Welcome to the Cumulus Support forum.

Latest Cumulus release v1.9.4 (build 1099) - 28 November 2014
Latest Cumulus MX release v3.0.0 build 3047 - 03 February 2019. See the Wiki for download

Homebuilt Wind Vane 1deg resolution

For discussion of DIY weather equipment - sensors, accessories, improvements to existing kit etc
daedalusminos
Posts: 15
Joined: Sun 23 Jan 2011 10:24 pm
Weather Station: TBD
Operating System: Windows XP SP2
Location: Oby, Norfolk

Homebuilt Wind Vane 1deg resolution

Post by daedalusminos » Sun 06 Feb 2011 10:01 am

Work is progressing fairly well on the windvane:
  • Circuit schematic complete (there's actually not too much to it!)
  • ucontroller code 90% complete (interfaces the hall-effect sensor to an I2C bus) - should be testing with the sensor this evening
  • Hardware design for bench testing 75% complete (this won't be weather-proof but otherwise fully functional)
Now, I'd like to eventually be able to interface this to cumulus but I have no idea of what protocols the current supported PWS's use. I believe for the windvane only the cardinal, ordinal and sub-ordinal (16 points total) are resolved, but is the data sent as degrees, or cardinal/ordinal/sub-ordinal points?
Ideally I'd like to send the degree data and have cumulus record (and display) this resolved to the nearest degree, failing that then I can decode degree data to cardinal/ordinal/sub-ordinal points with a resolution of 22.5degrees - not ideal, but atleast far more accurate than the reed switch method.

Jonathan

User avatar
steve
Cumulus Author
Posts: 26717
Joined: Mon 02 Jun 2008 6:49 pm
Weather Station: None
Operating System: None
Location: Vienne, France
Contact:

Re: Homebuilt Wind Vane 1deg resolution

Post by steve » Sun 06 Feb 2011 10:14 am

All of the stations which Cumulus works with send their bearings as degrees, even if they don't use all 360 (or 361) possible values. The Fine Offset stations only ever send one of 16 (if you're lucky) values.

The only way you could currently get data into Cumulus would be to generate an 'easyweather.dat' compatible file, but you would then be limited to 16 values.
Steve

User avatar
steve
Cumulus Author
Posts: 26717
Joined: Mon 02 Jun 2008 6:49 pm
Weather Station: None
Operating System: None
Location: Vienne, France
Contact:

Re: Homebuilt Wind Vane 1deg resolution

Post by steve » Sun 06 Feb 2011 10:19 am

steve wrote:The only way you could currently get data into Cumulus would be to generate an 'easyweather.dat' compatible file
This isn't quite true, of course. You could implement the protocol of an existing station somehow (i.e. emulate a real station)...
Steve

Gina
Posts: 1885
Joined: Sat 21 Feb 2009 12:41 pm
Weather Station: Nothing working ATM - making one
Operating System: OS X, Linux Mint, Win7 & XP
Location: Devon UK

Re: Homebuilt Wind Vane 1deg resolution

Post by Gina » Sun 06 Feb 2011 10:27 am

I've abandoned the reed switch system too - Try as I might, I couldn't get the 8 reed switch method to read 16 directions properly. (And to use 16 reed switches would just move the problem - there would be variable in-between positions.) Also, I found I got different readings depending on which way the vane moved due to the reed switches switching off at a lower magnetic flux than when switching on.

I'm going for optical measurement - 4 LEDs and 4 photo-transistors with a Gray encoded disc, giving 16 clear directions. A higher resolution would be perfectly possible but I think 16 directions (22.5 degrees) is adequate knowing how variable the wind direction is anyway. Even with magnetic damping there is still quite a lot of vane movement (though considerably better than without).

My wind vane unit is almost ready - I just have a bit of tweaking to do. With so many changes of design I've decided not to publish design or photos until I have a fully working setup. That's why I haven't posted much on the subject lately. I'm still using 1-wire as the connection method - it's so simple! There's quite a possibility that Steve will add 1-wire to Cumulus later (I'm hoping he will).
Gina

Sorry, no banner - weather station out of action. Hoping to be up and running with a new home-made one soon.

