Temperature data spikes revisited

[hans]

nice job,
looks like your patience and determination has paid off.
good move to solder the cable directly,that eliminates friction and signalloss
(and false signals).
all in all a nice "project".
1-wire sounds interresting too,offers lot's of possibilities to extend with sensors etc .lightning detection is interresting for me

[Gina]

Thank you

When it got a few degrees above zero today and with bright warm sunshine I decided to do the whole job - mast down, swapped wind vane unit, taped new cable to mast and put mast back up again. Then coiled up the spare cable in the Stevenson screen box and plugged the wind sensors back in. Weather station should now be back in operation with the new CAT5e UTP cable with proper termination both ends. Now to see it it's worked!

[Synewave]

Gina,

You are amazing (maybe a bit mad, going outside in this weather!) Well done on this break through.

[Gina]

Thank you Paul Oh yes, definitely a bit mad

[Gina]

Both consoles/systems set to receive the outdoor sensors so the test system, still set to 48 seconds logging has more chance of seeing any data spike that occurs. Been watching things closely with interest and trepidation... Have I really cured the spikes???

It was looking good for a few hours but NO - Not quite!!! AAaaarrrrggghhhhh.....

test-Temp-only7.png

I'm using all the newer kit now. The reappearance of a couple of (quite small) spikes corresponds with an increase in wind speed. I could only manage to get 10 mph out of my desk fan and now the real wind speed is peaking above that.

Unless I (or someone else) get any other ideas, this is it. Hardware testing is finished. I'm moving on to other things. Enough is enough!! The exercise has been far from a failure though - the spikes are much smaller and far less frequent and I've learnt a lot about the workings of Fine Offset weather stations.

To summarise my conclusions:-
1. The spike problem has been abated if not totally eliminated
2. A lot has been learnt about F.O. stations
3. The source of spikes is the anemometer in conjunction with the characteristics of the cable used to connect the wind sensors.
4. Screened cable seems worse than unscreened and properly terminated CAT5e UTP cable seems to help.

Eliminating remaining data errors from logged data will be tackled in software - as Steve has done in Cumulus.

[werribee_au]

I have been following this thread with great interest.

You, Gina, have not only learnt a lot about the FO stations, but by sharing your findings you have taught a us much about these stations and how they work.

Well done and Thank You

[Gina]

Thank you Gerard I'm very pleased to have helped de-mystify the Fine Offset station workings

[hans]

understanding you are tired of it,it is still naging me a bit.
so i am still searching and found a pdf off a reedsensor with some possibilities for protection against current spikes
http://www.elhinel.com.ar/camozzi-pdf/CSBCSC.pdf

you have done all except one,that is using a varistor.
(then one would be really out of posibilities)

it also shows you were spot on with the inductors of 1-2 mH.(cable longer than 10 meters)

as i understand correctly you now have the inductors combined with
an rc-circuit and still leaves some spikes.
maybe joined with a varistor it would kill off the last ones.

a very very simple one would be a bipolar bicolor led(or 2 ordinairy reverse connected) with a serial resistor set to around 3 volts ,connected to the anemometer pulse line connection and the battery negative on the pcb.
they would absorb peaks over 3 volts.

also the internal resistance of the batteries can influence these spikes.
a resistor of appr. 0.1M in series with the batteryplus might help

[beteljuice]

Going totally sideways ......

Is it plausable that the increased rotation (wind) speed is generating static ?

[hans]

yes,that still could be

[Gina]

understanding you are tired of it,it is still naging me a bit.
so i am still searching and found a pdf off a reedsensor with some possibilities for protection against current spikes
http://www.elhinel.com.ar/camozzi-pdf/CSBCSC.pdf

That's interesting

you have done all except one,that is using a varistor.
(then one would be really out of posibilities)

Yes, and I can't find any low enough voltage ones. I could try clipping the overshoot with a diode, I guess

it also shows you were spot on with the inductors of 1-2 mH.(cable longer than 10 meters)

as i understand correctly you now have the inductors combined with
an rc-circuit and still leaves some spikes.
maybe joined with a varistor it would kill off the last ones.

No, I found an inductor didn't help though that was not combined with the RC

a very very simple one would be a bipolar bicolor led(or 2 ordinairy reverse connected) with a serial resistor set to around 3 volts ,connected to the anemometer pulse line connection and the battery negative on the pcb.
they would absorb peaks over 3 volts.

Yes, see my reply above

also the internal resistance of the batteries can influence these spikes.

Yes, I have a 220uF electrolytic and 470nF ceramic capacitor across the battery terminals.

a resistor of appr. 0.1M in series with the batteryplus might help

Sorry, I don't understand what you mean there

[hans]

just a resistor between the battery and the pcb.

also been brewing of separating the anemometer reed from the circuit .
here is the idea;a galvanised separator
would it be possible to have an optocoupler between the reed and the pcb.
the reed would need its own power(maybe a AAA-penlight)

[Gina]

Going totally sideways ......

Is it plausable that the increased rotation (wind) speed is generating static ?

I don't know - the mast is earthed. But there is a foot or so of flat telephone cable (not twisted pair) from the anemometer to the wind vane unit.

I still have the other sensors and tx unit that I had originally, so further experiments are quite possible. I may be able to simulate higher wind speeds by using a coil run from a variable frequency oscillator to produce a pulsating magnetic field.

OK so this may not be the end of testing - just a break from it

[Gina]

just a resistor between the battery and the pcb.

also been brewing of separating the anemometer reed from the circuit .
here is the idea;a galvanised separator
would it be possible to have an optocoupler between the reed and the pcb.
the reed would need its own power(maybe a AAA-penlight)

Yes, I've been thinking along such lines myself, but shying away from too much complication.

[Gina]

Curiously, the spikes don't quite correspond with the highest wind speeds.

Screenshot-9.png