I’m a vibration analyst by trade and your method is spot on! The only additional thing we do with the “test” weight information is build a vector diagram on a polar plot to determine where the optimal final weight should be located. Then there is some math to determine the weight amount but that’s just some simple cross multiplication. Doing this all with a speaker and o’scope is just downright genius.
Spot on, search for graphic balancing method. You can solve your needed weight and place by doing 3 runs while moving the test weight in 120° increments around. The place your three influence circles meet gives you both the place and amount of correction weight. Using a speaker as a velocimeter is genius!
@@G5Ckxew for the 3point Methode, you don't need the phase angle, for vector Methode the phase angle of the vibration is needed that way you only need to do a single test run with a test weight to get the correction weight and position. Getting the phase angle or a reference is done by a keyphasor, or a laser module that detects a ref mark. As a DIY solution you would need anything that gives you a signal on the scope exactly once per rev, always at the same point.
Really interesting to see a speaker used that way. All you need is a sufficiently capable rotary encoder for position tracking and you've nearly caught up to industrial impeller balancing systems.
I think he wouldn't even need a rotary encoder. a simple photo diode to measure a fixed point per revolution should be sufficient for home application. With a dual Chanel scope he could overlay speaker and "zero point" and from there calculate the angle where mass needs to be taken off.
There are solutions that don't need to know position. Some Balance Technologies machines just have a strobe light so you can point it at the rotating object and it will freeze the point where weight needs to be added or removed. And with the amplitude of the vibrations you can set the correction radius and readout the grams required. This was all old school analog circuitry, too.
Wow Matthias! You never cease to amaze me! The way you combine intuitive understanding with real-world electronic know-how is clever and clear. When I was a kid and built slot-car motors, (circa 1964) I just chucked the armature into a drill, spun it, and lightly touched it with a magic marker. Then wherever the marks were thickest, I drilled a small hole in the armature; Reversed the armature, and did the same thing from the other side. That was as close to dynamic balancing as a cash-strapped kid could get, But it worked.! Of course I didn't have access to these electronic measuring devices then, so that method was a tedious iterative process. Even now, I would have never thought of using a speaker to generate a readable pulse into an oscilloscope. Brilliant idea!
With two speakers configured orthogonally and two channels on an oscilloscope, Lissajous curve could be displayed. You could directly find the vibration vector. If you further know the phase of the motor, you could even calculate and display the exact angle at which you should add or remove weight.
I love the low cost approach to everything you share. It is so refreshing in a world of "Look at all this cool stuff I got for free to show you, please go buy it".
I confused dynamic with active. I thought you were going to feed back the vibration into the weighted voice coil to damp them that way (with destructive interference). Your way makes way more sense. Particularly for the heavy cutter. Looking forward to the conclusion!
That impresses me the most in your videos is the didacts you have. Even when I know nothing about what you are experimenting with I'm always able to understand the concept. You would be a perfect college professor, for sure!
You could possibly get similar results using a 'vibration meter' app on your phone (for those of us without an oscilloscope). I dynamically balanced the drive pulley on my metal lathe by taping an old phone to the headstock and measuring the vibrations.
Me too, but I want an oscilloscope, its probably much more accurate. Vibration analyzers at mm/s^2 range are very expensive as well 😢. I balanced my angle grinder with my phone lol this is amazing.
If you have a small mirror and a laser pointer laying around, you could probably do this completely analog. Put a piece of paper on a wall far away to mark out the sweep of the laser. The added benefit is electrical noise would have no effect.
Man! That was a super cool video! I am astonished by your method, as always. I dream of being "your kind of smart" but the truth is, I'm a little lazy about trying to be. However! You explain the details so perfectly that I sit here and think to myself "That makes total sense. Of Course that's how that works!". I truly do understand and follow your explanations. I just, would never have gotten there myself. "That" is what perplexes me sometimes. That's the fun in it all... Right?! I have thoroughly enjoyed every one of your experiments over the years. The mice experiments were lots of fun! I appreciate your dedication to not only the true and sometimes brutally honest results but also to the storytelling arc. Thank you.
This looks like a job for a piezo transducer! Also thanks for the demo. I am about to buy one of those blowers and this will help a lot if it's unbalanced.
This is quality content. I like it a lot. Balancing fans and rotating things with a cheap diy friendly way and increasing their lifespan, lowering the noise and vibration is a really good thing. The earlier video you had with balancing a fan really helped me right after i bought a box fan (brand new) it wash shaking violently and dancing over the floor, but after a few minutws balancing it it became very silent and no noticeable vibration
On the driven end of the rotor I believe you were seeing a central frequency of 60hz and sidebands each side, offset by 10hz. I suspect it was caused by geometrical angular/offset misalignment between the motor and fan rotors if they are rigidly coupled together. Very smart use of the speaker to measure the waveform!
