Temperature data spikes revisited

[Gina]

The spikes don't seem to tie in with anything. I've checked the earthing and that's still intact. It seems to be the transmitter end as they compare perfectly between consoles.

This plot goes from 1700 on Wednesday 29 Sept 2010 to now - 1640 on Friday 1 October 2010.

48hrs-Temperature1.png

[nking]

It really is bizarre! Is it only on the temperature data? If everything is unplugged from the transmitter you will only get data for temperature and humidity – Have you had spikes with this configuration? With both old and new transmitters?

[Gina]

I tried it before with the old transmitter unit but not yet with the new one. Maybe I'll try that tomorrow. May take more than one day to prove it if there are no spikes though.

[Ned]

Great that you're determined to find out the cause. I'd just apply EWTempdiff in Cumulus to keep the sods under control
Since installing a CAT5 cable I haven't seen any spikes, but am still using EWT just in case

[Gina]

Wind and rain unplugged @ 15:15 BST 2/10/2010

[nking]

Fingers crossed Then if it's OK I guess you'll connect one item at a time. The problem now is how long do you leave it assuming the spikes don't recur.

[Gina]

Yes indeed. Just reconnected the rain gauge as it's started raining (actually @ 19:00). I'll leave the wind sensors disconnected for at least 24 hours. Or longer if it's raining hard!

[Gina]

So far so good...

24hrs-Temperature-only.png

[Gina]

Latest temperature plot - wind and rain sensors both connected up to 15:15 yesterday (2/10/2010) and rain re-connected at 19:00 yesterday.

24hrs-Temperature-only2.png

I decided to re-connect the rain and see if I got any spikes from that (as shown above). It had just started to rain and later it turned very heavy. Here's the rain plot.

24hrs-Rain2.png

Having the wind speed disconnected for so long showed up a bug in my software - I used the total wind over the period to scale the wind roses and had a divide by zero error when this was zero. Bug fixed for development version only ATM. I'll migrate the changes over to the working version shortly.

[Charlie]

Hi Gina,

It seems that whenever you have temperature spikes, I get them too. Perhaps it's the phase of the moon, sunspots, or sympathetic something or other. However, I did notice an interesting correlation for me, yesterday. We’ve had a steady breeze here for several days, and there’s plenty of noise spikes on my temperature line.

However, last evening, the wind died completely giving us about an hour of dead calm before picking up again. During this time, the temperature line was completely smooth. I’m now absolutely convinced that the wind sensor is the root cause of the bulk of the noise.

Unfortunately, changing to shielded cable is going to be tough, so I’m going to try filtering the leads at the transmitter end which is much more accessible. Quick question: The jack that come from the speed and direction sensors – any idea which pin is which? I’m sure I’ve seen this information posted before (I think by you), but I can’t seem to find it now…

[Gina]

The anemometer connects to the middle two pins and the wind vane resistor network and reed switches to the outer two pins. The anemometer closes a reed switch every time the wind buckets do a revolution. The wind vane unit has 8 reed switches which turn on depending on the wind direction. These determine the resistance between the two wires.

Details are in this thread https://cumulus.hosiene.co.uk/viewtopic.php?f=13&t=2685

[Gina]

Still no spikes...

24hrs-Temperature-a.png

[Gina]

42 hours without a spike... I think that'll do

42hrs-Temperature.png

[nking]

Looking good

[Gina]

Just popped out and plugged the wind sensors back in. Wind readings should resume from 11:00 BST 4/10/2010