Calibrating my anemometer for measuring really low wind speeds
Vložit
- čas přidán 5. 07. 2024
- I built a contraption to move my anemometer at known speeds for calibrating it with the external computer interface that I added to it.
Below 0.4 m/s, it no longer reads, so I have to rely on deflection of a hanging sheet of paper to measure that.
You could do a whole “calibrating my…” series. I bet it would kill.
Precisely.
Yes!
I know I’d what them all! G-d help me
“Precise junk”
Kill what?
This guy is always thinking....the world needs a lot more of him!
5:29 - the descending C-major triad is really satisfying.
wasn't trying to hit any particular musical notes!
Matthias, you would have made an outstanding physics teacher, your explanations and examples are so educational and interesting. You make it all look easy. Thanks for the video.
Haven’t even finished the video, but the sight and sound of this thing makes spinning up is delightful and hilarious. Thanks for the videos Mathias
I love the simple ASCII terminal plotter you use in all these videos. So useful! Would you be willing to share how you made that?
It's a single print command in his python data logging script - just calculate how many characters to print and do a 'print("#" *width)' and you're there!
What about the big numerical readout in the upper right?
If this is what he does to get his $20 anemometer working accurately
It's why manufacturers should send all of their devices to Matthias for review.
They'd get a comprehensive videoed review they could put on their web site and 2000 pages of test data supporting his review for their R & D department to digest.
it never cease to amaze me (it shouldn't) the level of ingenuity and creativity in your approach to problem solving. just keep posting your videos, i'll keep watching them.
You'll most likely reach the lower limits of those mechanical anemometers in this experiment. Have you considered ultrasonic anemometers? There's no minimum force to overcome at lower speeds, and they also give you direction! You can map and figure out how the air is moving through your house, e.g. is it being deflected off of walls, door etc.
You might consider buying a couple of the sensors to make your measurements... The experiment you propose to do is hard enough especially at low speeds... The reason you may want more than one of these to measure the wind's path because each time you walk through your experiment you disturb it and you have to wait for conditions to go back to baseline.
I know they're pricey, you can consider to DIY them... Each anemometer needs 2 pairs of Ultrasonic transducers a sender and a receiver (like the Ping sensor) and you measure the time of flight from a pulse generated by the transmitter to a receiver... then the same thing is done with the other pair at 90 degrees, then crunch the numbers, do some trig to calculate the angle, you get a vector of the wind (speed, direction).
I worked as a firmware engineer at a weather measurement company for 12 years, these are the very kind of problems in the field of measuring weather phenomena that comes up often, it's not as straightforward as people think... once in a while I get someone that can't believe it is that complex... and I'd say to them, if you think this stuff is easy, then clearly you don't understand the problem. For example: Solar irradiance needs cosine angle corrections, soil moisture sensors require AC signals to prevent migration of the sensing materials in the sensor, Rain has tipping buckets have errors at high rain rates and can't measure trace rain amounts... and last but not least... mechanical anemometers have low windspeed startup issues... not so easy after all... right?
Thank you for that fascinating rabbit hole! Half a dozen tabs later I'm reminded how little DSP I truly understand as a hobbyist
@@academicpachyderm5155 🙂I'm still learning. DSP Cool! I've just dabbled with it.
This makes me very happy inside. The amplified sound of a bit stepper motor is so satisfying.
I'm getting "decapitation hazard!" Mythbusters vibes from that arm. R.I.P. Grant, we miss you.
Hot wire anemometer for low speeds, just as other said in their comments. Those are used in small airplanes (gliders especially) and measure to 0.1 m/s speeds. Just 1 wire and resistance measuring ( plus "heating current").
Very interesting! What you need, for low wind speed is either an ultrasonic or hot wire anemometer (or an expensive laser Doppler anemometer), but the paper flap has low mass and seems sensitive enough for your project (and the principle has been used for ages, as a visual indication of wind speed on weathercocks - nevertheless a patent was granted for it in 1987).
Fascinating as always.
I don't know if it was your intention, but you proved the old saying, "One good turn deserves another"
I'm enjoying your fan experiments, and I am learning a lot, thank you. I live in a hot climate and have used fans coupled straight to a solar panel with a couple of resistors to move air in my tin sheds: when the sun shines, the fans blow. I hadn't got much further than that, but you are giving me inspiration to make it into something measurably useful :)
Welcome back to youtube! It had been a while since you posted that fan-distance from window video.