I proposed a system like this for using a speaker and diaphragm weight to dynamically balance motors about 20 years back on the electronics forums. The one thing you missed which would have been an improvment would be to place the whole assembly on an elastic suspension, which could have been as simple as a springy foam block. Your setup was damping the vibration into the table which reduced the measurement amplitude and accuracy. Very cool project! 👍
A guy at work once had a fan very similar to that where a blade had busted off and was no where to be found. He was trying very hard to fabricated a thin blade to replace it when I suggested we just count how many blades there were and snap one off from the opposite side. We counted and it was all good and snapped the opposing right off and it ran sweet.
Beautiful and insightful work, congratulations! The 69-ish Hz you described at timestamp 5:08 is possibly due to the aliasing between the 60Hz network frequency and the 50-ish Hz of the rotational speed. The aliasing or beating frequency generates a pair of side bands (+10 and -10 Hz) The 10Hz can be also observed at the 10Hz region of the spectra and as a sideband of every 60Hz harmonics. If the system responds well to higher frequencies, you may also find a series of 100-110-120-130 -140Hz and so on.
Have you thought about marking a zero position with reflective tape and use an optical pick up to mark the zero position on your oscilloscope? That would make finding the “location” of the imbalance easier than trial and error.
My God, dude. Your methods are simply inspired! What an awesome approach. I've yet to see a video from you where I did not learn something. Bloody brilliant.
This is fascinating. Definitely beyond the scope of any of my projects but it's interesting to read all the comments from all these knowledgeable viewers.
Very interesting! I'm thinking of combining this to Denis' approach (channel "the GADGETS playlist") to dynamic balancing, ie, using a microphone (or speaker) to listen to the vibrations and reference it using a photocell to figure out where the weight should go - and perhaps combine it with the Y approach, ie, instead of adding weight at position 0, add it (for instance) at positions -3 and +3. Denis uses an accelerometer to measure the vibrations, not sure which option would be more accurate.
in the 1980s the company I worked for looked at analysing vibration from the heavy industrial machinery it made in an attempt to give the operator maintenance information much earlier than his ears could detect. It actually worked quite well but few end users would pay a worthwhile price for it.
There's a mentality of "don't fix what isn't broken", which seems to stop them from considering that "Everything is gradually breaking, so it WILL break at some point. Fix it when it's most effective to do so". I've worked with a number of factories and suggested that premise as well. Advising that a lot of their stoppages and repair costs could be reduced, AND would offset the cost of any system upgrades by an order of magnitude. Not a single one was willing pay for it except where I just did it anyway, and demonstrated it, and billed them.
With this technique, you have to make sure the resonant frequency of the speaker is not the same as the DUT. Otherwise the vibrations will look a lot stronger than really is.
@@AmirASD The more mass, the more inertia, the lower the frequency of resonance. Same thing with inductance. The more henrys, the lower the resonant frequency.
Does it really matter? The frequency response will be nonlinear but we only care about relative changes. As long as the DUT frequency remains constant I think it would still work fine.
I have some 3D-printed flywheels on small RC brushless motors that I need to balance, and this method looks like it would be ideal and uses stuff I already have on hand. Thanks for the tip!
Very clever! I imagine that squirrel cage fans are pretty tolerant of imbalances given that they all collect dust, and any dust buildup will likely throw off even the most carefully balanced fan. However, a balanced fan will give off less fan noise and that's always a blessing. Good show on the use of low cost test gear to analyze a problem!
The dust will be evenly distributed around the evenly sized and evenly spaced blades of a squirrel cage fan, so it doesn't affect the balance much. Initial balancing really protects the machinery and makes it last longer. There are industrial balancers that balance absolutely everything in, for example, a pulp mill. I once pulled a single leaf out of an automobile's squirrel cage blower that was making a horrible noise and that fixed it perfectly.
30 years ago we use to characterize disk drive motor samples with a similar set up sensitive and expensive B&K gear. Today, you can pull the 9 axis of freedom board off your drone and couple that with your favorite eval board (pi, arduino, stm32) and throw together a pretty darn good sensor system. Couple that with all the free firmware that deals with static and dynamic behavior and you will know more about anything you want to measure. The vibrations in the handle of your weed eater, lawn mower, shovel, hammer or vibrations in the motors... you could spend your whole summer on this..
That's a pertty interesting idea to use a speaker as an accelerometer. Ive seen them used as microphones, but never for low frequency. It's not a calibrated measurement, but for some quantitative analysis it's more than good enough.
my father has a balancing machine for big rollers and ducts. Althought its a little bit more sophisticated with a control computer etc., it messures the vibration on both sides of the tube (its laying there on bearings) and always knows at what position the duct is with the help of a light sensor and a glued on tag, after you run it it tells you the spots that need weight automatically.
As a Vibe Tech by trade, what you got at the Driven End where the motor is, was Sidebanding. Looking at the data you got, I'd say that you might have an eccentricity present, however as for 60hz being the Center Frequency, I can't currently figure that part out.
i might actually use this method. ive made fans and fast rotation devices before and many times simple field balancing techniques will not suffice. it also shows the thousands of uses for signal analyzation of electro-mechanical components using an oscilloscope.
