Inspeed Vortex & HB Anemometer Control Board odd behaviour when it rains?

In order of preference, no join, good join, , , , , , , bad join.

Conundrum…

I’m halfway through the job. The grounding wire is attached at the top end and strapped to the pole most of the way down. Then I stopped for some lunch. I couldn’t resist taking a look at the figures. Even though it’s dry and calm the graph the vortex was reporting winds of over 100mph (when it wasn’t ‘out of range’). Ah. I thought, that will be because I’ve attached the grounding wire at the top, but not bonded it to the earth rod at the bottom, so I’ve just created an even bigger antenna.

So, I nipped out and attached the grounding wire to the earth rod. Came back, checked the speed. 0 mph. :D. But it seems that the 0mph was just chance. The next reading, and quite a few since have been back to way way too high or ‘out of range’. :frowning: It seems to go in ‘fits and starts’.

20090711122542 wind8 129 0 0 0192
20090711122642 wind8 139 0 0 0189
20090711122743 wind8 140 0 0 0191
20090711122842 wind8 100 586 330 0159
20090711123142 wind8 118 577 293 0165
20090711123342 wind8 144 581 489 0167
20090711123442 wind8 93 581 219 0164
20090711123542 wind8 121 577 313 0162
20090711123642 wind8 125 577 241 0162
20090711123742 wind8 144 0 33 0192
20090711123842 wind8 144 577 101 0164
20090711123942 wind8 119 0 0 0193
20090711124042 wind8 119 0 0 0193
20090711124142 wind8 118 586 444 0165

So do I

i) Leave the grounding wire on, on the basis that it is “A Good Thing”, raise the pole and accept that I’ll have nonesense at least till I get to try fitting a low pass filter.

or

ii) Remove the grounding wire as an idea that didn’t work and at least have reasonable data while it’s not raining (till I get a chance to try out fitting a low pass filter).

Hmmm! At the moment I’m tending towards i)

I’m a little interested in what would be the case if I had the grounding wire up the pole, but NOT attached to the anemometer. Unfortunately it’s so well attached at the top that removing it temporarily is not practical (I’d almost be starting again from when I lowered the mast).

Another thought is that the grounding wire runs along side the anemometer/reed switch wire, but there is not a lot I can do about that.

Decisions, decisions!

Oh well off out not to do a bit more of something…

i)

That makes us unanimous. :smiley:

I’ve just finished replacing the connection between the Inspeed leads and the extension. Previously a connector block in a small IP55 box strapped to the pole with cable ties. Now with soldered joins, protected by heat shrink and liquid electrical tape. I just popped in to check all is okay at the computer end. Off back out now to complete strapping the grounding wire to the mast, then time to raise it again.

w0mbat: I was um-ming and ah-hing a bit about whether or not to beef up the e-Vane. It seems pretty stout, but I can see what you mean about the possibility for vibration and I ended up going your way, attaching a ‘strut’ made of 12mm aluminium angle. A piccy may follow later…

This is all getting very interesting…Your 1st precipitation test is closing rapidly!!!..lets hope all the theories are correct… :slight_smile:

Mast now up, and the Postie has just arrived with a parcel from Maplin. :smiley:
Hmmm. Do I have time to try creating a quick Low Pass Filter…

The anemometer wire connects to “+5V” and a “Counter” connections on the Anemometer Controler Board (ACB).

I know the capacitor goes across, and the resistor goes in series, but I’m not sure which ‘leg’ is the better home for the resistor, and which (R or C) should be closer to the ACB?

My guess is R on the +5 leg and the C closer to the ACB, but that really is a guess!

Off now to furtle in the parcel to see if I can work out which cap is which (it’s a mixed bag).

Having taken another look at the circuit, I opted for R on the “counter” leg and the C closer to the ACB. I finished wiring that up just over 5 minutes ago, and…

20090711171242 wind8 43 558 452 0172
20090711171342 wind8 39 489 442 0174
20090711171442 wind8 30 469 443 0176
20090711171642 wind8 13 586 60 0173
20090711171742 wind8 65 212 18 0185
20090711171842 wind8 34 298 35 0181
20090711171942 wind8 53 484 111 0174
20090711172042 wind8 35 577 157 0173
20090711172143 wind8 42 394 8 0178
20090711172242 wind8 42 16 5 0199
20090711172343 wind8 28 63 9 0199
20090711172442 wind8 48 11 2 0200
20090711172542 wind8 72 14 1 0200
20090711172643 wind8 87 3 0 0198
20090711172742 wind8 59 18 2 0198

(]ISO date time] [sensorId] [direction] [gust x 10] [averag x 10] [windchill].

So there appears to be a very definite effect. :smiley:

Unfortunately, with the present combination too much of an effect (expected wind ~8mph, gust ~15mph, vortex => wind < 1mph, gust ~3mph). :frowning:

Maybe the R should have gone on the other leg?

