low power analog wind sensor / anemometer

[crice]

Hello all, I am in need of some design advice.

I am designing a low power anemometer. The main hurdle here is that I can work with only a 2.5V supply (with 3V as the absolute maximum because I am running on 2 AA batteries). Also, I need to have a voltage output. So, if I were to look at the output voltage of the anemometer, I can correlate the voltage to a wind speed. I only need to detect wind speeds of roughly 150 fpm to 1200 fpm (sorry I am from the US ). I have toyed with many ideas of how to approach this problem.

First, I thought of using a DC brushless fan to generate voltage. When a wind forces a dc fan to spin, it generates a small but measurable voltage. I thought of somehow correlating wind speed to the output voltage in this way, but I could see no discernable pattern or correlation. Also, the fan was not very sensitive to low wind speeds because of the internal metal coils. I tested this with a 40x40x6mm Sunon DC fan. Anyone have ideas? This would be the most ideal solution for my design, because it requires no extra circuitry to function. All it needs is a DC fan.

My next idea was to build the homemade easter egg anemometer as documented here: https://www.otherpower.com/anemometer.html
However, the problem with this anemometer is that LM2917 frequency-voltage chip they use requires a very high supply voltage. Is there a low power version of this chip? I cannot seem to find a frequency-voltage chip that needs less than a 9V supply.

This is where I am stuck. I have been researching some analog tachometers, but I am not sure how they apply to my design. So my dilemma is this, I need to design an anemometer that outputs a voltage and only uses a 2.5V supply.
Any ideas would be greatly appreciated!

 

[philba]

any permanent magnet motor will generate a voltage when spun so make your cup anemometer and attach it to the shaft of a small dc motor. the fan won't work for low wind speeds because it was designed to be efficient at something like 3K rpm.

Does the output have to be a DC voltage? With a micro, you could just make a photo interruptor or hall effect device and count pulses. If you pick the correct counting period, the number of pulses would be the wind speed.

 

[crice]

Thanks for your reply.

The output should preferably be a DC voltage. I'm interfacing this anemometer with another device, and the most ideal situation for me would be to "poll" the output voltage at any given time to obtain the correct windspeed (after doing calibrations).

I forgot to add that I also thought about what you had mentioned, philba. I played with the idea of using a reed switch + written software to count the number of pulses from a spinning set of cups or fan. The main problem here is that I think I would have to keep my device on at all times to obtain an accurate reading of the speed. I believe for the DC voltage case, I can wake my device up, get the voltage output reading, then make the device go back to sleep so that I can save precious battery life.

EDIT: What I'm most interested in is a frequency-to-voltage chip that requires a very low voltage supply.

 

[philba]

Thanks for your reply.

The output should preferably be a DC voltage. I'm interfacing this anemometer with another device, and the most ideal situation for me would be to "poll" the output voltage at any given time to obtain the correct windspeed (after doing calibrations).

I forgot to add that I also thought about what you had mentioned, philba. I played with the idea of using a reed switch + written software to count the number of pulses from a spinning set of cups or fan. The main problem here is that I think I would have to keep my device on at all times to obtain an accurate reading of the speed. I believe for the DC voltage case, I can wake my device up, get the voltage output reading, then make the device go back to sleep so that I can save precious battery life.

EDIT: What I'm most interested in is a frequency-to-voltage chip that requires a very low voltage supply.

actually, you could power up the device, wait for a pulse and then start counting. once the count period has elapsed, put it back to sleep. no need to always be on.

with the generator approach, you could use the generator to recharge a battery and get unlimited life.

Also, I've been playing with TI MSP430 microcontrollers. they are ultra low power devices and I see how I could make an anemometer that would be super low power - running off of a 3V coin cell for years.

 

[dknguyen]

The output should preferably be a DC voltage.

Does the output have to be a DC voltage?

DC motors put out a DC voltage when back driven, while AC motors put out an AC voltage when back-driven. The internal commutation does that. So using a DC motor as a tachometer would give you a DC output.

If you need a to count pulses for frequency but need the signal to be unipolar, maybe use an AC motor with a diode overvoltage clamp or something to produce a square wave.

 

[Paul Obrien]

Have you considered the Dallas 1-wire weather station the power comes from the data line. https://www.aagelectronica.com/aag/index.html?target=p_1.html&lang=en-us

 

[philba]

DC motors put out a DC voltage when back driven, while AC motors put out an AC voltage when back-driven. The internal commutation does that. So using a DC motor as a tachometer would give you a DC output.

If you need a to count pulses for frequency but need the signal to be unipolar, maybe use an AC motor with a diode overvoltage clamp or something to produce a square wave.

