Cheap project anemometer

[laika]

I have an idea for a project which involves network-connected transducers. Not wanting to stick extra holes in the hull or cut the wires of my existing transducers I'd like to acquire a cheap anemometer for a proof of concept (not one of those handheld things: something which does speed and direction). Output can be analogue or (if in a format I can read and understand) digital: it will be fed into a development board of some type, translated and fed out over IP.

I have no real experience with anemometers and docs for stuff I've been looking at on ebay don't necessarily tell me much about how data are output from the transducer to a (superfluous for my needs) supplied display.

There must be a few people on here with hobby meteorology set-ups. Can anyone make a recommendation? Cheap is good (I might end up partially butchering it) doesn't need to be high quality or especially accurate (this is just PoC), "readily available on ebay" also good.

 

[lw395]

Windspeed is no problem, it's just so many pulses per second= so many mph/knots etc.
Wind direction is more of a problem, lots of 'giftware' units giving direction as 'digital' to the nearest compass point or worse.

Proper 'yacht' units often have either an NMEA output or a sine/cos or three phase output.
For proof of concept, you could emulate the speed transducer with a pulse generator.
The direction is harder, you need to make some choices.

 

[laika]

It does need to be a physical Thing rather than an emulation: The signal-to-data translation is the least interesting part of the project. The significant part is the nature of network connectivity and the the part which gives it interest value is the implementation using a physical Thing (nobody really cares about the Internet of Virtual Things).

NMEA-0183 output I can live with, N2K not. Analogue is fine but I neither own an oscilloscope nor have more than basic electronics skills, so a documented output I can translate to digital without too much detective work would be good. If NMEA-0183 or similar digital output, one update every 5 secs is about minimum, 1Hz or better ideal.

 

[lw395]

Friction free angle sensors are notably hard to come by at low price.
Particularly accurate ones.
I wondered about rotating a magnet over a Honeywell compass IC.
Optical encoders are available, look for 'shaft encoder' or 'rotary encoder'.

 

[jneale]

Ah the memories. I used to build my own instruments many years ago in the 70s including making my own sensors

For a wind speed sensor - look at maplin.co.uk/p/maplin-replacement-anemometer-for-n96gy-n09qr
I suspect it might give a digial output, if not then stick a small magnet to the underside of the spider to just miss a magnetic transducer stuck on that. Any fuel flow, speedometer transducer should be ok, but the smaller the better. If in doubt make your own with a coild of wire around a nail.
I used to use an old jaguar rev counter meter for the display as all the elecky bits were built in

Wind direction is a bit more awkward. I think an early Nasa product used a vane wih a mgnet which raised a metal plate to bridge a couple of contacts across the circumference. it was a messy afair and only accurate to 10 - 15 degrees.
I built one with a 360 degree low friction 50k potentiometer giving an analogue output which was quite good.

Have fun

 

[lw395]

There are some nice mag encoder chips around now.
About USD9.50, from Digikey.
You might find it cheaper.
http://www.digikey.com/product-search/en?mpart=AS5043-ASST&vendor=961
Resolution a fraction of a degree, just by rotating a magnet above a chip!

 

[lw395]

http://www.madscientisthut.com/Shop... / Robotic Sensors&user4=Rotary Encoder&xm=on

$20 for assembled pcb.

 

[laika]

Many thanks for all this. I should have emphasised that my abilities descend quite rapidly from systems programming (not too shabby) down through electronics (pretty basic) to somewhere not quite at fabrication (hopeless+no workshop). As the thrust of the project was the networking side I was hoping to acquire something that was pre-built with just an output to decode but i have been investigating all the recommendations and will try to build if I can't reasonably buy: more to the point this is all good info for future projects too.

fwiw commercial wind direction indicators (other than super-cheap very low res ones) that I found seem to use a potentiometer as jneale suggested.

 

[lw395]

.....
fwiw commercial wind direction indicators (other than super-cheap very low res ones) that I found seem to use a potentiometer as jneale suggested.

Do you have links for any of those?
The only off the shelf one I've had apart was a B&G mhu, which was a magnetic job with (IIRC) a ferrite moving in a load of coils, with some electronics which I never worked out. I just fixed the corroded bits...

I've heard of others working like an AC servo 'synchro'.

You might want a big old windwane to turn a pot in F1.

 

[lw395]

I've assuming the magnetic ones need zero torque to turn the magnet, could that be a dodgy assumption?

 

[laika]

Do you have links for any of those?

Here's the Davis one:
http://www.lexingtonwx.com/anemometer/

 

[lw395]

Here's the Davis one:
http://www.lexingtonwx.com/anemometer/

Nasty.
http://www.mouser.com/ds/2/54/6539-346704.pdf

340degree electrical angle.
I suppose that might be OK if you have a wind direction that never happens, like dead ahead on a sailing boat ....
But most of us like the instruments not to have a 20 degree dead zone.
Also, not particularly cheap, around £10 in small quantities plus VAT.

Mechanical life seems quite high at 10 million shaft revs, but I'd guess that's not many years of fretting back and forth between wsw and ssw on my roof?