The current arrangement is very crude, little more than the components plus connector blocks. Sort of electronics meets civil engineering!

Rather than mess further, I’ll leave this combination (R = 1K ohm, C = 1uF, if I’ve read the R correctly ) and see how it responds when the rain arrives.

Back out now to batten down the control box for the expected rain…

The wind is steady and stronger now, and I’m just geting zeros (presumably nothing low enough to pass the filter).

<fx: thinks> I wonder if I’m a factor of 100 out somewhere and I’m actually filtering at 1.6Hz, i.e. 4mph! :lol: :lol:
Capacitor reads 1uF
Resistor Brown Black Red, which I think means 10 x 100 => 1000 ohm?

Cap from counter input to ground (on the board)
Res from the counter input to the wire going to the anemometer.

Brown Black Red = 1K.

Do you have a smaller cap?

Res is wired that way. Cap isn’t (it’s across the counter wires, so from counter input to ‘+5’.
I can try swapping that over next opportunity I get.

Yes, down to 100pF.

Is there any mileage in seeing what is the maximum recorded gust, inferring from that the corner frequency, and making use of that for the next iteration?

So far the highest recorded 3 second gust is 7.4mph (~3Hz), of which there have been several instances. There was an unrepeated higher figure in the 1st couple of minutes, but I might still have been making/breaking connections at that point (I didn’t disconnect the board while fitting the components).

PS. As mentioned earlier the “vortex with a serial plug” appears to come fitted with just a 1uF capacitor across the wires to/from the anemometer.

Let’s get the cap in the right place first, but you can try a smaller one, 0.1 maybe, although I think 1.0 should be OK (esp. if they use that value in the other version). It’s not an electrolytic is it?

Yes, electrolytic.

Of the ones I have 1uF and larger are electrolytic. Smaller than 1uF are polyester film, except the 2 smallest (470pF & 100pF) which are ceramic.

For clarity for others? :wink: See attached piccies for internal connectivity of the capacitor in a “Vortex serial” connection…note that there is an unconnected tail pair that goes nowhere.


Vortex serial.jpg

Vortex serial unconnected tail.jpg

Vortex serial diag.jpg

Probably makes no difference because it’s a 0 -5v pulse but I don’t like to use electrolytics in this kind of application, I would use your biggest non-electrolytic instead.

And I now see that the capacitor goes to pin 5 (‘signal ground’) rather than ‘across’ to the other leg of the anemometer pair. Thank you.

Thank you.

There’s an even chance it might be dry enough this afternoon to try out one or two variations.

Maybe I should start by moving the existing capacitor from +5 to GND, seeing what difference that makes, then looking at changing the capacitor, and if necessary resistor?

I realize that this approach is unlikely to work well because it assumes that the input is actively switching between ground (0) and 5 volts charging and discharging the cap. In reality it is switching between 5 volts and an open circuit so it’s likely the cap will not be fully discharged between hits of 5 volts and the voltage at the counter input will rise to some undefined and non-zero value. Darn, the sort of problem that can be resolved in minutes hands-on with an oscilloscope :frowning:

There are pull-downs on the board so it is switching between +5v and ground on the controller board. When the reed switch in the anemometer is open then the counter input is pulled to ground through a 1 Meg resistor, then when the reed switch closes the counter input is pulled to +5v.

Eric

Aha, there’s skyewright’s 1 Hz theory validated, 1 uF + 1 meg = 1 Second. In that case 1 to 5 nF should do it nicely.

Assuming the 5v isn’t parasitic, intutively I would be inclined to connect 5v through a resistor to counter in, and have the reed from counter in to ground.

BINGO!
I believe we have a result.

In a gap between showers I changed the wiring over so that the cap is now between Counter A and GND.
I also swapped the electrolytic 1uF for a polyester film 100nF, and bumped the resistor (still between anemometer and counter A) up from 1K ohm to 10K ohm to compensate.

Now in wind conditions that are variable but just strong enough to keep everything turning, the Vortex is giving an average over an hour that is just a touch higher than the OS WMR928 3m below it (i.e. a 7m) on the same mast. That seems to me like a good indicator of stability. :smiley:

The Vortex is giving substantially better max gusts, but that is not unexpected (indeed it’s what I hoped for), not least because the Vortex is measuring over 3 seconds and the WMR928 over 14 seconds. The highest gust for the hour is 13.6mph for the vortex and 8.1 mph for the WMR928. I think that if that could in anyway be attributed to false counts (as opposed to better position and shorter evaluation time) then the false count would also affect the average, which isn’t happening.

Naturally I’ll keep monitoring, but I think we have something useful. :smiley: :smiley:

Thanks to all, but most especially to niko who came up with the solution.