Interesting to learn house the air inside the house circulates. I’ll stay tuned due to sheer interest.👍
I clicked on this video on accident. Never would i ve thought that watching someone calibratin a wind speedopmeter could be so entertaining
I was thinking model train. A long straight section can be filmed with a metered background, then you have a scale to calibrate to. Add the RPi to the train and access via WiFi or record to read later.
Your programming skills is awesome .
Nice tonic triad on the piece of paper.
You are genius, so elegant solution
That is something I would hav loved doing or ever tried to do. Well done.
Haven’t got a clue what you’re talking about but I’m hooked😂
The random stuff channel is so good. Thanks Matthias
I love these experiments. I can see you are getting more and more confident with electronics and I am all for it ;)
sadly, the electronics topics don't get the same kind of views the woodworking videos get
@@matthiasrandomstuff2221 Wich is slightly understandable, you built your channel on woodworking after all. But I do think people with more diverse interests will be interested in your more techy content, like me !
@@matthiasrandomstuff2221Mouse videos do the best.
Nice work as always. Thanks
The fan experiments are great!
This was really cool, thank you for sharing!
5:30 It's beautiful to hear the maths behind simple intervals.
Ooooooo! So it's a series! I ❤️ a good series!
love your experiments. thanks
It might be worth considering using plastic instead of paper for your flap device. The paper catching humidity may impact the results, although perhaps too marginally to care about.
Great stuff as always.
You amaze me!
I'm a fan of this series
I see what you did there.
This kind of nerdy content is what youtube needs more of! I love it!
Love your problem solving. To avoid the twisting cable, maybe use an ESP32 and battery setup. To sync with stepper pulses, you can use a LED on the stepper pulses that flashes into a photo transistor connected to the ESP32. The other option is to time sync everything, I guess +/1ms is good enough.
Nice idea on the flapping paper to measure low air speed.
or just untwist it every time. Way less effort. Plus I wanted to do a reverse pass anyway
I was thinking the same thing, only with a Raspberry Pi and a power bank.
You make the most interesting experiments
Can’t wait for the next installment
This has a chance of becoming Matthias main channel.
It's pretty cool to see all these experiments, you can now write your own firmware y start selling accurate anemometers!
1:54 Time to start using string interpolation =) print("Initiate sweep, deg=",deg," freq=",freq) can be print(f"Initiate sweep, deg={deg} freq={freq}") (which I think is neater) and the same with the old % and the .format ones.
I'm kind of jealous actually, I got so used at writing software "the correct way" (with object-oriented programming, unit testing, TDD and all that stuff) that I wish I could just wing it lol The only times I do it that way is during programming competitions where speed is paramount.
I'm not a python expert, I will have to keep that print thign in mind. Python is great for just winging it. A lot of "parameters" I just put in the code. The code IS the configuration file.
@@matthiasrandomstuff2221 "use the source Luke" seems appropriate (I remember this from someones .sig file a long time ago).
You can even do f"Initiate sweep, {deg=} {freq=}" to save typing the variable names twice.
something pleasing about deriving an equation for a sensor output in excel
And yes, until this very day Matthias is still combing his hair with a balloon and making anemometers that can measure a fart. 😂
Pretty cool. A friend had an idea about how to measure low currents in a room using sound waves and calculating Doppler shifts. With an apparatus of known dimensions it should be possible to calculate speed and direction. Having several of them in a room, at different frequencies, might give a "picture" of the air currents. I don't think it was ever built, but I've always wondered now with fast computers how simple it might be (at the time he theorized about this 2400bps modems were in vogue).
This is how some house gas meters measure gas flow. Total use is the integral.
Doppler shift is a working but "time of flight" is the method of pro weather station sensors. See from a comment here.
Cool experiment, as always... measuring very low whatevers is always challenging... what about a big balsa wood propeller running on lubricated nylon or some low friction bearing? Now you have a way to calibrate it...
Going all the way for the fans, I see.
I have need for a similar sensor. I was thinking of using a "whisker" instead of the paper, with a small strain gauge to measure the force on it.
