Arduino Based Weather Station

[Gina]

Simple and complicated ones on thingiverse   I just used a standard funnel on the top to increase the collection area.  There's at lot about my earlier weather station projects on another forum site.  One of my threads.

Edited January 3, 2017 by Gina

[Stub Mandrel]

My idea is to put it in a box with a mini-funnel on top, to go under a big (6") funnel. Need to add a movement detector, a blade for an opto switch, I think, glued to bottom of the bucket so it has a flat bottom!

[Gina]

That looks fine   I may well make myself a new one - can't be any worse that the Fine Offset brand cheapo one   I should be able to produce something rather more accurate with any luck.  I agree with an optical sensor - magnet and reed switch will slightly impede the motion.  An alternative would be Hall switch with magnet - there is no iron in the Hall device to pull on the magnet and a Hall switch is simpler than LED and phototransistor.

Edited January 3, 2017 by Gina

[JamesF]

4 hours ago, Gina said:

OTOH do we actually need accurate rain measurement?

"None", "A bit", "A lot", "Find the plans for the ark"

James

[Gina]

Yes, pretty much   I may just see if the FO rain gauge is working and possibly make my own later.  One thing wanted is a bit of heating when the temperature drops below freezing so that snow or freezing rain can be catered for.

Just printed the wind vane hub.  Was about to change over to white for the encoder but found the white wouldn't go through the 2mm PTFE tube in spite of claiming to be 1.75mm   Wonder if the translucent filament would work alright - should be almost as reflective.  Or there's pink.

[MarkyD]

I now use xBee's.  Very easy to use to send serial data across wirelessly.

[Gina]

I've used similar things in the past.

[Gina]

Been looking at the whole design again to see if I can make use of knowledge I have gained since I last looked at it.

I don't need to use WiFi if I attach the mast to my observatory which is what I'm planning.  Currently the mast is fastened to a fence post to the NE of the scope position and does get in the way of imaging at times.  With it placed to the north or NW it will be out of the way - I never seem to find anything to image in that direction.  Power will be supplied from my observatory battery backed 13.8v supply and I have gigabit Ethernet available back to the house.

I'm not really happy with using discontinued parts as there would be a problem of replacement in case of failure.  Also, although the circuitry is built and was working, I have only tried it at room temperature and cannot say it will work when it's very cold though according to the chip specs, it should do.  Also, I would prefer circuitry I can publish as a working system for anyone else to copy if they want to.  As such, I'm reviewing the whole thing.

As far as the wind instruments are concerned probably the worst cast of discontinued devices is the 1-wire counter - a great shame as this is a 32bit binary counter meaning it could count pulses from the anemometer and not overflow for 4,000 years even if we had constant winds of 100mph!  I'm going to see if the pulses can be counted in software.

Edited January 4, 2017 by Gina

[Gina]

Checked up on the approximate pulses per second versus mph and I get about 1p/s for 1mph so 100mph would produce 100 pulses per second or 10ms minimum time between pulses.  I reckon even the Arduino should be able to handle this from its interrupt system.

[MarkyD]

13 minutes ago, Gina said:

Checked up on the approximate pulses per second versus mph and I get about 1p/s for 1mph so 100mph would produce 100 pulses per second or 10ms minimum time between pulses.  I reckon even the Arduino should be able to handle this from its interrupt system.

pulseIn() or pulseInLong() 

[Stub Mandrel]

Oh dear. Printed the bucket part and the sides didn't come out at all - I made them too thin

[Gina]

How thin did you make them?

[Gina]

The other thing I'm not happy with is the wind direction encoder - a lot of variation in the photo-transistor current, at least so far.  Maybe photo-reflective sensors are not the best way to go after all.  To use the Gray encoder means I can't use slot type photo-detectors - the sender and receiver parts need to be separated.  This was the way I did it at first.

I'm considering another idea.  A slotted disc and software counter with two photo-detectors to allow up/down counting (bi-phase).  Much simpler mechanically but would it be reliable enough or might it lose counts and get "out of sync"?  I wonder.

[Gina]

A problem with opto-sensors is that the output current varies quite a bit and also is sensitive to noise.  Maybe that's why most commercial weather stations seem to use magnets and reed switches.  An alternative would be Hall switches, which I'm rather fond of I have to say   I think they are more consistent than reed switches as to the distance at which they operate.  Maybe I could get away with just 8 and make use of overlap.  I think some experimenting is indicated   They're cheap anyway

[Stub Mandrel]

39 minutes ago, Gina said:

How thin did you make them?

About 0.1mm >oops!<

[Gina]

Ah yes well...  TBH I wouldn't go below 1mm but even that needs a good printer.

[tekkydave]

I have one of These which I bought when I had some money

It has continuous analogue output accurate to a degree but I don't know how it works internally. I'm certainly not taking it apart as I could never afford to replace it now. I also have an Inspeed anemometer. I bought them from Hobby Boards a few years ago (now no longer trading) and they came with a 1-wire controller board.

[Gina]

Very nice

[Gina]

Been doing some research.  The photo-reflective sensors I'm using for the wind direction encoder are infra-red and it seems that black ABS reflects IR - doh   I think the ideal way to make an encoder disc that works with reflective IR sensors would be to carefully cut black velour flocking material to the shape of a Gray encoder and stick it onto aluminium foil or a thin aluminium disc.  OR cut suitable holes in aluminium sheet and put black velour behind it.

Actually, I guess I could print a fairly thin Gray encoder disc, glue kitchen foil to it and cut out the holes.  Put black velour behind it and there you are   Maybe a light coloured ABS would reflect enough light on its own.

Think I'll make up a test rig with one of the IR reflective sensors and measure the collector current with various colours of filament.  I can also test the reflectivity of FLO flocking material

[Gina]

I've moved my Pilot 3D printer off the table and onto a cabinet in the corner where the UP Plus 2 printer stood.  Also removed the fume cabinet ready for a new bigger one for the Pilot.  That leaves the table free for experiments.

[Stub Mandrel]

My gray code disc is printed on OHP acetate, but works by transmission rather than reflection.

[Stub Mandrel]

You may be interested in the quality of print from the Dremel. The bottom one is a 0.5 module gear so about 34mm across, it meshes well with a stock gear.

[Gina]

That looks fine

[Gina]

Yes,  I think transmission would be much better than reflection.  I look into that again.

Edited January 4, 2017 by Gina

[Gina]

Post disappeared again