Cumulus Support

That's an interesting discovery.

I made a poor sort of extension to the weather vane with tape and stiffeners, but it made no difference to my flapping wind vane, which it certainly ought to have done if the vane's size was the cause.

If we need to split the level of the vane and the anemometer, we'd need some kind of end cap for each of the legs if I recall correctly how it went together. It would be pretty easy to make one of those out of aluminium strip - probably, or even a piece of plywood, but that would be less good since you can't bend it to fit around the pole.

EDIT:

Thinking about this, maybe the top beam is in one piece and just fitted down onto the pole. I can't remember really. Must be dementia setting in.

YES - It is a one piece beam - or looks like it is. Much harder to fix then.

How about cutting off the part of the beam holding the anemometer, just next to the upright, inverting it and fixing it to the underside of the beam holding the wind vane? You then have a more 'standard' arrangement.

Before anyone carries out major surgery on the beam, we could test Charlie's theory about the effect of the anemometer on the vane, by tying up the anemometer with an elastic band and then observing for a bit. Cumulus graphs would provide definitive copy of the effect. Except maybe it is set to nullify wind direction during periods of calm.

EvilV wrote: Except maybe it is set to nullify wind direction during periods of calm.

The average direction will come out as zero, but the 'instantaneous' direction is only zero if you choose that option.

steve wrote:

EvilV wrote: Except maybe it is set to nullify wind direction during periods of calm.

The average direction will come out as zero, but the 'instantaneous' direction is only zero if you choose that option.

Thanks Steve - yes I went upstairs and checked it on the machine that runs Cumulus.


RESULT OF TEST:

Ok - I dropped my pole and put an elastic band around the anemometer to stop it.

The vane was flapping about as usual with the cups held still. I gave it ten minutes and then reversed the operation, taking off the tail extension I'd put on. If anything, I think the vane is moving less than it was before and even when the cups were held still.

So - my system is now back in native mode - just as Fine Offset Intended.

I didn't bother to consult the graphs - it was completely obvious just watching it.

steve wrote:How about cutting off the part of the beam holding the anemometer, just next to the upright, inverting it and fixing it to the underside of the beam holding the wind vane? You then have a more 'standard' arrangement.

I haven't looked at the underside of the anemometer lately but wouldn't wet get into the works if the anemometer were turned upside down?

I'm thinking I might take off my extended tail and counterweight. I've just been watching it and it's just done a 360 degree!! I'm thinking it's not really any better with the extra tail! As it is at present you might just as well spin a top with 16 points on it

Gina wrote:I haven't looked at the underside of the anemometer lately but wouldn't wet get into the works if the anemometer were turned upside down?

Yes, it did look that way to me.

Just been outside, taken my pole down and removed the tail extension and counterweight. I think it's actually a bit better now - it certainly doesn't seem any worse.

While I had the pole down I looked at the underside of the anemometer. It has a base plate with lower surface flush with the tubular part (goes inside), fastened with three screws. The base plate has a hole for the cable. Seems to me that it would be quite easy to seal it. And if the bearing is a ball race, it should be happy enough upside down. When we get some more settled weather I might take it down again and dismantle it and see if there's any reason why it should not be inverted.

how about taking it where the problem is ,the windvane.(responding to the anemometer though)
by putting a piece of pipe between that and the T-shaped holder you can raise it to your liking,out of range of the anemometer.
30-50 cm should be enough.
will test it myself,but for now it's rain and wind here.

It isn't the anemometer. I stopped mine with an elastic band and the vane was still flapping about.

https://cumulus.hosiene.co.uk/viewtopic.p ... 976#p20976

Also, if anyone did want to split the wind vane and anemometer, they will need to lengthen the cable that links them. It is far too short to move the two apart as it is now.

Why not do what the low cost pointer instrument makers do, that is put a spot of silicone oil of a suitable viscosity ( the more viscous then greater degree of damping) on the pivot points. if you don't like it you can clean it off with surgical spirit or meth.
Dengle

EvilV - the elastic might have only changed the problem, not removed it (great idea though!). If you removed the anemometer completely and just let it hang for a bit, that would be a better test, I think.
Like Gina, I have watched in light air and seen the vane "twitch" whenever one of the cups goes by.

dme235 - high viscosity oil will help, but it's a pain to get to the bearing to add it. You risk breaking the unit trying to get the pressure fit pieces apart. It's also potentially a problem for those of us with extreme climates. A typical year here will see temperatures swing between -30C and +30C, and I've recorded extremes of -38C and +41C over the years. I can't think of any lubricant that will work consistently over that range.

hans - I like your idea of raising the vane with a short section of pipe. I've got a scrap bit close to the right size maybe 25 or 30 cm long. I'll try that when I have a spare minute.

grease won't help,already done that,no succes.
actually,just finished doing the extention.
only i just gone a little further.
i have raised the vane with a spare pipe i had from the ws3000(same as wh 1081,but no winddirection and no usb)
this pipe is half the length of the pipe from the wh1081(which also is in 2 pieces)
like evilv said you must lengthen the cable,so i took the anemometer from my old ws3000,which cable is 1,5 meter long.
then i also cut off the tail of the vane,(the oval part),left the spine of it intact and attached a small strip of aluminum (12 cm) to it with small screws.
at the end i attached the wing,(looks like a plane tail mirrored)
it does respond to turbulence (it's on a balcony) but it's more steady,not the lunnatic on steroids it was before.
later i can post a pic if someone want to se it,but for now i don't have much time,maybe later in the evening,else tommorow.
still it's now wait and see,if not,we can rule this out too.