Classic Mathias Wandel. Love it :)
dont dare give this man free time
I wonder if an alternative would be to disperse some chemical marker in the air that a device can measure. If you know the diffusion speed from a static air experiment, then you could calculate that out of the concentrations measured around the house and get values for the drift.
Aerospace engineer, swinging through the same air is not ideal, but in this case i wouldn’t worry too terribly much. As a non-lifting surface only the drag is really perturbing the measurements in future sweeps but for just a couple I think it’s unlikely to be a major contributor to your error here. They used to have rotating boom equivalents to wind tunnels that worked ok but did face problems with re-entering the perturbed air
Beaufort scale recalibrated to candle flame using a light sensor connected to a raspberry pi to measure flicker.
Thank you for counting the toothpicks for the rest of us. :D
Thank you.
he is the most interesting man on the internet
Put a conical cylinder in front of the meter. The increased speed would be easier to measure. Divide the result by the amount the cone mechanically increased it to obtain true air speed.
I just tried that. It made maybe a 5% difference. Not worth it.
Have you tried using a large cpu fan and reading the pulses? You could use your rig to calibrate it. I assume it will have a higher breakaway torque because of the coils but would be neat to see a diy wind speed sensor. Love you low speed analog solution and test apparatus!
I have playeed around with fans. But it needs to be a 5" ball bearing fan, and the motor part needs to be removed for it to spin freely
Thats cool.
Don't know why, but I find the giant arm incredibly funny 😅
Love the escalation. I'm betting this started with "it won't take that long" to "I need another camer!" :) I agree with others who said "here's how I calibrated X" series. Love it!
Matt, for low speeds instead of deflecting a sheet of paper, you could consider a static orifice with a microphone to measure sound pressure volume, but you would probably have to calibrate that! U of W alumni.
those slow speeds don't make enough turbulence to be heard
@@matthiasrandomstuff2221 It doesn't have to cause turbulence! I think the idea was measuring the static pressure the wind would be causing directly from the signal, rather than the oscillations from turbulence. It not causing turbulence is a good thing!
P.S.: Especially with a large surface attached to the microphone (i.e. a sheet of rigid plastic or even paper attached to a piezo), it should be easy enough to measure the static pressure. I'm thinking about doing that experiment myself now out of curiosity! I've been wanting to measure the pressure coming out of those little ultrasonic humidifier devices for a few months, and this sounds perfect for that.
He’s one step closer to making his own halidron collider
I really like your clever ideas. Do you have plans for your low wind speed anemometer?
Meanwhile I’m trying to figure out how to open a bag of Cheetos without it ripping apart. And I’m failing at that…
i would love to see a video of the reaction of your family to all that. we love to see your videos but they have to live surrounded by strange contraptions (unnecessary) Lol!!
Engineer at work.
Wow!
Would be neat to get a smoke machine or really just anything that produces somewhat clean smoke in there to visualize the air movements
5:27 nice descending arpeggio!
Holy crap matthias.
But boy does this thing sound coool
I feel like I learn more about RIM here than from the movie.
Wouldn't it be easier to fill a room with "smoke" or with "prism" light and that way capture the air movement via dust or particles suspended in the air? I think the error factors are going to creep in at those super slow speeds any way to try to minimize it. Someone walking upstairs or downstairs or random knocks or creaking or settling or just temp and humidity changes might cause brief error readings right ? 😀
I think that works great for visualization, both for conveying the movement and figuring out roughly where air is moving, but the problem is quantifying it.
There's probably me, you and 10 other folks that get this :)
Very interesting. Why 2 meters? It seems that your life would have been easier with a shorter radius. Also, FWIW, if you soak the rims of the holes in your paper sensor with thin cyanoacrylate and allow it to harden (out of the apparatus that is), it will stiffen and be more robust.
So you plan to map airflow in your home? I think this is a great project. Thank you.
the long radius allows more time for the aerodynamic drag to dissipate before the next turn.
Neat
Do you take in account the current air pressure? Measurements will be different under different air pressure conditions, because of the density of the air. Not sure if humidity has a significant effect on the ‘performance’ of the fan at the same speed.
yes, that will make a difference, but the variation in air density is smaller than the accuracy that I'm aiming for
@@matthiasrandomstuff2221 understood 👍
Yes more fan science🌬
I've watched Matthias Wandel for years, but I bet I haven't 'Liked' more than a handful of his videos. Time for me to get my ducks in a row.