So about the 120Hz humming: A transformer iron core expands and shrings with the magnetic forces involved. An electric sinus signal means a plus and another negative wave or going from 0 to max to 0 to min and back to 0. So that's why a transformer core hums with double the frqeuency of the electric signal. In most (if not all) parts of Europe it's a 100Hz humming, because we use 50Hz electrically.
With the addition of either an optical sensor tach, on the scope, the phase of the vibration relative to the shaft position, can be used to find the offset and amplitude of the balance for accurately finding the weight and position for the correction. This is how a tire shop wheel balancer works.
Thank you so much for this. So well explained and demonstrated. The only thing I'd like more information on is the weights added to the speaker. How much? The purpose? Etc.
Imterestimg video. In the past I've wondered if there's some way to balance a ceiling fan besides by trial and error. I have to believe this approach could be adapted for that.
Brilliant! Could I suggest using a piezo transducer instead of the speaker. It will likely be much more responsive and less likely to add sympathetic impulses as a piezo has zero moving parts(no coil, cone, etc..). Also, if possible within reason, when you attempt to balance your cutter head, make a stand for it and spin it with a damper of some sort between motor and head. Won't have to be a strong motor, and perhaps something as simple as latex tubing would work to couple the motor and head and provide reasonable damping to allow you to only measure the head and not the motor. Maybe....
Like others have said, quite clever using a weighted speaker as a contact mic, but I was expecting something like noise-cancelling headphone technology, where a vibrating mass cancels out the vibration of the tool in real-time 😄
nice idea to try and error with pulsing at the same frequency that the imbalance vibration, delay represent angle of rotor and amplitude power in speaker represent a easy variable weight! thks man
@@sebastienl2140 probably you need a tach pulse to timely find the high vibration spot then feed a cancelling waveform with the speaker exactly opposite phase
You may want to check over by the motor with the fan stalled to see if you are picking up magnetic hum from the motor windings. Also the speaker probably only outputs vibration that coincides with the line of the voice coil movement. Nice experiment and thanks very much !!
dynamic balancers often work using a strobe light, trigger a strobe of light on the highest peak value, then you mark the shaft with degrees and turn in on, the strobe will make the exact location of the peak imbalance visible. a possibly better rig would be using a piezo speaker as a sensor supporting the shaft, that would be a much more precise way of doing it
As I understand, with a fan blade, or perhaps a cutting tool, the vibration that happens is a result of the sum of the imbalance of center of mass, and imbalance of the work done by each blade. Measuring vibration and compensating rotational balance may work better when the blades are only cutting air. There may be only so much you can do when the blades are cutting wood. For your jointer, you may have to confirm balance of sharpness and depth of cut, then finesse with vibration reduction with center of mass balance technique. Mostly conjecture.
You might use a speed control to vary the speed of the thing you're trying to minimize vibration. How ever bad something vibrates, there's likely a slower speed that it's vibration is more dramatic and easy to detect. This is where the rotating mass' holder is sympathetic with the frequency of vibration caused by imbalance. It should make your detection mechanism much more sensitive. However smooth things get at this speed, they'll be that smooth or smoother at higher speeds.
I have an old computerized tire balancer that works in a similar fashion. However, it first calculates static balance using a tachometer. With a horizontal axis, the wheel will accelerate when the heavy side is turning down, and decelerate when it is on the up. Back to your device, you could improve your trial and error approach by connecting the vibration trace with something that correlates to the phase of the wheel. I would suggest putting on a strip of reflective tape and using a laser and photoresistor to pick that up as a second trace on your scope.
yes, I keep getting this comment over and over. I didn't do it, not becasue I didn't think of it, but because it wasn't worth the development effort for what I wanted to accomplish.
In(genius) as usual. But there is still the challenge of dealing with the heavy, high sped, spinning jointer head. Flashing knives! How will you attach weights? Eagerly awaiting the next video in this series. But stay safe.
It's very interesting, not sure if an accelerometer would be more adapt to this test, or the transducer is better... Maybe they can be used together to increase accuracy. Cheers!
if it comes to balancing the blade thing you probably will either have to balance it every time you replace a bit but you could probably do it with different length screws because i doubt you can shim it
I wonder if you could make a small preamp circuit and run the speaker output to a microphone input of a phone. I know there are spectrum analyzer apps for phones, so it might be the cheapest possible solution.
I wonder if you could add reflective tape 180 degrees out and somehow block one then pick up when the other passes a sensor. Then when plotted with the 2nd channel, would it tell you approximately where the heavy spot is in relation to the tape?
With a light sensor or a construction of tape and a piece of wire, you could measure the phase and then find the right spot to balance with less trial and error. I'm really looking forward to the next video!
yes, I keep getting this comment over and over. I didn't do it, not becasue I didn't think of it, but because it wasn't worth the development effort for what I wanted to accomplish.
