SparkFun Electronics Forum

I've been looking to find some type of directions to build a dynamic balancing machine for electric ducted fan assemblies. I've seen a few people make them out of speakers and piezo discs. Some even use cell phones to try and balance them with sound. I was thinking of more of something electronic using accelerometer or something. Has anyone built something like this or have any suggestions on doing so?

I'd like to build something like these but seems it could be easier.
http://www.bobsworkshop.co.uk/outdoor-f ... n-balancer
http://www.turbinemuseum.de/Gasturbines ... _tool.html

Interesting project. Seems the device in the 2'nd link uses piezo elements to sense the side to side motion induced by the imbalance. Am I correct in this ? I think the 1'st machine senses in the up/down direction (how exactly ?), though I'm not sure how/if the differential sensing used in #2 is done ? I note that auto tire/wheel balancers I've used do some other sensing so that weights can be applied on both the inner and outer edges of the wheel. Is this additional step needed or desired in your usage ? I'm not sure how to do it (I have a guess) but "we" might as well get all the reqs established up front. Do you have some idea of the size of the residual imbalance you want ?

FWIW, if I'm understanding the techniques used, then yes, I think accelerometers might work also ... assuming the accelerations would be large enough to be accurately sensed by the affordable devices now available. But if simple, cheap piezo's can be used, why use accelerometers ?

EDIT : the wiki on balancing machines had a good link to understand the very basics. Go to the FAQ here. Seems like a single plane balance would suffice.

Another issue I can think of using accelerometers is the speed at which you can get readings.
The sampling rate needs to be high enough to resolve the rotational position of the un-balance.

Definitely an interesting project so please keep us posted on the progress.

Well I'm trying to balance small electric ducted fans and I just through the x-y-z accelerometer sold here would be an easy fit but of course I have no idea what I'm doing. I've built some simple electronic devices and made my own pcb's so i figured I could tackle this but may be over my head. I assumed you could take a reading with the motor off on the accelerometer and then if you put a mark on the spinner and take readings at that exact time/spot you could calculate the vector of the imbalance??

barkster wrote:I assumed you could take a reading with the motor off on the accelerometer and then if you put a mark on the spinner and take readings at that exact time/spot you could calculate the vector of the imbalance??

The imbalance should result in a oscillating motion when sensed in any 1 axis orthogonal to the spin axis. An X/Y accelerometer, in theory, mounted to the housing should give you this in both sensing axes. Not sure how that helps (vs 1 axis) but there it is. The problems still remain :
1) You need to correlate the sensed accel to the position of the rotor. The balancer in one of you links used a strobe and a mark to do this. What would you do ?
2) The accelerations need to be large enough to be sensed accurately and the fan housing must be held to some fixture. That's some mechanical design and too stiff a connection btw the two will diminish the accelerations ... perhaps too much to be sensed by what's affordable for accelerometers. You'll need to decide how small a residual imbalance is good enough and what acceleration that will produce given your mechanical design and whether accelerometers can sense that level accurately enough.
3) The accelerometer must sample quickly enough to be able to resolve a sufficiently small rotational angle. Depending on what RPM you balance at, that could be a problem.

No doubt there are other challenges to be met but these will do for a start.

sounds a bit over my head. I'll just buy the $200 one the guy from uk sells and see how he's doing it though I've heard lots of complaints of his being hard to use. That's why I was thinking an accelerometer would work better but guess not. Thanks for the help