I was asking if the interface needed to be a DC voltage level. nothing to do with how that level got generated. my point was, that seems like a somewhat inconvenient choice of interfaces.

 

[HiTech]

hey crice, I read that Easter egg anemometer article weeks back. I started to scavenge a motor from an old vcr, but things got unexpectedly difficult in saving it in one piece while removing the attached hardware. I'll keep watch of this thread to learn of your outcome. Maybe it will aid my efforts. So far the project is on the back burner due to other priorities.

 

[Klaus]

OK, you got a lot of answers which are, as usual here, on the complicated side.
Think simple and you may get something useful and workable within your low voltage restrictions.
The simplest anemometer that coud provide an electrical output is a horizontally hinged vane. You build it to the expected windstrength, sizewise and in robustness.
All you need now is a 10K or so servopotentiometer, which is coupled to the hinge pin. The potentiometer is connected as a voltage divider and you calibrate the deflection of the vane for the wind strength. This will be non linear most likely.
Why a servo pot? They have minimum drag due the ball bearing shaft and special wipers, which offer minimal load to the vane. You coluld try an ordinary pot first if you find one that has very little resistanc to movement by hand.
have fun,
Klaus

Hello all, I am in need of some design advice.

I am designing a low power anemometer. The main hurdle here is that I can work with only a 2.5V supply (with 3V as the absolute maximum because I am running on 2 AA batteries). Also, I need to have a voltage output. So, if I were to look at the output voltage of the anemometer, I can correlate the voltage to a wind speed. I only need to detect wind speeds of roughly 150 fpm to 1200 fpm (sorry I am from the US ). I have toyed with many ideas of how to approach this problem.

First, I thought of using a DC brushless fan to generate voltage. When a wind forces a dc fan to spin, it generates a small but measurable voltage. I thought of somehow correlating wind speed to the output voltage in this way, but I could see no discernable pattern or correlation. Also, the fan was not very sensitive to low wind speeds because of the internal metal coils. I tested this with a 40x40x6mm Sunon DC fan. Anyone have ideas? This would be the most ideal solution for my design, because it requires no extra circuitry to function. All it needs is a DC fan.

My next idea was to build the homemade easter egg anemometer as documented here: https://www.otherpower.com/anemometer.html
However, the problem with this anemometer is that LM2917 frequency-voltage chip they use requires a very high supply voltage. Is there a low power version of this chip? I cannot seem to find a frequency-voltage chip that needs less than a 9V supply.

This is where I am stuck. I have been researching some analog tachometers, but I am not sure how they apply to my design. So my dilemma is this, I need to design an anemometer that outputs a voltage and only uses a 2.5V supply.
Any ideas would be greatly appreciated!

 

[philba]

That's an interesting idea but has some issues.

- weatherproofing the pot. not impossible but needs to be able to withstand any weather, including moisture/humidity/...
- you will only use a small part the pot's swing - say 45 degrees.
- above a certain speed you will have value compression rendering it useless. for example, the difference between 30 and 31 KPH may well be less than the noise level.
- the assembly needs to pivot to face the wind. more complexity.
- what effect will rain have on this? it will make the vane heavier than thus not deflect at the same level.
- I don't see long term reliability in that design. pot wear, hinge pin wear, corrosion, dirt, ...

I disagree that timing pulses is complex. No moving parts other than the actual cup assembly. The electronics can be sealed to the weather, analog noise is not an issue, no need to pivot into the wind direction and the upper limit on speed is basically the cup assembly's structural integrity.

 

[Klaus]

Well, if you think about it, you would also have to:

weatherproof the shaft of a rotating cup anemometer (I built one and its NOT an easy job if the tiny ball bearings should last. Its been going on my roof for over 20 years now and it does need maintenance every 6-8 years or so).

You can use a simple plastic gear set to increase the pot travel 2 or 3 times.

You lost me with your value compression statement but I would challenge you
to build a rotating cup anemometer that can reliably differentiate between 30 & 31 Kph. You are doing very well if you get reliable increments of 5 units over the whole range. Keep in mind that most mortals do NOT have access to a wind tunnel to calibrate the thing.

Keeping a vane facing to the wind is no big deal, its done all the time on horizontal axis wind generators.

Rain does affect the rotating cups too, the effect on a vane can be minimised by a water repellant surface.

Prevention of wear & dirt ingress, corrosion, etc. is common to ALL electromechanical anemometers.

All I did is suggest a simple system that can cope with the limited supply voltage. You would be battling to drive the timing pulse electronics from just a little 3V battery as the original poster specified.
A vane is certainly MUCH simpler to build than a reliable rotating cup anemometer. it should be quite sensitive if the vane is partially balanced. The original poster DID specify a range of windspeed and it could be tailored to that.