User avatar
steve
Cumulus Author
Posts: 26717
Joined: Mon 02 Jun 2008 6:49 pm
Weather Station: None
Operating System: None
Location: Vienne, France
Contact:

Re: Homebuilt Wind Vane 1deg resolution

Post by steve » Sun 06 Feb 2011 10:30 am

Gina wrote:There's quite a possibility that Steve will add 1-wire to Cumulus later (I'm hoping he will).
Yes, 1-wire is certainly in the 'eventually' category, rather than 'never'.
Steve

daedalusminos
Posts: 15
Joined: Sun 23 Jan 2011 10:24 pm
Weather Station: TBD
Operating System: Windows XP SP2
Location: Oby, Norfolk

Re: Homebuilt Wind Vane 1deg resolution

Post by daedalusminos » Sun 06 Feb 2011 10:53 am

steve wrote:This isn't quite true, of course. You could implement the protocol of an existing station somehow (i.e. emulate a real station)...
That's eventually what I intend to do, emulate the PC to logger link. As mentioned before (somewhere), it's a shame there isn't a generic station comms protocol.
Gina wrote:I'm going for optical measurement - 4 LEDs and 4 photo-transistors with a Gray encoded disc, giving 16 clear directions.
I'd considered opto, only really ruled it out due to disc-size/sensor-alignment trade off plus I wanted to play with the hall effect device!

From a mechanical point of view, I think it's important to keep the mass near the axle and thus reduce the inertia of the vane (making it more responsive and less likely to 'freewheel')

Gina
Posts: 1885
Joined: Sat 21 Feb 2009 12:41 pm
Weather Station: Nothing working ATM - making one
Operating System: OS X, Linux Mint, Win7 & XP
Location: Devon UK

Re: Homebuilt Wind Vane 1deg resolution

Post by Gina » Sun 06 Feb 2011 11:50 am

I have tried to keep the mass down though I could have used a smaller encoding disc. I might produce a more compact design later - all this is an ongoing experiment in weather sensors and measurement methods. I'm currently using a spoilt CD-R with the coating scraped off to make the transparent bits. The LEDs and photo-transistors are also quite large, being 5mm diameter and 9 & 11 mm high. I have some tiny SMD reflective IR opto chips but they need more LED current and I'm hoping I might be able to use parasitic power from the 1-wire network with the low current LEDs and high sensitivity PTs I'm using. ATM the LED current required is a bit too much but I'm hoping some tweaking will enable me to reduce it.

I thought of Hall effect devices but didn't know where to source them and also thought the implementation would be complicated. I shall be most interested to see your final design. Good work :)
Gina

Sorry, no banner - weather station out of action. Hoping to be up and running with a new home-made one soon.

daedalusminos
Posts: 15
Joined: Sun 23 Jan 2011 10:24 pm
Weather Station: TBD
Operating System: Windows XP SP2
Location: Oby, Norfolk

Re: Homebuilt Wind Vane 1deg resolution

Post by daedalusminos » Sun 06 Feb 2011 12:20 pm

I'd be interested to know if you plan to keep the LED's on permanently and log when a direction change occurs though the phototransistors, or are you planning on sampling the phototransitors every so often and log the direction at each sample interval. If the latter, what sample interval are you proposing to use?

I've always planned to have battery powered sensors (probably with solar top up charging) but I still need to keep current draw to a minimum. The hall sensor I'm using draws approx 15mA whilst measuring so I need to shut this down and power up momentarily to take a reading, but I'm wondering what sample interval is realistic?

Gina
Posts: 1885
Joined: Sat 21 Feb 2009 12:41 pm
Weather Station: Nothing working ATM - making one
Operating System: OS X, Linux Mint, Win7 & XP
Location: Devon UK

Re: Homebuilt Wind Vane 1deg resolution

Post by Gina » Sun 06 Feb 2011 12:48 pm

daedalusminos wrote:I'd be interested to know if you plan to keep the LED's on permanently and log when a direction change occurs though the phototransistors, or are you planning on sampling the phototransitors every so often and log the direction at each sample interval. If the latter, what sample interval are you proposing to use?
I'm planning on having the LEDs on continuously and reading the PTs something like every 3 or 10 seconds. I haven't decided how often to sample. For the anemometer I'm thinking of something like a few seconds to catch gusts but I think the wind direction could be sampled less often - not sure. I might try to build in some sort of averaging for the wind direction in the software, in which case more frequent sampling could be an advantage.
I've always planned to have battery powered sensors (probably with solar top up charging) but I still need to keep current draw to a minimum. The hall sensor I'm using draws approx 15mA whilst measuring so I need to shut this down and power up momentarily to take a reading, but I'm wondering what sample interval is realistic?
Hmm... yes. 15mA is quite a lot for battery power. I have been thinking of the possibility of using intermittent power for the LEDs. I could add a 1-wire PIO chip to switch the LED power - they're well up to the job. Makes the software slightly more complicated, of course, because I'd have to power-up, read data then power-down. Not that difficult and may be the way to go. I could then use 1-wire parasitic power and a large electric capacitor (or a battery).

