm rain per hour!

Rainfall glitches are quite common with these stations and can have many causes. If you upload the monthly log from when this occurred I can see what happened to the rain counter. Sometimes it just goes up for various reasons and stays there, but sometimes it can temporarily appear to go up and then go down again - or vice versa - but by then it's too late, Cumulus has already added the rainfall. I've put some checks in 1.8.9 for this - which version are you using?

Recently I got a temperature reading of 25.1'c,

There's a setting in the cumulus.ini file called EWTempDiff which can help if the temperature reading suddenly jumps and then goes back down again. See the wiki for how to use it.

I have reconnnected EasyWeather, it downloaded the data for that period and it doesn't seem to show any such weirdness.

there's a difference between live data and what gets recorded in the station logger, which are just periodic snapshots of the data. If the glitch is temporary, then it may well not get logged, so running Easyweather after the fact doesn't really prove anything (except perhaps it suggests that the rain counter glitch was temporary).

Is this related to the fact that I get "Lost Sensor Contact" as an error quite regularly (twice a week maybe)?

Possibly. If the signal is borderline, then it's possible that data transmissions are being corrupted. I don't know anything about the radio protocol and whether it has any error checking, but the data sent from the console to the PC has no checksums or anything, and bits do get flipped occasionally. You may well be getting many more bad readings than the ones you are seeing, Cumulus filters the more obvious garbage out.

Hi Ian,

I had a similar problem recently. I can't help you with the reason why it happens, but Steve and daj gave me advice on what to do to correct the glitches when they happen - see here for the thread:
https://cumulus.hosiene.co.uk/viewtopic.php?f=4&t=2314

I was getting glitches in my outdoor readings which became much worse when I moved the transmitter/temp/humidity unit lower. But since moving the console higher up and in line-of-sight through a window, all my glitches have disappeared. I don't think the range on these units is anywhere as good as they make out! Beware of anything that conducts electricity in the transmission path. And also anything that produces electrical interference - especially electric fences! They can produce a bucketfull of rain a second without a drop of water in sight!!

iwilliamsd wrote:I am getting occasional glitches in the data - meaning extremes are totally inaccurate. One day I had a reading of 853mm rain per hour!

You should be so lucky! I've got a WH-1091 and Cumulus 1.8.8 and I had a rain reading of over 15000m per hour. Just today, it's reporting 7074mm per hour.

I'm resetting the records every week or so.

What's strange is there are no files under C:/Program Files/Cumulus/Data

Is that normal?

qrqeather wrote:What's strange is there are no files under C:/Program Files/Cumulus/Data

Is that normal?

On Vista and Windows 7 it is: http://wiki.sandaysoft.com/a/FAQ#I_can. ... a_files.21

Hi All
I think it is the signal strength that causes all sorts of problems and that seems to be related to the supply voltage of the batteries. I have powered mine with a mains 3.2V power supply and everything seems much more accurate. The only time I had problems was when I was running a greenhouse fan heater on an extension cable that went past the transmitter. So my guess is that the unit is very susceptible to interference and gets worse if the battery voltage drops due to cold or just age. The phantom rain amounts are sometimes so great it cannot always be the bucket tipping - it would need to be literally buzzing.
Regards
Alan

apenwith wrote:The phantom rain amounts are sometimes so great it cannot always be the bucket tipping - it would need to be literally buzzing.

Yes, I agree. I'm sure the huge rainfall amounts are down to some sort of interference/corruption; a few mm when it's windy is probably down to the bucket tipping.

in response to the original request re 'bad data'

i run Easyweather and Cumulus software BUT NOT at the same time.

This gives 2 data sets so corrections are easy.

It works for me.

As for battery issues, I am still on the same TX set as when i started in June 2008. They are Duracell and seem good so far. Quite surprised to find NO issues with recent -10C temps. I will change them in June anyway but suspect that the drain is well within a longer life. I havent looked at recharging yet but its well covered in the Forum.

As for the radio (RF) side of things, 'line-of-site' will always be best at this frequency (UHF) so unobstructed and high is the order.

Lastly, if Easyweather is reading data OK then Cumulus should.

philcdav wrote:Lastly, if Easyweather is reading data OK then Cumulus should.

Except that you can't run them both at the same time and they operate in completely different ways, so that temporary glitches are almost certain to be logged by Cumulus, and quite likely to be not logged by Easyweather.

As I see it, there is no way Cumulus could add glitches on it's own. The glitches must be recorded by the weather station itself. Reading data from the console is very unlikely to err - the USB connections are unlikely to cause any problem (unless a plug is almost hanging out). Interference or lack of radio signal will be the cause of glitches in my experience.

Gina wrote:the USB connections are unlikely to cause any problem (unless a plug is almost hanging out).

I know that some people use USB cables with ferrite chokes (is that the right word?) which they believe improve the connection between the PC and the console. Whether it really works or not I have no idea; I'm a software man and have spent my adult life working around problems caused by hardware. Why I'm doing it now as a hobby as well is hard to comprehend, really

The ferrites round the cables do not affect the data transfer in any way but they help to reduce radiated radio interference from the cable. Practically all computer or other digital data cables have these ferrite rings. Digital equipment produces a lot of electromagnetic radiation which can interfere with other electronic or radio equipment. If you put an ordinary transistor radio near a computer or it's cables you'll hear the buzz of digital data. A lot of effort has gone into reducing this interference as much as possible and there are international standards that need to be satisfied in this respect. You can see the labels that confirm compliance with these standards.