Don't forget the motor is an electromagnet, and the speaker is also to some extent a receiver of magnetic fields. Some of that hum on the motor end could be electromagnetic coupling rather than actual vibration.
Brilliant idea using a speaker as a sensor. I guess the coil moves up/down relative to the permanent magnet mounted in the speaker chassis. And this movement induces a voltage that reflects the amount of coil velocity/direction relative to the magnet. Is the coil the stationary part, and the magnet the moving part? If so, you could possibly magnify the readings by adding some mass to the coil/speaker cone (it stays more stationary, increasing the relative motion to the magnet).
The coil is the stationary part here. Ideally I'd push the speaker membrane against the shaking object, but I can't mount the speaker by the membrane. But this works ok for frequencies below maybe 300 hz.
Perhaps some of the guesswork could be eliminated by syncing the scope to the rotational position of the shaft. A little dab of white or black paint on the perimeter and an opto-sensor could detect that. You would also have to determine the phase relationship between the heavy side of the shaft and the 'scope reading. With no electrical load on the speaker the scope voltage would be related to the velocity of the speaker cone (relative to the speaker frame), which is the first derivative of the position of the speaker cone. The latter is proportional to the acceleration of the speaker frame (due to the spring-like nature of the cone mount). If I got that right the scope is showing the first derivative of the acceleration. This might be more trouble than the trial-and-error approach for a one-off job though.
yes, I keep getting this comment over and over. I didn't do it, not becasue I didn't think of it, but because it wasn't worth the development effort for what I wanted to accomplish.
I wouldn’t mind a video about how exactly you connect your various probes to your computer and what software you use for this oscilloscope function. Might seem basic to you, but I haven’t seen anything on UA-cam about these things.
Really interesting, would try it out myself, why thought you add a metal washer into the speaker? Not to peak up voice/noise? Isn't that blocking the membrane?
I suspect when measuring the motor side you were not seeing motor vibration but rather inductive coupling from the core of the motor to the voice coil of the speaker.
I’m a vibration analyst by trade and your method is spot on! The only additional thing we do with the “test” weight information is build a vector diagram on a polar plot to determine where the optimal final weight should be located. Then there is some math to determine the weight amount but that’s just some simple cross multiplication. Doing this all with a speaker and o’scope is just downright genius.
Spot on, search for graphic balancing method. You can solve your needed weight and place by doing 3 runs while moving the test weight in 120° increments around. The place your three influence circles meet gives you both the place and amount of correction weight.
Using a speaker as a velocimeter is genius!
Search for this on UA-cam:
3 point balancing method
What is your preferred method to apply an encoder? Or does it even matter?
@@G5Ckxew for the 3point Methode, you don't need the phase angle, for vector Methode the phase angle of the vibration is needed that way you only need to do a single test run with a test weight to get the correction weight and position. Getting the phase angle or a reference is done by a keyphasor, or a laser module that detects a ref mark. As a DIY solution you would need anything that gives you a signal on the scope exactly once per rev, always at the same point.
Cross multiplication, is that the one listed by the same name on Wikipedia (I was initially assuming the cross product instead)?
Really interesting to see a speaker used that way. All you need is a sufficiently capable rotary encoder for position tracking and you've nearly caught up to industrial impeller balancing systems.
It is! I guess you could use a voltmeter instead of a graph, but it would take more patience.
I think he wouldn't even need a rotary encoder.
a simple photo diode to measure a fixed point per revolution should be sufficient for home application.
With a dual Chanel scope he could overlay speaker and "zero point" and from there calculate the angle where mass needs to be taken off.
@@fabianrudzewski9027 A simple 5v hall effect sensor would also be a good choice when working with metal objects.
phase of output signal from speaker is a non linear function of frequency, it can be compensated but it's not linear
There are solutions that don't need to know position. Some Balance Technologies machines just have a strobe light so you can point it at the rotating object and it will freeze the point where weight needs to be added or removed. And with the amplitude of the vibrations you can set the correction radius and readout the grams required. This was all old school analog circuitry, too.
This is a pretty clever method! Thanks for sharing with us.
Mind blown. It would never occur to me that there was an issue, let alone following such a logical approach to fixing it.
Wow Matthias! You never cease to amaze me! The way you combine intuitive understanding with real-world electronic know-how is clever and clear. When I was a kid and built slot-car motors, (circa 1964) I just chucked the armature into a drill, spun it, and lightly touched it with a magic marker. Then wherever the marks were thickest, I drilled a small hole in the armature; Reversed the armature, and did the same thing from the other side. That was as close to dynamic balancing as a cash-strapped kid could get, But it worked.! Of course I didn't have access to these electronic measuring devices then, so that method was a tedious iterative process. Even now, I would have never thought of using a speaker to generate a readable pulse into an oscilloscope. Brilliant idea!