Now, if you could figure out a way to get an electronic readout from these ball in a tube venturi anemometers then you would have cracked the simplicity challenge and I' tip my hat to you

Klaus

That's an interesting idea but has some issues.

- weatherproofing the pot. not impossible but needs to be able to withstand any weather, including moisture/humidity/...
- you will only use a small part the pot's swing - say 45 degrees.
- above a certain speed you will have value compression rendering it useless. for example, the difference between 30 and 31 KPH may well be less than the noise level.
- the assembly needs to pivot to face the wind. more complexity.
- what effect will rain have on this? it will make the vane heavier than thus not deflect at the same level.
- I don't see long term reliability in that design. pot wear, hinge pin wear, corrosion, dirt, ...

I disagree that timing pulses is complex. No moving parts other than the actual cup assembly. The electronics can be sealed to the weather, analog noise is not an issue, no need to pivot into the wind direction and the upper limit on speed is basically the cup assembly's structural integrity.

 

[philba]

Well, if you think about it, you would also have to:

weatherproof the shaft of a rotating cup anemometer (I built one and its NOT an easy job if the tiny ball bearings should last. Its been going on my roof for over 20 years now and it does need maintenance every 6-8 years or so).

You can use a simple plastic gear set to increase the pot travel 2 or 3 times.

You lost me with your value compression statement but I would challenge you
to build a rotating cup anemometer that can reliably differentiate between 30 & 31 Kph. You are doing very well if you get reliable increments of 5 units over the whole range. Keep in mind that most mortals do NOT have access to a wind tunnel to calibrate the thing.

Keeping a vane facing to the wind is no big deal, its done all the time on horizontal axis wind generators.

Rain does affect the rotating cups too, the effect on a vane can be minimised by a water repellant surface.

Prevention of wear & dirt ingress, corrosion, etc. is common to ALL electromechanical anemometers.

All I did is suggest a simple system that can cope with the limited supply voltage. You would be battling to drive the timing pulse electronics from just a little 3V battery as the original poster specified.
A vane is certainly MUCH simpler to build than a reliable rotating cup anemometer. it should be quite sensitive if the vane is partially balanced. The original poster DID specify a range of windspeed and it could be tailored to that.


Now, if you could figure out a way to get an electronic readout from these ball in a tube venturi anemometers then you would have cracked the simplicity challenge and I' tip my hat to you

Klaus

the compression issue is that the reading are non-linear. as the deflection increases, the force pushing back against the wind increases and thus increasingly more force is needed to deflect the vane. as the deflection increases, the "sail area" decreases as well. above a certain wind velocity there will be no deflection. the vane deflection approaches this point asymptotically. as the vane approaches this asymptote the differential readings will decrease. Since this is an analog system, noise will come into play. when the difference between 2 readings (my example between 30 and 31) is less than the system noise level you can't distinguish.

with a pulse measurement technique, it is trivially easy to very accurately measure the rotational frequency. not only can it discern between 30 and 31, it could discern between 30.9 and 31, if that matters much. Yes, you will need to do filtering as the wind velocity is constantly changing. And, yes, we do agree that calibration is an issue for both designs.

the problem with the rain issue on a vane is that it also changes the force on the vane. I'm not even sure how to model it though.

more mechanical stuff = complexity and unreliability. I agree that both approaches have corrosion issues but fewer moving parts is better, imo. that's why I think a rotating cup approach is better.

by the way. roller blade bearings are pretty good. they are very cheap, take side force well, are sealed and last a long time.

 

[Klaus]

Thanks for enlightening me by by explaining what you meant re the word 'compression'. You might check back that I did mention non linearity of the vane output, its just a question of terminology then.

Agree with the finite wind speed a vane would work reliably at. It should be possible to fine tune the device for a certain range.

Yes, I know all about the pulse output technique, remember me mentioning that I have such an anemometer on my roof for a very long time?
What I disagree is the practicality of making the pulse from a supply as low as 2.5V. The average opto slot switch needs a little more to work reliably.

A pot set up as a voltage divider could give a suitable analog output. No fancy IC chips required at all. Minimum current draw if the pot value chosen is high enough.
The servo pot I use in my wind DIRECTION sensor has 360 degree mechanical and ~ 350 degree electrical rotation. Easily good enough for the 16 wind directions I'm using it for with my LED readout. Downside, it WAS expensive.
It would work just as well on a horizontally hinged vane, though the only about 60 degrees deflection of the vane might give 1:3 up gearing an advantage.