Hi Folks. I know I've found the right forum from earlier postings, pictures from inside this wind direction sensor, etc.
I'll start with posing my question, then follow up with some background for those of you who are interested.

What's the simplest way for me to connect this wind direction sensor into my small picaxe computer on a robot boat ? At present I'm testing my picaxe software, which can monitor the analogue voltage on a spare input pin, using a 10k potentiometer, connected across the +5v 0v supply, and with it's slider connected to the input pin.

I've already snipped wires, tested with a meter on ohms, and pulled the unit apart - confirming it is the same (spare unit for a WH1080) as the product discussed here. It seems to me that the forth (yellow?) wire is for when the product is used to detect wind speed. For my use, I assume I'll use the green, black and possibly red wires.

If I can connect it as a "variable resistor", ideally in the ballpark of 0 to 10k, then my software inside the picaxe will need to test for maybe 16 voltage thresholds and translate these to rotational position. Thanks for that earlier posting with resistance vs rotation: N 33k, NNE 6.58k ... W 120k ... NNW 21.9k etc.

From earlier postings on this thread, it seems to me that the simplest solution might be for the green wire to go to my input pin, and for something like a 100k pull up resistor from there to the +5v. I can live with just 22.5 degree accuracy, but it seems it could be twice this - subject to limitations discussed earlier on this thread. I'm advised from the picaxe forum that my resistors at the input side should be in the 10k ballpark, but I've not investigated limitations in the picaxe.

I anticipate having to add some components, even if only the odd resistor. If this cannot be done, for my application, then I could imaging wiring my own resistors into those 8 reed relays - if that were needed. Exploiting this product seems a much quicker way of meeting my need for the simple autopilot on my robot boat to get an approximate direction of wind, relative to boat heading, so it can sail close to the wind. Yes - it is a small yacht. Reliability and simplicity are the key features needed, but more accuracy could be useful. Before finding this product and getting one to test, I had thought I might have to build something similar from scratch - either using reed relays or a hall effect chip.

Now some background: you will find lots of information and an amusing "Snoopy Sails!" video on my hobby robot boat page http://www.gpss.co.uk/autop.htm This is part of a "friendly international competition" to be the first robot boat to sail itself across the Atlantic. Most of those participating are academic outfits, including from UK, France, Switzerland, Portugal, USA, Canada. For me it is simply a hobby project, occupying part of my time for over three years. We started with a Pocket PC based autopilot, but right now I am experimenting with a much simpler and more primitive picaxe processor based solution. It has just three parts: GPS --->picaxe---->rudder servo. We are testing the basic steering and navigation logic now. I then want to add a wind direction sensor - and that's the reason for this posting. I only ordered it yesterday, off e-bay, and it arrived this morning. It looks as if I can make good use of it.

Thanks for the earlier postings, and thanks in advance for any ideas.
Either via this forum or direct to me on gpss@compuserve.com
Robin
http://www.gpss.co.uk

Hi Robin,

With its poor design of magnetic circuit giving small "intermediate" directions (NNE, ENE, etc.) and no rotational damping, the FO sensor is perhaps not a good choice to start a "homebuilt" project. But as you have :

The FO transmitter doesn't use a "proper" A/D converter but a current to frequency (actually resistance to time delay) converter, so the resistor values were chosen (I refrain from using the word designed ) in a kind of logarithmic sequence. However, if you use an A/D converter to monitor the voltage across the FO sensor (just the two wires connected to the "direction" reeds) when it is fed by a 10k ohm resitor from the A/D "reference" voltage, then you should be able to detect the 16 directions. As the "intermediate" directions (with two reeds closed) must always have a lower resistance than those either side, the obvious method is to use a "lookup table" to decode the 16 directions.

I've actually been trying (with some success) to improve the uniformity of the 16 directions by adding small "super" magnets to the rotor and hope to post more details soon. But for now you might be interested in the 'scope waveform that I obtained across a normal sensor when fed via a 10k resistor from a voltage source, with the vane spun at about 1 rev per second. The 8 "major" levels (between the markers) are obvious but the intermediate directions (with two reeds closed) are hardly more than a small "undershoot".

Cheers, Alan.