1-wire counter with reed switch a long way away?

[skyewright]

I've finally got my Inspeed vortex (anemometer) and e-Vane (direction indicator) up on the mast and wires run from the mast to the computers.

I don't have the 1-wire board at the mast end yet, but as well as running FTP through the pipe for the 1-wire system I also ran some 6 core alarm/phone wire so I've temprarily used that to make a direct connection between the vortex and e-Vane and a HB Anemometer Control Board (ACB). The wire run is a bit over 50m (~160'), and avoids proximity to power cables. The ACB is part of a 1-wire system being read using OWFS plus custom software on an ALIX-1D computer. The e-Vane is working fine, but the counter seems to be over reporting. When the anemometer is still (Not common here, but we have unusally calm weather at present) the reading is steady, but when rotating the count is increasing much more that one per rotation (around 20 times too much).

If I switch the vortex reed switch's pair to a different counter board the same over counting effect is observed.

If I attach a simple "bicycle computer" to the vortex reed switch's pair the computer reports exactly what I'd expect (there is an OS WMR928 anemometer on the same mast).

This arrangement isn't what I intend for the long run, and the combination worked fine on the bench so I expect all to be okay once I get a waterproof box fitted at the foot of the mast, but an inquiring mind would like to know if there is any obvious reason for what I'm seeing? Should it be okay to attach a counter board to a switch so far away?

David

Edit: I just tried attaching a LabJack U3 counter to the vortex reed switch's pair and that similarly over counted (I could see the anemometer rotating at just under than 1Hz, but the counter was increasing far faster than that).

 

[ericvic]

I've finally got my Inspeed vortex (anemometer) and e-Vane (direction indicator) up on the mast and wires run from the mast to the computers.

I don't have the 1-wire board at the mast end yet, but as well as running FTP through the pipe for the 1-wire system I also ran some 6 core alarm/phone wire so I've temprarily used that to make a direct connection between the vortex and e-Vane and a HB Anemometer Control Board (ACB). The wire run is a bit over 50m (~160'), and avoids proximity to power cables. The ACB is part of a 1-wire system being read using OWFS plus custom software on an ALIX-1D computer. The e-Vane is working fine, but the counter seems to be over reporting. When the anemometer is still (Not common here, but we have unusally calm weather at present) the reading is steady, but when rotating the count is increasing much more that one per rotation (around 20 times too much).

If I switch the vortex reed switch's pair to a different counter board the same over counting effect is observed.

If I attach a simple "bicycle computer" to the vortex reed switch's pair the computer reports exactly what I'd expect (there is an OS WMR928 anemometer on the same mast).

This arrangement isn't what I intend for the long run, and the combination worked fine on the bench so I expect all to be okay once I get a waterproof box fitted at the foot of the mast, but an inquiring mind would like to know if there is any obvious reason for what I'm seeing? Should it be okay to attach a counter board to a switch so far away?

David

Edit: I just tried attaching a LabJack U3 counter to the vortex reed switch's pair and that similarly over counted (I could see the anemometer rotating at just under than 1Hz, but the counter was increasing far faster than that).

160' of wire is like a big antenna and with the inputs to the counter being high impedance it would be easy for it to count some noise on the line.

Eric

 

[skyewright]

160' of wire is like a big antenna and with the inputs to the counter being high impedance it would be easy for it to count some noise on the line.

Thank you.

David

 

[skyewright]

Just rounding this off...

The full 1-wire system is now in place (i.e. the Anemometer Control Board is now at the bottom of the mast, only 10m (~33') from the vortex & e-Vane) and both anemometer and direction are reporting well.

David

 

[ericvic]

Just rounding this off...

The full 1-wire system is now in place (i.e. the Anemometer Control Board is now at the bottom of the mast, only 10m (~33') from the vortex & e-Vane) and both anemometer and direction are reporting well.

David

Good to hear.

 

[skyewright]

The full 1-wire system is now in place (i.e. the Anemometer Control Board is now at the bottom of the mast, only 10m (~33') from the vortex & e-Vane) and both anemometer and direction are reporting well.

I may have spoken too soon...

Generally the new arrangement is working well giving sensible but 'better' results compared to the OS WMR928 that is lower down the same pole. However twice now the anemometer counter has gone suddenly flipped to a behaviour in which it increments at around 10 times the rate attributable to rotation of the anemometer. I'm not talking about the odd spike now and then, once this behaviour starts it hangs around for a while (an hour or more) then just as suddenly normal behaviour resumes. Attached is a graph of 5 minute averages for the vortex and WMR928. The dip at 15:30 represents rebooting the computer to see if that made any difference.

I'm not aware of anything strange happening to the computer or in the environment before during or at the end of the event. Reporting from the e-Vane was fine throughout.