With two speakers configured orthogonally and two channels on an oscilloscope, Lissajous curve could be displayed. You could directly find the vibration vector. If you further know the phase of the motor, you could even calculate and display the exact angle at which you should add or remove weight.
I love the low cost approach to everything you share. It is so refreshing in a world of "Look at all this cool stuff I got for free to show you, please go buy it".
I confused dynamic with active. I thought you were going to feed back the vibration into the weighted voice coil to damp them that way (with destructive interference). Your way makes way more sense. Particularly for the heavy cutter. Looking forward to the conclusion!
Yhea. That is how I interpreted the descripton.
That would be really freakin cool. Like an engine with a software ballance shaft.
That impresses me the most in your videos is the didacts you have. Even when I know nothing about what you are experimenting with I'm always able to understand the concept.
You would be a perfect college professor, for sure!
Wow, I'm blown away by the analysis and the use of the speaker is ingenious. Thank you for the video.
You could possibly get similar results using a 'vibration meter' app on your phone (for those of us without an oscilloscope).
I dynamically balanced the drive pulley on my metal lathe by taping an old phone to the headstock and measuring the vibrations.
Me too, but I want an oscilloscope, its probably much more accurate. Vibration analyzers at mm/s^2 range are very expensive as well 😢. I balanced my angle grinder with my phone lol this is amazing.
If you have a small mirror and a laser pointer laying around, you could probably do this completely analog. Put a piece of paper on a wall far away to mark out the sweep of the laser. The added benefit is electrical noise would have no effect.
Mirrors have mass
I use this method to dynamically balance props on drones.
@@MadScientist267 The mirror goes on the static part so makes very little difference.
That's freaking cool
@@DullPoints I have a really old vid on my channel if you want to see how it's done. 🙂
I never knew how balancing could be done before this. Fascinating video.
Man! That was a super cool video! I am astonished by your method, as always. I dream of being "your kind of smart" but the truth is, I'm a little lazy about trying to be. However! You explain the details so perfectly that I sit here and think to myself "That makes total sense. Of Course that's how that works!". I truly do understand and follow your explanations. I just, would never have gotten there myself. "That" is what perplexes me sometimes. That's the fun in it all... Right?! I have thoroughly enjoyed every one of your experiments over the years. The mice experiments were lots of fun! I appreciate your dedication to not only the true and sometimes brutally honest results but also to the storytelling arc. Thank you.
This looks like a job for a piezo transducer! Also thanks for the demo. I am about to buy one of those blowers and this will help a lot if it's unbalanced.
This is quality content.
I like it a lot.
Balancing fans and rotating things with a cheap diy friendly way and increasing their lifespan, lowering the noise and vibration is a really good thing.
The earlier video you had with balancing a fan really helped me right after i bought a box fan (brand new) it wash shaking violently and dancing over the floor, but after a few minutws balancing it it became very silent and no noticeable vibration
On the driven end of the rotor I believe you were seeing a central frequency of 60hz and sidebands each side, offset by 10hz. I suspect it was caused by geometrical angular/offset misalignment between the motor and fan rotors if they are rigidly coupled together. Very smart use of the speaker to measure the waveform!
I proposed a system like this for using a speaker and diaphragm weight to dynamically balance motors about 20 years back on the electronics forums.
The one thing you missed which would have been an improvment would be to place the whole assembly on an elastic suspension, which could have been as simple as a springy foam block. Your setup was damping the vibration into the table which reduced the measurement amplitude and accuracy.
Very cool project! 👍
I'm very excited to see the process done with your planer!
A guy at work once had a fan very similar to that where a blade had busted off and was no where to be found. He was trying very hard to fabricated a thin blade to replace it when I suggested we just count how many blades there were and snap one off from the opposite side. We counted and it was all good and snapped the opposing right off and it ran sweet.
Beautiful and insightful work, congratulations!
The 69-ish Hz you described at timestamp 5:08 is possibly due to the aliasing between the 60Hz network frequency and the 50-ish Hz of the rotational speed.
The aliasing or beating frequency generates a pair of side bands (+10 and -10 Hz)
The 10Hz can be also observed at the 10Hz region of the spectra and as a sideband of every 60Hz harmonics.
If the system responds well to higher frequencies, you may also find a series of 100-110-120-130 -140Hz and so on.
this is also what creates the really cool "whirr" sound when induction motors start up!
@@polymetric2614 , the “whirr” is outstanding when you change speeds of a 2/4 poles motor 😊😊
Nice and simple approach to solving this complex problem. I like how you use a speaker instead of a fancy pant accelerometer.
Have you thought about marking a zero position with reflective tape and use an optical pick up to mark the zero position on your oscilloscope? That would make finding the “location” of the imbalance easier than trial and error.
Once you figure out the delays from vibration to the speaker pickup, however adding a weight should show that easily enough.
Absolutely the best video i have seen in the past thousands
My God, dude. Your methods are simply inspired! What an awesome approach. I've yet to see a video from you where I did not learn something. Bloody brilliant.