So many ideas and ways of doing things! :lol:
Gina

Sorry, no banner - weather station out of action. Hoping to be up and running with a new home-made one soon.

User avatar
hans
Posts: 171
Joined: Sun 18 Jan 2009 9:27 pm
Weather Station: wh1080 /ws4000(alecto) XP-sp2
Operating System: Windows XP SP3
Location: leiden,Netherlands
Contact:

Re: Homebuilt Wind Vane 1deg resolution

Post by hans » Mon 07 Feb 2011 6:33 pm

Maybe instead of leds you could "consider optical interrupter switches".
Image
tricky thing is you would need 3 different sizes to read the disk,or make 3 separete disks for each code track.
you could find them in an old vcr,usually used for the tape load
mechanism.
don't know what current they draw,but could be less than normal leds

Gina
Posts: 1885
Joined: Sat 21 Feb 2009 12:41 pm
Weather Station: Nothing working ATM - making one
Operating System: OS X, Linux Mint, Win7 & XP
Location: Devon UK

Re: Homebuilt Wind Vane 1deg resolution

Post by Gina » Mon 07 Feb 2011 8:40 pm

Yes, I looked into that but the ones I found used more current that the "low current, high output" LEDs I'm using. And with the high sensitivity photo-transistors I may be able to lower the LED current even further. To use a single disc I found some photo-reflector devices - very tiny but used too much current. I'm hoping to use parasitic power from the 1-wire network. I'm still experimenting.
Gina

Sorry, no banner - weather station out of action. Hoping to be up and running with a new home-made one soon.

daedalusminos
Posts: 15
Joined: Sun 23 Jan 2011 10:24 pm
Weather Station: TBD
Operating System: Windows XP SP2
Location: Oby, Norfolk

Re: Homebuilt Wind Vane 1deg resolution

Post by daedalusminos » Thu 10 Feb 2011 10:02 pm

Just a quick update on the windvane.

I've got communication with the hall sensor via I2C.

Now working on the mechanics for the bench test model (double bearing) - top bearing plate machined, bottom bearing and sensor mount this weekend hopefully. Still need to source suitable magnet (diametrically magnetised).

Current consumption looks good as the sensor powers up very quickly, I'm estimating 1-2mA current draw with a sample rate of 1/s. This si probably overkill and could be dropped to 1/5s.

Jonathan

gliderguider
Posts: 12
Joined: Tue 20 Oct 2009 8:19 am
Weather Station: ws-2315
Operating System: windows 7 professional
Location: Mount Evelyn, Victoria, Australia

Re: Homebuilt Wind Vane 1deg resolution

Post by gliderguider » Fri 11 Feb 2011 6:22 am

Just a thought.
The windvane on the lacrosse units have a disk with 4 slots for the gray code. The interesting thing that keeps the component count and the current consumption down is that they use a led for each slot, but only a single phototransistor. there is a light pipe with 4 little nubs on it below the slots. When the windvane needs to sense position it strobes the leds in sequence and senses each of them in sequence. The code is a little more complex, but it keeps the count of i/o pins down, the leds are off most of the time, and only one led on at a time for low consumption.

Gina
Posts: 1885
Joined: Sat 21 Feb 2009 12:41 pm
Weather Station: Nothing working ATM - making one
Operating System: OS X, Linux Mint, Win7 & XP
Location: Devon UK

Re: Homebuilt Wind Vane 1deg resolution

Post by Gina » Fri 11 Feb 2011 9:52 am

That's interesting :) I think I'll pulse the LEDs to save current. I thought of a light pipe but couldn't think of any way to make it myself. I was going to do it the other way round - use one LED to feed four photo-transistors. It's best to sample all four slots at the same time to avoid the possibility of the vane moving between measurements. I have no problem reading four lines at the same time using a 4 input chip (DS2450, 1-wire, 4 channel A to D). That part is already built - it's just the mechanics of it I've yet to finish.
Gina

Sorry, no banner - weather station out of action. Hoping to be up and running with a new home-made one soon.

User avatar
Super-T
Posts: 806
Joined: Tue 09 Sep 2008 3:37 am
Weather Station: wh-1081
Operating System: Weather Laptop - Windows 10 Pro
Location: Auckland, New Zealand
Contact:

Re: Homebuilt Wind Vane 1deg resolution

Post by Super-T » Fri 11 Feb 2011 9:59 am

Basically what you are doing is building an encoder. You could print a mask onto clear film and count pulses. You would also need to know the direction that the vane is moving....backwards or forwards and count pulses as appropriate.
I guess this may use too much power but you would have great accuracy, depending on the number of pulses per rev. 360 might be a good number :-)

Post Reply