Any ideas for diagnosis or prevention?

PS. I'm aware that the version of the vortex that comes fitted with a serial plug has a capacitor fitted into the plug between the wires (a K5U 105M EC1 according to a guy I know who has one, 1uF +/- 20%?). Could that be of significance?

David

 

[ericvic]

Do you have any radiating antenna on your mast or do any of your neighbors?
33' is still a pretty long cable. Is there anyway you could mount the controller board closer to the Inspeed to keep the cable length to a minimum?
You could try putting a capacitor between the counter input and ground at the controller board to see if that helps.

Eric

 

[skyewright]

Do you have any radiating antenna on your mast

Only a standard OS WMR928 anemometer Tx unit ~6m (19') up the mast. I have other OS sensors too, but none within ~30m(100') of the mast.

or do any of your neighbors?

There are no radio hams in the area if that is what you mean. This is a rural area. Just folk with normal domestic equipment, e.g. DECT phones, and even then the nearest neighbour is ~150m from the mast (and weren't at home when we had the blip yesterday afternoon). We are not in range of any cell phone transmissions (i.e. mobile phones don't work here). There is a telecoms (BT, the national phone network supplier) relay station around a mile away, that is line of sight to the top 20' of the mast, but I believe it uses directional antenna (I've certainly never heard of interference problems from it, even from people closer to it).

33' is still a pretty long cable.

The vortex comes with at least 25', with website options for up to 200', and comments that it has been used at over 1500'.

Is there anyway you could mount the controller board closer to the Inspeed to keep the cable length to a minimum?

I could probably mount a small box at around 8', leaving just the standard 25' of cable. I'd not be keen on going higher, both for accessibility and because part of the appeal of the vortex is its small cross section (at 10m it is likely to be seeing winds of 80mph or higher in the winter).

If you feel that the problem is the vortex to ACB wire acting as an antenna, might it help to replace part or most of that wire with twisted pair? I have available UTP patch or FTP solid. Neither twisted pair cable is outdoor rated, but I could route it down the N side of the mast for minimum exposure to UV.

You could try putting a capacitor between the counter input and ground at the controller board to see if that helps.

It's not something I have to hand, but I could order one if it stood a chance of helping. What's the theory behind its usefulness?

Regards
David

[Edit: It's happening again, and I believe I may have spotted a theme. Each time the probelm has occurred (3 times now) it has coincided with light rain. I hadn't noticed that pattern before because the 1st time was in the middle of the night. This is just a small amount (fraction of a mm) of light rain. There is no obvious sign of electrical activity associated with the rain.
So far as I'm aware all connections are waterproof (the vortex end is as factory supplied, the standard 25' wire is joined to an extension inside a small IP55 box and the joint is sealed with liquid electrical tape, the extension joins to the ACB inside an IP65 box, with the connectors coated in dielectric grease. In both IP boxes the eVane connections are on the same cable and connector blocks as the vortex ones).]
[Second Edit: Rain stopped (0.5mm over ~1hr). False readings stopped shortly after. Could the rain have some sort of effect without actual ingress to connections?]

 

[skyewright]

[Second Edit: Rain stopped (0.5mm over ~1hr). False readings stopped shortly after. Could the rain have some sort of effect without actual ingress to connections?]

Asking around, the idea of 'precipitation static' has been raised as a possible explanation, maybe exacerbated by the fact that the mast is a fibreglass one.
Has anyone else come across that sort of thing?
Cures? Shielded cable? Grounding the anemometer bracket?

David

 

[ericvic]

Asking around, the idea of 'precipitation static' has been raised as a possible explanation, maybe exacerbated by the fact that the mast is a fibreglass one.
Has anyone else come across that sort of thing?
Cures? Shielded cable? Grounding the anemometer bracket?

David

I have heard of 'precipitation static' in relation to aircraft but I am not sure what might help. If static is causing the problem then some way to dissipate it would be the thing to do. Grounding the anemometer bracket may help, shielded cable may help if you ground the shield. Fiberglass can definitely build up a static charge on its skin. Sorry I'm not being much help with this issue so far

Eric

 

[BraveSirRobbin]

Disclaimer: The following advice may be of no help and not applicable to your situation

At work I'm responsible for some very remote monitoring stations that measure various weather paramaters. One of the parameters was a rain gauge tipping bucket that simply provided contact closures whenever a tip (0.01") of rain was registered. The problem I had is this was being recorded via a high speed counter on a data logger.

I had a problem with one station where, during a precipitation event, many counts would get registered instead of just "one". I had a heck of a time troubleshooting this (as it did not happen during "normal" conditions).

Here are some of the things I did/found:

There was a lot of surface laid power cable that was going to a nearby air monitoring station (noisey pumps) and it's grounds were connected to my station. A phone line box was also connected to my station and it also had surface laid cable.

