Arduino and 1-wire based home built weather station.

[steve]

And how's the observatory coming along? Finished by now, I expect - any photos?

[Gina]

Hi Steve and everyone - I'm back (at last) Much has happened since I last posted here and I have a number of updates to post. Bottom line is that I shall be resurrecting this project shortly and have already been planning a local wall display in my living room.

Steve, hope you are well and still going strong Not read the general posts yet. The observatory was completed and is in use - well it is when the weather permits which hasn't been very often in the last months or more Of course, when I say completed, there are always improvements that can be made but it has been well operational for a couple of years now. If you don't mind threads off topic of weather, I post a brief report.

[steve]

If you don't mind threads off topic of weather, I post a brief report.

Yes please, I'd find that interesting. We have a (cheap) telescope but little chance to use it because of the weather.

[Gina]

I've 3D printed a number of parts for the weather station mostly for the wind instruments. Here is a new anemometer. A rather poor photo on the table and perched on my anglepoise type light/magnifier. I can post photos of the individual parts if anyone is interested.

Anemometer_22.jpg

Anemometer_23.jpg

[Gina]

I've been giving some thought to the sensor system for the anemometer. Currently the idea is a magnet on the rotor and reed switch underneath in the fixed part. This gives two closures per revolution as the reed switch closes with either N-S or S-N magnetic field. With one revolution per second estimated to represent 3mph this means a pulse per second for each 1.5mph. So the wind speed reading is only resolved to 1.5mph if a period of 1s is used for counting. I wonder if 1.5s would be short enough to detect gusts - that would give a 1mph resolution. Average speed could have a much higher resolution by using a longer counting period. Does anyone have any info on duration and form of wind gusts?

Another thing I noticed with magnetic sensing is that the anemometer rotor tends to "lock on" to the positions at either end of the reed switch due to magnetic attraction. This reduces the wind speed at which the anemometer starts to turn so I might change over to optical sensing. Also, I've heard of reed switches failing in this application with millions of operations a week. Optical sensors have no such problem.

Optical sensing could use a multi-slot disc on the rotor and give faster and more accurate wind speed readings.

[mcrossley]

Gina,

For the 'cogging' issue, the solution I have generally read about is to move the reed switch further away from the magnet or use a weaker magnet. It only needs a weak field to close the relay.

For your windspeed, an alternative approach may be to time the interval between pulses, rather than count pulses in a period. You would probably want to do a little smoothing with a rolling average of a few readings.

[Gina]

Thanks Mark

I could try a weaker magnet or moving it further away as you suggest.

I had thought of measuring duration of each half revolution rather that counting pulses - it's certainly a possibility though I haven't worked out just how I would do it. I've been thinking of this idea for the rain gauge too - counting bucket tips is not very accurate. I'll give it more thought and see what I can come up with.

[laulau]

Or a hall effect sensor instead of the reed switch like in the recent Davis anemometer !

[Gina]

Yes indeed I have a few Hall effect sensors.

[Gina]

I think I shall use a hall effect device and three magnets which will fit well into the rotor that takes the cup spokes. Here is a photo of the anemometer parts. To change over to a hall effect device requires redesigning and printing two parts of the anemometer- the rotor to take two more magnets and the fixed bearing support currently with a slot for a reed switch (and printed in pink ABS) which will take the hall effect device.

Anemometer_27.jpg

Rotor with one magnet. This is actually separate from the spokes and cups and can be redesigned and printed twith three holes for magnets.

Anemometer_24.jpg

Same with ring to keep wet out

Anemometer_25.jpg

Pillar with bearing. The part printed in pink is separate, just a tight fit and currently with a groove to take a reed switch. This will be replaced with another print to hold a hall effect device.

Anemometer_26.jpg

[Gina]

I've made a new rotor taking 3 tiny "supermagnets" a mere 3mm cube, and just modified the stator by drilling a hole for the hall effect chip wires. Made a test rig with a LED and series resistor to check that everything worked. It did and I have reassembled the anemometer and tried it out with the test rig and a room fan some distance away on lowest setting. Anemometer is working well even with a very low air flow and several seconds between pulses representing 0.5mph or less. Taking 3 as the anemometer factor, 3mph will give 1 pulse per second with 3 magnets and I estimate the time between pulses was at least 2s corresponding to 0.5mph or less. I'll post some photos and the test circuit diagram tomorrow.

[Gina]

Here are some photos of the mod for using the hall effect device. I designed and printed a new rotor taking 3 miniature "supermagnets" a mere 3mm cube but just did a small mod on the stator, drilling a hole for the hall effect chip wires. I connected up a test rig with LED and resistor to test the hall effect device operation. It worked well and was pretty well balanced. I tested the slow running ability and the anemometer was working with a "light air" from a room fan. The LED was flashing slower than every couple of seconds so with with three magnets and 1 pulse per second corresponding to roughly 1mph that meant less than 0.5 mph.

Firstly here's a circuit diagram for the test rig.

Test Rig 01.JPG

Hall effect device mounted on stator.

Anemometer_28.jpg

And this is the rotor with three magnets.

Anemometer_29.jpg

Using the test rig with just the rotor to check that the magnets clear the hall chip and that it works before further assembly.

Anemometer_30.jpg

Cups added to rotor to check clearance of inner ends of spokes.

Anemometer_31.jpg

Finally, the fully assembled anemometer set up to check free running and see how sensitive it would be to very light breezes.

Anemometer_32.jpg

[Gina]

Just looked back through this thread and find I didn't post the circuit diagram for the wind instruments so I think It's time I went through and discussed it. Where possible, I'm using 1-wire devices to make the measurements as this uses a multi-device bus rather like USB or I2C etc. but using just one data wire plus ground. Twisted pair wires as used in CAT5 ethernet cable work fine for the network and devices can be used at up to 100m. As I recall - 1-wire for weaather stations was discussed elsewhere in this forum.

There is problem though - several 1-wire devices have been discontinued by the manufacturers Maxim (previously called Dallas) and they are not producing any alternative. With least two of these devices being used for weather stations, this leaves us in the mire

I'll just stop for a moment and look at what was posted about 1-wire by myself and others.

[Gina]

Yes, found a thread I started about 1-wire devices. Not quite up-to-date as Maxim have discontinued the DS2450 Believe it or not there are still several 1-wire chips left in production

I've been through the list and just the following six devices seem of much use to me :-
DS18B20 Digital thermometer
DS2406 Single or Dual Addresable Switch
DS2408 8 channel PIO
DS2413 Dual-channel programmable I/O
DS2417 1-Wire Time Chip with Interrupt
DS2438 Smart Battery Monitor containing 10v A/D, 250mv A/D, thermometer, elapsed time counter

Two of the three 1-wire devices I had in my wind instruments circuit are discontinued and the third is not needed with +5v available at the mast top. I'm considering a complete change of design

This was my first design.

Wind Module Circuit 01.jpg

And with +5v available I cut out the DS2413.

Wind Module Circuit 02.jpg

[Gina]

Question now is - do I continue using obsolete components for which no replacements are available or change the design to use available and currently supported ones?