A simple practical demonstration like this is great for a lay person like myself
This is fascinating. Definitely beyond the scope of any of my projects but it's interesting to read all the comments from all these knowledgeable viewers.
Very interesting! I'm thinking of combining this to Denis' approach (channel "the GADGETS playlist") to dynamic balancing, ie, using a microphone (or speaker) to listen to the vibrations and reference it using a photocell to figure out where the weight should go - and perhaps combine it with the Y approach, ie, instead of adding weight at position 0, add it (for instance) at positions -3 and +3.
Denis uses an accelerometer to measure the vibrations, not sure which option would be more accurate.
Your video was very helpful. I was able to balance 3D printed impeller with cheap oscilloscope
*checks notes* Dynamic Systems Laboratory, Junior year! Super clever implementation of the 'accelerometer!' love your videos, WoodGear.
Fascinating approach to dynamic balancing! Cant wait to see the Shelix cutter head balancing :-)
Matthias blows my mind every time I watch one of his videos.
in the 1980s the company I worked for looked at analysing vibration from the heavy industrial machinery it made in an attempt to give the operator maintenance information much earlier than his ears could detect. It actually worked quite well but few end users would pay a worthwhile price for it.
There's a mentality of "don't fix what isn't broken", which seems to stop them from considering that "Everything is gradually breaking, so it WILL break at some point. Fix it when it's most effective to do so".
I've worked with a number of factories and suggested that premise as well. Advising that a lot of their stoppages and repair costs could be reduced, AND would offset the cost of any system upgrades by an order of magnitude. Not a single one was willing pay for it except where I just did it anyway, and demonstrated it, and billed them.
With this technique, you have to make sure the resonant frequency of the speaker is not the same as the DUT.
Otherwise the vibrations will look a lot stronger than really is.
I'd imagine a speaker having very little resonance (given their purpose)
But, doesn't it change that, if you add weights to the diaphragm?
@@AmirASD The more mass, the more inertia, the lower the frequency of resonance.
Same thing with inductance. The more henrys, the lower the resonant frequency.
I suspect that adding an electrical load on the speaker (perhaps a resistor equal to the speaker's nominal impedance) would damp any resonance.
Does it really matter? The frequency response will be nonlinear but we only care about relative changes. As long as the DUT frequency remains constant I think it would still work fine.
I could watch this a billion times and still have no idea how you did that.
Very cool. You never cease to impress, Matthias.
I have some 3D-printed flywheels on small RC brushless motors that I need to balance, and this method looks like it would be ideal and uses stuff I already have on hand. Thanks for the tip!
Very clever! I imagine that squirrel cage fans are pretty tolerant of imbalances given that they all collect dust, and any dust buildup will likely throw off even the most carefully balanced fan. However, a balanced fan will give off less fan noise and that's always a blessing. Good show on the use of low cost test gear to analyze a problem!
The dust will be evenly distributed around the evenly sized and evenly spaced blades of a squirrel cage fan, so it doesn't affect the balance much. Initial balancing really protects the machinery and makes it last longer. There are industrial balancers that balance absolutely everything in, for example, a pulp mill. I once pulled a single leaf out of an automobile's squirrel cage blower that was making a horrible noise and that fixed it perfectly.
WOW!!! Way cool. Get HOT on balancing that shaper/planer! Im very interested! Looks like a very cool project!
30 years ago we use to characterize disk drive motor samples with a similar set up sensitive and expensive B&K gear. Today, you can pull the 9 axis of freedom board off your drone and couple that with your favorite eval board (pi, arduino, stm32) and throw together a pretty darn good sensor system. Couple that with all the free firmware that deals with static and dynamic behavior and you will know more about anything you want to measure. The vibrations in the handle of your weed eater, lawn mower, shovel, hammer or vibrations in the motors... you could spend your whole summer on this..
KWR, Clifton Precision?
That's a pertty interesting idea to use a speaker as an accelerometer. Ive seen them used as microphones, but never for low frequency. It's not a calibrated measurement, but for some quantitative analysis it's more than good enough.
Thank you very much for this very interesting video on DIY balancing.
my father has a balancing machine for big rollers and ducts. Althought its a little bit more sophisticated with a control computer etc., it messures the vibration on both sides of the tube (its laying there on bearings) and always knows at what position the duct is with the help of a light sensor and a glued on tag, after you run it it tells you the spots that need weight automatically.
Fantastic idea without expensive equipment, love it!
Very impressed with everything you do. Thanks for sharing.
Very interesting topic and clever idea of using a speaker. Nerdy stuff as usual ;)
As a Vibe Tech by trade, what you got at the Driven End where the motor is, was Sidebanding. Looking at the data you got, I'd say that you might have an eccentricity present, however as for 60hz being the Center Frequency, I can't currently figure that part out.