Nice setup. Have you considered not to measure the frequency at low speeds but rather the time in between pulses from the sensor? The ratio between max windspeed and very slow speed might be too big for accurate frequency measurements (in one frequency range). An alternative methode might be to increase the sensor pickup at low windspeeds by increasing the number of reflectionstrips?
timing indivisual pulses is risky as the motino may not be that smooth. But I accumulate at least 40 transitions or 3 seconds worth to make sure I have a good frequency reading.
@@matthiasrandomstuff2221 Thanks for the response. If the highest windspeed is lets say 25m/s and the lowest that you want to measure is 0.1m/s than this gives a frequency range ratio of about 250. To measure this accurate over the full range with lets say .01m/s accuracy you need most likely to change sample time (or gate time). Do you do this in the way you measure the frequency?
The sound as the rotor starts up … :)
You can use slip ring so your wire doesn't get tangled
Good on you for testing that . . . merchandise. Ahem.
When we suggested an OnlyFans channel we weren't being literal!
You can also use smoke machine and lasers to see how air fills home
Could you also take measurements with a mass airflow sensor from a car? That might give your more precise measurements.
Just go in circles: A management solution turned engineering solution.😂
wizardry
Think the quadratic would have also the force of gravity like a pendulum right? That might fix your curve.
At 5:29, did you intentionally pick speeds that would make the motor hum in a major triad? :)
No, I just picked nice multiples of each other
Instead of a peace of paper, maybe you could use thin material with much lower weight. And add a little weight on the bottom, for stability. In addition you could trying to create a cupped shape, so you receive more force, that will result in more accuracy.
I could also have used a longer strip of paper for more sensitivity. But at some point it just gets too twitchy
This paper sheet is too hard. Cotton yarn is extremely sensitive and used in soaring planes (gliders) in front of the pilot to indicate side slip.
If a fan blows in the other room, and no one is around to see a piece of paper move 2mm does it still move?
Yes because there is a remote camera on it.
5:29 sounds like a chord or a key signature
could make a larger prop for the meter, or add a collecting duct.
The collection duct is a neat idea. But making a prop tht runs easily enough is a challenge. I tried to make a better one, but the one in the meter runs extremely lightly already.
I tried a cone attached to the meter. Surprisingly, it made almost no difference, maybe 5%
This measuring device only measuring the speed in 1 direction. You have any thoughts about a 360 degree measuring device? I’m thinking about a balloon upside down, with a needle glued on the bottom, and a circular scale. So you can measure the actual speed and direction in 1 go.
I could just have two vanes 90 degrees different angles
Would it be possible to measure the static pressure on a paper disc? Instead of deflection?
with a sensitive enough scale, yes. but deflection is much easier
perhaps a smoke test would give you a visual representation of how the air is moving behind the fan
Smoke is a constant method in car industry for wind tunnel test observation; turbulence detection.
This is what happens at the Wandel household when the wife and kids go off for summer vacation (I'm supposing)
Ha, they are indeed away this week, but I filmed most of the video last week!
Looks complicated can’t you put it on the end of a pipe and blow air through at a given rate .
Air circulating smoke or a thermal camera.
I hung 9 temperature sensors in 3 lines from the ceiling floor and ceiling were 20 deg c different. Halfway between floor and ceiling was half way between two temps.
When petting aircon fan on recirculate all sensors went the the half temp , after turning fan off all temps split again to original readings.
Made some cool graphs
Why not a Venturi tube?
Isn’t the angle of the fan pointing 90 degrees, effecting the measurements? You could easily test it by giving the measuring fan a slide angle inward. My prediction is a few procent more fan speed, with a few degrees angle inward of the circle.
do the experiment, and you will see your prediction disproven.
@@matthiasrandomstuff2221 I only know that if I put my wind speed meter on a angle in the air direction, I’m getting lower readings.
yes, cosine error. But being 10 degrees off is not going to make a noticeable difference
This is fantastic. I'm trying to thin of a suitable way to caliprate a torque wrench to +/- 0.5NM accuracy, but all the methods I've found involve using a fish scale. Whilst I accept that I'm going to have to accept some third party source of truth, there must be a better way to do it?
I'd approach that by hanging weights on a lever arm turned by the torque wrench.