The grounds and shields were improperly wired. Also there was a long (over thirty ft) cable between the station and tipping bucket.


Here is what I did and the combination solved the problem.

I eliminated all grounds going to other systems and removed the telephone box from my station. I replaced the ten gauge ground wire with a large copper cable that went from my station to a ground rod driven in the ground. This was the ONLY ground wire connected to my station.

I ran a shielded cable to the tipping bucket and only grounded that shield at the station end (the shield at the tipping bucket end was not terminated to ground to prevent ground loops).

I installed a 100 ohm resistor in series with the cable (one conductor of the pair) to help with the long run's capacitance.

I think one of two things are happening (or maybe the combination is biting you):

You are getting a static buildup which 'snaps' when the pulse is detected on the anemometer via your recording system. The cable capacitance is making this condition worse. This static discharge looks like a lot of additional pulses to the recording system. (Sort of equivalent to switch bounce). You mentioned in your ealier post that this condition only happened during a rain event, which would also lead me to believe your indeed do have a static problem of some sort.

I don't know if any of this info will help you like I stated above, just wanted to mention some of these items to see if it would trigger a possible cause for your situation.

 

[skyewright]

I have heard of 'precipitation static' in relation to aircraft but I am not sure what might help.

It's apparently a problem in the radio ham community too, especially for those with fibregalss masts...

Sorry I'm not being much help with this issue so far

I'm don't think that either the ACB or the vortex are in themselves in anyway to blame. I just seem to have discovered an interesting corner case! I am hopeful that a solution is just a matter of finding time to apply various grounding and anti-static measures. When we have a solution, we'll know what to suggest if it happens again; and in the meantime I have a pretty effective, if unintentional, rainfall detection system!

I don't expect to be able to do anything till the weekend, but it the weather is okay then my current plan is to lower the mast, then:

i) Run a wire from the anemometer bracket to an earthing rod at the base that I'd already got ready for use with a HB lightning detector.
ii) (Probably) replace most of the vortex to ACB wire run with FTP (foil shielded twisted pair).

For the grounding run I have some "2.5 Twin & Earth" (what we use in the UK for a domestic ring main) stashed up in the loft from when the house was last rewired. That's has 3 good thick conductors - maybe thick enough that I can arrange for one or more cores to reach out and past the anemometer so that it is not itself the high point, though I'm not sure if that would be worth it?

The FTP may be belt and braces, but I figure that it is unlikely to make matters worse, and it would be sad to raise the mast and discover that I then needed to lower it and try something else. If I do that then bonding the FTP shield to the ground would seem like a good idea too, as you suggested.

One other thing I wondered about was if it might be an idea to bond the (metal) spiders & guys to the grounding wire too?

If any of that sounds daft, or if you (or anyone) have other comments I'd be happy to hear.

David

 

[skyewright]

Disclaimer: The following advice may be of no help and not applicable to your situation
<SNIP>
I don't know if any of this info will help you like I stated above, just wanted to mention some of these items to see if it would trigger a possible cause for your situation.

Thank you. Yes I think it may be relevant and helpful.

I hadn't come across the resistor idea (and not being an electronics guy would never have thought of it). I'll hold it in reserve for the present; it would be quite easy to retro-fit later if necessary, as it could at least be done at ground level.

If anyone knows I'd be very interested to know what might be the electronic role of the capacitor that I mentioned earlier as being fitted across the wires on vortex systems that just plug into a serial port (and, I have since discovered, to the vortex / windworks hardware combination).

David

 

[skyewright]

I've put a few turns of the anemometer switch pair around a ferrite ring (which I happed to have available) just before the connection to the ACB. It would be interesting to see if that had an effect on its own, but from the forecast it may be several days till we see rain again. Maybe I'll have the grounding in place too by then.

It has also been suggested that I try adding a low pass filter, e.g. 1K series resistor and a 1 microfarad cap across the pair to cut of anything greater than around 160Hz (40Hz would be 100mph!).

Does that sound like a reasonable measure?

David

 

[skyewright]

Update:
The ferrite ring had little if any effect on its own

Today I added a grounding wire from anemometer bracket to an earthing rod. The effect of that was to make the overcounting full time, i.e. not just when raining.

I then fitted a 1uF capacitor between the anemometer wires (+5 to counter A) and a 1K ohm resistor in series between the anemometer switch and the counter A side of the capacitor.

The aim was to achieve a 160Hz low pass R-C filter. In practice it seems to be filtering very effectively but at around 3Hz!

By the time I reached that stage I needed to pack in. I've left the R-C filter in place.

The filter approach looks like the way to go, but the value obviously need refinement!

Are the connections for the capacitor correct, or should it be across GND and counter A?

Suggestions or observations welcome.