Ac induced noise on the speaker would be my guess (hum)
i might actually use this method. ive made fans and fast rotation devices before and many times simple field balancing techniques will not suffice. it also shows the thousands of uses for signal analyzation of electro-mechanical components using an oscilloscope.
Clever boots. I look forward to the joiner balancing video.
I would love to see your take on the DML speakers.
Thanks for this, quite interesting.
So about the 120Hz humming:
A transformer iron core expands and shrings with the magnetic forces involved. An electric sinus signal means a plus and another negative wave or going from 0 to max to 0 to min and back to 0. So that's why a transformer core hums with double the frqeuency of the electric signal.
In most (if not all) parts of Europe it's a 100Hz humming, because we use 50Hz electrically.
With the addition of either an optical sensor tach, on the scope, the phase of the vibration relative to the shaft position, can be used to find the offset and amplitude of the balance for accurately finding the weight and position for the correction. This is how a tire shop wheel balancer works.
That is a pretty cool idea. I really appreciate your sharing that.
Thank you so much for this. So well explained and demonstrated. The only thing I'd like more information on is the weights added to the speaker. How much? The purpose? Etc.
Imterestimg video. In the past I've wondered if there's some way to balance a ceiling fan besides by trial and error. I have to believe this approach could be adapted for that.
This should be on your main channel. Clever idea.
it's not woodworking by any stretch
@@matthiasrandomstuff2221 Just add a few shots making the bracket et viola! Great video, thanks.
That's cool, would like to see further improvements on this concept.
Brilliant! Could I suggest using a piezo transducer instead of the speaker. It will likely be much more responsive and less likely to add sympathetic impulses as a piezo has zero moving parts(no coil, cone, etc..). Also, if possible within reason, when you attempt to balance your cutter head, make a stand for it and spin it with a damper of some sort between motor and head. Won't have to be a strong motor, and perhaps something as simple as latex tubing would work to couple the motor and head and provide reasonable damping to allow you to only measure the head and not the motor. Maybe....
my first thought was pizo as well, but I didn't ahve one handy.
@@matthiasrandomstuff2221 , a door buzzer is a cheap and readily available piezo disc.
Wow!! I never would have thought to check the balance of these things. Makes perfect sense though! I love these videos!
👏👏
it was close enough for using already. I just wanted to see if my method works.
Like others have said, quite clever using a weighted speaker as a contact mic, but I was expecting something like noise-cancelling headphone technology, where a vibrating mass cancels out the vibration of the tool in real-time 😄
nice idea to try and error with pulsing at the same frequency that the imbalance vibration, delay represent angle of rotor and amplitude power in speaker represent a easy variable weight! thks man
@@sebastienl2140 probably you need a tach pulse to timely find the high vibration spot then feed a cancelling waveform with the speaker exactly opposite phase
Amazing! Anyday I will try to use this method!
That's a brilliant idea , seat of your pants engineering . Food for thought , thanks for posting the vid.
You may want to check over by the motor with the fan stalled to see if you are picking up magnetic hum from the motor windings. Also the speaker probably only outputs vibration that coincides with the line of the voice coil movement. Nice experiment and thanks very much !!
that's what I was demonstrating in the video
dynamic balancers often work using a strobe light, trigger a strobe of light on the highest peak value, then you mark the shaft with degrees and turn in on, the strobe will make the exact location of the peak imbalance visible.
a possibly better rig would be using a piezo speaker as a sensor supporting the shaft, that would be a much more precise way of doing it
As I understand, with a fan blade, or perhaps a cutting tool, the vibration that happens is a result of the sum of the imbalance of center of mass, and imbalance of the work done by each blade.
Measuring vibration and compensating rotational balance may work better when the blades are only cutting air. There may be only so much you can do when the blades are cutting wood. For your jointer, you may have to confirm balance of sharpness and depth of cut, then finesse with vibration reduction with center of mass balance technique. Mostly conjecture.
You might use a speed control to vary the speed of the thing you're trying to minimize vibration. How ever bad something vibrates, there's likely a slower speed that it's vibration is more dramatic and easy to detect. This is where the rotating mass' holder is sympathetic with the frequency of vibration caused by imbalance. It should make your detection mechanism much more sensitive. However smooth things get at this speed, they'll be that smooth or smoother at higher speeds.
Fine work by you and exposition for us.
You made an electronic stethoscope to diagnose an illness with that fan :-) nice work.
I tried with a piezo speaker and some weight glued to it. Worked great also for high frequency vibrations.
Interesting stuff. I always wondered how wheel ballancing machine got thier data. And crank shaft ballancers.
I have an old computerized tire balancer that works in a similar fashion. However, it first calculates static balance using a tachometer. With a horizontal axis, the wheel will accelerate when the heavy side is turning down, and decelerate when it is on the up.
Back to your device, you could improve your trial and error approach by connecting the vibration trace with something that correlates to the phase of the wheel. I would suggest putting on a strip of reflective tape and using a laser and photoresistor to pick that up as a second trace on your scope.
yes, I keep getting this comment over and over. I didn't do it, not becasue I didn't think of it, but because it wasn't worth the development effort for what I wanted to accomplish.
