Cumulus Support

Hi Gina,

I have not noticed any timing diagrams from your latest experiments so perhaps you can just mention if you saw changes? I was thinking originally that changing the electrical characteristics would change the timing - perhaps to the point where the procesor runs out of time to make the measurement and returns a strange value.

Also you last comment made me think that perhaps having all the conductors in the same cable might be coupling noise spikes from the anemometer into the vane circuit. Another task for the trusty 'scope! If so, a fairly large cap right out at the anemometer reed switch (or very near by) should quiet that down..

Charlie wrote:I have not noticed any timing diagrams from your latest experiments so perhaps you can just mention if you saw changes? I was thinking originally that changing the electrical characteristics would change the timing - perhaps to the point where the procesor runs out of time to make the measurement and returns a strange value.

No, I didn't see any changes in waveforms at all, except that with the wind vane circuit disconnected the timing timed out at 1 second and the processor shut off the charging when the capacitor was only slightly charged. I'll post some more waveforms shortly.

Also you last comment made me think that perhaps having all the conductors in the same cable might be coupling noise spikes from the anemometer into the vane circuit. Another task for the trusty 'scope! If so, a fairly large cap right out at the anemometer reed switch (or very near by) should quiet that down..

The cable I'm using has two twisted pairs with the pairs screened individually ie. from each other. Although the colours are different, I was careful to use one pair for the anemometer and the other pair for the wind vane. So the pairs should be well shielded from each other. A tinned copper braid then covers the screened pairs and then a grey PVC covering over that. So it should be better than Ethernet CAT5e UPT cable though I have bought a 250ft reel of CAT5e UPT cable for my 1-wire experiments, so I can try a few yards of that (or even the complete coil of it as both ends are accessible). CAT5e cable is cheaper than anything else, probably because of the quantities used. CAT3 cable is quite expensive. But I digress...

Current test setup. Waveform on the "Measurement" line - common connection to timing capacitor. Wind vane pointing west giving 120K resistance. Wind vane pulse width measured 445ms. And this is the resulting temperature plot for this setup so far. (6 hours)

Having got some spikes from the arrangement above, I decided i wanted something more so at 19:30 this evening I connected a 100nF ceramic capacitor across the anemometer (as was suggested) in addition to the RF choke feeding the long cable. Circuit diagram below...Here's result so far...

Overnight run looks good so I'm trying it now without the choke. The large peak after 0830 was when I opened the curtains and sunshine fell on the thermistor.

No spikes for a couple of hours so I'm now looking at the proper outdoor sensors with the 48 second sampling rate. We have a 3-4mph breeze and increasing so I'm expecting to see a spike of two in the next few hours. Then I plan to make use of the fine, sunny weather to lower the mast and add a 100nF capacitor to the anemometer. Actually, I think I'll modify the new one and replace the old anemometer with it - that would be easier.

Here's plot from start of outdoor logging until now. Plenty of spikes!

Soldered 100nF ceramic capacitor onto circuit board of no.2 anemometer and reassembled. Took mast down and swapped anemometers and put mast back up. I had just taken the mast down when down came the snow!!! So I had to change anemometers and put mast back up in a snow storm It's not quite upright but will have to do for now - I'll fix it when we have some better weather.

Still getting spikes May be a little better but not a full and complete cure. Will we ever get one? - I really don't know. I think I'll disassociate the test console/computer system (48 sec logging) from the outdoor sensors and go back to indoor testing. Just a matter of powering up the other transmitter and resetting the console to lock onto that one instead. I have a 250ft coil of cat5e UTP cable instead of the 2 pair screened cable. No extra components in circuit for the initial test. It will be interesting to see what happens Cat5e UTP cable is very cheap

I admire your determination to fix this problem! Hope you can find a solution (in finite time)

Thanks Dave I don't intend spending a lot of time on this. Fortunately, it doesn't use much time doing things, it's mainly just a few minutes fiddling with components then several hours left to itself while the results are logged.

It would be nice to find a solution for cat5e UTP cable as that's very readily available at low cost. And if I can kill the spikes while using the whole 250ft coil of cable, that should satisfy everyone's needs

I'll be looking at terminating the cable at it's 100 ohm characteristic impedance for higher frequencies. I know a bit about transmission lines (long cables) but more with proper line drivers and receivers. There's nothing in the low frequency performance of the cable that could affect the system - it's got to be a high frequency problem of some sort.

09:50 27 Nov 2010 - added 100 ohm resistor in series with 470nF capacitor at sensor end of cable across anemometer.

10:30 27 Nov 2010 - still spikes so added high frequency termination at the other end of the cable as well (across anemometer). Results up to 17:00 Is it premature to say the results look good? Six and a half hours now with the cable properly terminated at both ends.

Been checking up on the waveforms with the new setup. The wind vane ones were just the same but the anemometer quite different :-

Thank you Now i just need to reduce the capacitors otherwise the maximum wind speed will be something like 20-25 mph! I just used what came to hand but I have a full E3 range of ceramic capacitors from 10pF to 220nF, so no problem availability-wise.

I'll see what the working range is. I calculate that the top end would be about 100nF for winds up to 100 mph but I have no way of simulating 100 mph winds.

I'm keeping my fingers crossed Gina - I want to separate the transmitter from the wind sensors too.