Yes, I agree with you, a rotating cup device IS a better choice but:
a), its not easy to make with just 2.5V battery supply.
b) it requires access to some fine machining facility and a precision scale to balance the cups & arms. I have that, you may have it but does the original poster?


Thanks for the roller blade wheel bearing tip, I'll keep that in mind for my next project requiring durable bearings. BTW, fully sealed bearings have far too much drag for a sensitive anemometer, IMO. I removed the non exposed seal and washed out the grease, replacing it with fine machine oil. The cups spin up at the slightest breeze.

Klaus

PS. somebody mentioned above that one gets DC from a driven DC motor and AC from an AC motor. Not neccesarily so, certainly with AC motors, as a field current has to be supplied first (not possible on sqirrel cage rotors). Only PM (permanent magnet) motors give a direct output voltage but they suffer from cogging which is detrimental on an anemometer.
A PM generator with no iron at the stator would be my choice for this application. Challenge is to build one small enough with little inertia, to respond quickly to wind changes.

the compression issue is that the reading are non-linear. as the deflection increases, the force pushing back against the wind increases and thus increasingly more force is needed to deflect the vane. as the deflection increases, the "sail area" decreases as well. above a certain wind velocity there will be no deflection. the vane deflection approaches this point asymptotically. as the vane approaches this asymptote the differential readings will decrease. Since this is an analog system, noise will come into play. when the difference between 2 readings (my example between 30 and 31) is less than the system noise level you can't distinguish.

with a pulse measurement technique, it is trivially easy to very accurately measure the rotational frequency. not only can it discern between 30 and 31, it could discern between 30.9 and 31, if that matters much. Yes, you will need to do filtering as the wind velocity is constantly changing. And, yes, we do agree that calibration is an issue for both designs.

the problem with the rain issue on a vane is that it also changes the force on the vane. I'm not even sure how to model it though.

more mechanical stuff = complexity and unreliability. I agree that both approaches have corrosion issues but fewer moving parts is better, imo. that's why I think a rotating cup approach is better.

by the way. roller blade bearings are pretty good. they are very cheap, take side force well, are sealed and last a long time.

 

[philba]

You can certainly get pulses with 2.5V (actually 2 AA cells). lots of HE sensors work at 2.5V and have very low idle current. allegro makes several that would work.

 

[stevez]

Consider a strain gage in a bridge arrangement to produce DC output directly. Mount the strain gage on a mast so that the wind pressure bends the mast (slightly). A vane or tail fin - like an old weather vane would have to be used to keep the wind direction/strain gage orientation proper. Things to consider would be the flexibility of the mast and mass - so that it doesn't whip around adding a lot of noise to the process.

Keep in mind that measuring the velocity of air with a sensitive and fast instrument can produce an awful lot of very noisy data. Wind currents close to the ground are quite variable in velocity and direction. The rotating cup or other mechanical parts do filter, in a way, this noise or variability because of their mass. You might need to add some electronics to accomplish the same thing.

 

[crice]

Thank you for all your input.

However I'm sorry, I forgot to update my original post. This anemometer that I am designing will be used inside air ducts within buildings or houses, hence the very small range of wind speeds. I won't have to worry about natural, outdoor corrosion or related problems.

I've been looking into magnetic reed switch sensors to provide the necessary pulses for the frequency-voltage IC. Anyone have experience with these?

 

[Nigel Goodwin]

Thank you for all your input.

However I'm sorry, I forgot to update my original post. This anemometer that I am designing will be used inside air ducts within buildings or houses, hence the very small range of wind speeds. I won't have to worry about natural, outdoor corrosion or related problems.

I've been looking into magnetic reed switch sensors to provide the necessary pulses for the frequency-voltage IC. Anyone have experience with these?

They work fine, but you have the added weight of the magnet to move with your 'fan'.

Have you considered ultrasonic methods?, no moving parts!.

 

[crice]

They work fine, but you have the added weight of the magnet to move with your 'fan'.

Have you considered ultrasonic methods?, no moving parts!.

Can you elaborate on the ultrasonic methods?

 

[Nigel Goodwin]

Can you elaborate on the ultrasonic methods?

It was invented by a British yachtsman a number of years back, google finds plenty of hits, try **broken link removed** which explains it quite well.

The magazine EPE also presented a PIC based project a few years ago!.

 

[audioguru]

The analog part of the circuit uses slow LM324 opamps that work poorly at ultrasonic frequencies, and their input polarities are backwards.

The input capacitor's value is so tiny that it passes frequencies only above about 1MHz and the feedback capacitor's value is so big that it cuts frequencies above about 7kHz.