You are a genius 👏. Other method may be using a high speed MEMS acelerometer (1 khz or more) with an esp32 or Stm or RPiPico.
In(genius) as usual. But there is still the challenge of dealing with the heavy, high sped, spinning jointer head. Flashing knives! How will you attach weights? Eagerly awaiting the next video in this series. But stay safe.
Excellent idea! 👍
How about using two identical speakers with weights at every spindle end and dual channel oscilloscope?
Interesting, maybe a second oszilloskope channel could be added to (maybe optically) get a pulse from a known position on the rotor.
I would give this a try with a piezo transducer.
It's very interesting, not sure if an accelerometer would be more adapt to this test, or the transducer is better... Maybe they can be used together to increase accuracy. Cheers!
What!?! I was so into this problem and how much it would work/IF it would work for the jointer. That was a great cliffhanger. Argh!
Brilliant, Matthias! 😃
Looking forward to see the process on the jointer!
Stay safe there with your family! 🖖😊
I'm excited for the followup!
if it comes to balancing the blade thing you probably will either have to balance it every time you replace a bit but you could probably do it with different length screws because i doubt you can shim it
Try to connect "phase meter" to the second channel of the oscilloscope, using optical or hall sensor
I wonder if you could make a small preamp circuit and run the speaker output to a microphone input of a phone. I know there are spectrum analyzer apps for phones, so it might be the cheapest possible solution.
I thought about using an audio input, but the problem is, audio inputs may adjust the gain if necessary, and that messes with my measurements
I wonder if you could add reflective tape 180 degrees out and somehow block one then pick up when the other passes a sensor. Then when plotted with the 2nd channel, would it tell you approximately where the heavy spot is in relation to the tape?
With a light sensor or a construction of tape and a piece of wire, you could measure the phase and then find the right spot to balance with less trial and error. I'm really looking forward to the next video!
yes, I keep getting this comment over and over. I didn't do it, not becasue I didn't think of it, but because it wasn't worth the development effort for what I wanted to accomplish.
@@matthiasrandomstuff2221 Sorry, it's not easy to keep track of similar comments. Maybe I should have refreshed the page before commenting.
Now THAT is really neat. Thanks for sharing.
Maybe a variable speed drive for the jointer? Keep the speed down until you are comfortable with running it faster.
Don't forget the motor is an electromagnet, and the speaker is also to some extent a receiver of magnetic fields. Some of that hum on the motor end could be electromagnetic coupling rather than actual vibration.
Brilliant idea using a speaker as a sensor. I guess the coil moves up/down relative to the permanent magnet mounted in the speaker chassis. And this movement induces a voltage that reflects the amount of coil velocity/direction relative to the magnet. Is the coil the stationary part, and the magnet the moving part? If so, you could possibly magnify the readings by adding some mass to the coil/speaker cone (it stays more stationary, increasing the relative motion to the magnet).
The coil is the stationary part here. Ideally I'd push the speaker membrane against the shaking object, but I can't mount the speaker by the membrane. But this works ok for frequencies below maybe 300 hz.
Perhaps some of the guesswork could be eliminated by syncing the scope to the rotational position of the shaft. A little dab of white or black paint on the perimeter and an opto-sensor could detect that.
You would also have to determine the phase relationship between the heavy side of the shaft and the 'scope reading. With no electrical load on the speaker the scope voltage would be related to the velocity of the speaker cone (relative to the speaker frame), which is the first derivative of the position of the speaker cone. The latter is proportional to the acceleration of the speaker frame (due to the spring-like nature of the cone mount). If I got that right the scope is showing the first derivative of the acceleration.
This might be more trouble than the trial-and-error approach for a one-off job though.
yes, I keep getting this comment over and over. I didn't do it, not becasue I didn't think of it, but because it wasn't worth the development effort for what I wanted to accomplish.
I wouldn’t mind a video about how exactly you connect your various probes to your computer and what software you use for this oscilloscope function. Might seem basic to you, but I haven’t seen anything on UA-cam about these things.
see my video about the USB scope that I use
In this case just using the microphone input and some sound-oscilloscope software would have been enough.
I don’t always get what you’re actually doing but it always looks so cool
Really interesting, would try it out myself, why thought you add a metal washer into the speaker? Not to peak up voice/noise? Isn't that blocking the membrane?
You are a genius gentleman!
Very informative video 👌👌
You should probably use an ADXL345 accelerometer to analyze vibrations and armonics. Speaker with a mass has his own elastic resonance.
Mind blown. How do I always end up learning something in every one of your videos?
That's just brilliant diy!
Okay this is actually awesome. I've actually "Saved" this video... which I never do. 😂
Clever guy, another great video
I suspect when measuring the motor side you were not seeing motor vibration but rather inductive coupling from the core of the motor to the voice coil of the speaker.