Temperature data spikes revisited

[Gina]

I think the batteries powering the opto-isolator have gone flat. The anemometer output started going and now it's gone completely. That's two AA alkaline cells. I'll go out shortly and change them. Coffee break ATM.

[Gina]

Oh dear, no joy Two brand new Duracells and still no wind speed! I did find it very wet from condensation inside the Stevenson screen so it's probably corrosion. We have had several days of fog, some of it with no breeze, and the humidity has been at maximum for 2 or 3 days. Everything outdoors is dripping with condensation - not the ideal environment for electronics! I'm going to have to coat or pot any new electronic stuff I put outdoors, even if it is covered.

[Gina]

I poked it about, rubbed the battery and battery carrier contacts etc. and it's fine now. It hadn't flattened the batteries in a week or so - they show full voltage.

[serowe]

Something else to consider - are the batteries working ALL the time - especially when the temperature gets really low? IOW are they within their stated operating temperature limits?

[mikew]

Hi all. I have had a lot of trouble with temp. spikes however one of the rechargeable alkaline batteries supplied with my unit (has a solar panel) leaked and the transmitter stopped working. Luckily I was supplied two transmitters so I have swapped them over and straight away the spikes have gone (I hope I am not tempting fate by writing this). The spikes used to be ±10°C, now nothing. I did notice before that when the wind was blowing the spikes were worse but today we have a 20 - 30 km/h wind and no spikes. I'm not sure if this may help with the diagnosis of this problem or not

[Gina]

There are several causes of the spikes. Low battery voltage can cause them but also if you extend the cable from wind sensors to transmitter unit, the anemometer can cause spikes. The cure I found was to isolate the anemometer signal with an opto-coupler (see earlier posts). I still have a problem with interference from the electric fence despite doing everything possible to improve the fencer. The fencer unit itself is a long way from the weather station and has a really good earth. I have given up trying to stop this source of interference and will remove the spikes in the software.

Cumulus can already do this and now that I have finished using my second console for testing I have set up the beta version of Cumulus on my XP machine and will be testing that. Cumulus is much more advanced than my Linux hosted software - I'll be using and testing both while also developing mine further.

[m0clx]

Hi all
just for info - have erratic-ish temp readings too - and tried the 'snubber' 100ohm resistor and 0,5µF capacitor across the anemometer read switch and the temp readings are much smoother than they were.
Am going to put together a better sun screen and shield the cable inside the support post (elec fence nearby) just to be sure.
really appreciate all the discussions and suggestions on the forum
rgds
George

[Gina]

I'm planning to replace my mast sometime and run the cable down inside too. Our electric fence is currently away from it and only causing the occasional spike. I think the wind gauge cable was picking up interference too as the spikes are less since I accidentally pulled the cable out of the plug! Just one of my many maintenance jobs around the place. The anemometer feed has failed too, another thing to sort out. But not in this dreadful weather!!!

[Super-T]

I get the occasional .3 up to .5 degrees C short spikes and what I have noticed is that it is normally reasonably breezy when these spikes happen. Perhaps the transmitter unit is particularly busy with the wind and may overlap slightly into the temperature "window". Just an observation

[Gina]

Weather station experiments and development have taken a back seat of late to building an observatory. But will continue once the observatory is built. The latter includes a "warm room" which will contain computers and also be used as a workshop. The weather station software will be running on PCs in the warm room in addition to webcams for astrophotography etc.

What I have noticed in clearing the site for the observatory, including moving my F.O. rain gauge, is that the spikes produced by the electric fence disappeared when the rain gauge was unplugged from the transmitter unit. I have now moved the rain gauge to the bank right on the edge of our property and only a couple of feet from next door's electric fence. Now, having plugged the gauge back in, the spikes have come back with a vengeance!

CONCLUSION :- Any cable connected to the transmitter box will pick up interference and cause spikes in the temperature plot. So screening or RFI counter-measures will be required on the rain gauge input too.