How to dynamically balance EDF unit with a cell phone - Page 2 - RC Groups
Jun 30, 2010, 01:15 PM
Ugly planes do not fly
Quote:
 Originally Posted by jzuniga Ok this may sound like some dumb questions but... 1) What would an equivalent "app" be? I have an iphone and all I found was an app called "vibration" that takes X, Y, Z readings? there is some kind of "sec/div" value also given? would this work? 2) Are the "vibration" values in your diagram, the average of the X, Y, Z readings? I see in the photo of your droid it too looks like it's giving these readings, like an oscilloscope? Thanks for your help...;-) Z
I am not familiar with Apple market of applications.

To answer your question about data readings: the software provides four columns of data per reading: Time, X, Y, Z. In fact you can ignore time altogether. You are not interested in absolute measurements of the vibration values either, see below why. All you need is two columns out of four with vibration data along axis X and Z.

If you take a closer look at the picture with the phone you'll see that I mounted phone at 45 degrees. This angle "sort of compensates" weight factor of the phone when measurements are done along axis X and Z. You do not need to measure axis Y, because this is movement along the motor shaft.

Now that you have two columns of data, you sum them up as vectors to get resulting force using this formula:

Result force vector = SQRT(X*X + Z*Z)

This will give you one column of data. To get one number from the series of data that can be compared with another series of the data you calculate standard deviation of the series. Standard deviation is volatility or vibration level in engineering terms.

Vibration = StdDev(result force vector)

Now when you have this number you can compare it with number from another measurement.

I hope this explanation helps.
Jun 30, 2010, 03:01 PM
Screw the wind, let's fly!
Quote:
 Originally Posted by _Sergey_ I am not familiar with Apple market of applications. To answer your question about data readings: the software provides four columns of data per reading: Time, X, Y, Z. In fact you can ignore time altogether. You are not interested in absolute measurements of the vibration values either, see below why. All you need is two columns out of four with vibration data along axis X and Z. If you take a closer look at the picture with the phone you'll see that I mounted phone at 45 degrees. This angle "sort of compensates" weight factor of the phone when measurements are done along axis X and Z. You do not need to measure axis Y, because this is movement along the motor shaft. Now that you have two columns of data, you sum them up as vectors to get resulting force using this formula: Result force vector = SQRT(X*X + Z*Z) This will give you one column of data. To get one number from the series of data that can be compared with another series of the data you calculate standard deviation of the series. Standard deviation is volatility or vibration level in engineering terms. Vibration = StdDev(result force vector) Now when you have this number you can compare it with number from another measurement. I hope this explanation helps.
See what ya started?
Jul 01, 2010, 12:13 PM
Ugly planes do not fly

# Calculations example

There were questions about calculations so I created an example. Please see attached picture.

### Images

 Jul 01, 2010, 01:07 PM Registered User Thanks Sergey! Z
 Jul 01, 2010, 01:16 PM Retired C-5 Flight Engineer Awesome! Funny thing.....I fly C-5's for a living and we have this type of "program" on the jet for our engine vibration reporting. We basically run through this type of procedure if we get an engine vib indication...pretty cool!!
 Jul 21, 2010, 12:05 AM Registered User Thank you so much for this tech info! I just bought a Freewing F16. I have a couple of friends who own Freewing SU34's. I mentioned to them I wanted to balance my fan and one said he wouldnt waste his time for the minimal gains it would achieve. This is going to be fun! Last edited by Fubar69; Jul 21, 2010 at 12:16 AM.
 Jul 21, 2010, 12:46 AM Registered User I have to ask this out of curiosity. Although a little more tedious, if you went more granular on the initial check of the motor to impeller vibration, say every 20 or even 10 degree's instead of 45, logic would say that you would get a better result by getting it closer to the spot where vibrations are lowest correct? Yes I know it would be harder and more exacting but I wonder if the increase in efficiency would be worth the trouble?
 Jul 21, 2010, 07:16 AM Registered User This should be a sticky... Questions - In step 2 it looks like there's a normal static balance of the rotor, am I right to conclude that the area of least vibration on the motor is where we're supposed to align the balanced rotor? - I have an HTC Touch Pro 2 running WM 6.5 and it has accelerometers do you know of any sensor apps for this phone? - Any vid of your 380? Would love to hear the sounds of the fans TIA Last edited by erh7771; Jul 21, 2010 at 07:39 AM.
Jul 21, 2010, 11:08 AM
Ugly planes do not fly
Quote:
 Originally Posted by Fubar69 I have to ask this out of curiosity. Although a little more tedious, if you went more granular on the initial check of the motor to impeller vibration, say every 20 or even 10 degree's instead of 45, logic would say that you would get a better result by getting it closer to the spot where vibrations are lowest correct? Yes I know it would be harder and more exacting but I wonder if the increase in efficiency would be worth the trouble?
You are correct about precision. There are also ways how to achieve more exact angle without it being too tedious. To increase precision one extra step is needed. After lowest vibration angle A(min) is identified in step 1 follow step 1a:

Step 1a:
a) Position fan at the angle A(min)-30 degrees
b) Take measurement
c) Rotate fan by +10 degrees
d) Repeat steps b-c 7 times

The result should look like a U-shaped curve with the bottom point is the minimum point of vibration with accuracy of 10 degrees with just extra seven measurements.

Theoretically step 1a can be repeated with new A(min) and +- 10 degrees spread to narrow down the minimum point to a three degree angle. But it may not be practical. First it is much harder to manually rotate fan by three degrees. Secondly there is a measurement error that can be larger than the result making single measurement impractical. There are ways to include this error as a factor in the calculation but this is going to be very tedious job with dozens of extra measurements and may require some sort of automation, where computer or phone controls the throttle, not human.

Jul 21, 2010, 11:36 AM
Ugly planes do not fly
Quote:
 Originally Posted by erh7771 - Any vid of your 380? Would love to hear the sounds of the fans TIA
Here are two videos. First one is the maiden flight with unbalanced motors. Check take off and cruise sound starting at 0:55. Second is after engines are balanced. Don't mind the landing on both videos, it's a coincidence.

Unbalanced fans:
 A380 Maiden Flight (3 min 33 sec)

Balanced fans:
 Airbus A-380 RC Jet (1 min 38 sec)
Jul 21, 2010, 11:58 AM
Ugly planes do not fly
Quote:
 Originally Posted by erh7771 - In step 2 it looks like there's a normal static balance of the rotor, am I right to conclude that the area of least vibration on the motor is where we're supposed to align the balanced rotor?
Mostly correct with exception there is no fan balancing. In fact, if the fan is statically balanced the dynamic balancing won't work. The fan must be disbalanced from the beginning, otherwise it is impossible to find motor's heaviest point. Then heaviest point on the fan aligned agaist heaviest point on the motor.

Quote:
 Originally Posted by erh7771 - I have an HTC Touch Pro 2 running WM 6.5 and it has accelerometers do you know of any sensor apps for this phone?
I am not sure. Check this application if it works on your phone: http://www.lucanatech.com/sensor_insider
 Jul 21, 2010, 02:51 PM Yes..ok..maybe..lol..... Sergey....would this vibration meter work for balancing a fan unit? looks like it is wired for the y-axis? http://www.dimensionengineering.com/...tion_meter.htm -Dave
Jul 21, 2010, 02:53 PM
Registered User
Quote:
 Originally Posted by _Sergey_ Here are two videos. First one is the maiden flight with unbalanced motors. Check take off and cruise sound starting at 0:55. Second is after engines are balanced. Don't mind the landing on both videos, it's a coincidence.
On the second vid the wheels rolling on the asphalt were louder than all 4 EDFs!!!

I wanted a Sprint Evo anyway....sounds like an excuse to buy one if I can't find any sensor software for windows mobile.

Further questions

- The Excel spread sheet, did that come with your software or did you do the calculations and the spider chart yourself?

If you did them yourself are you willing to sell the sheet? 5 bucks?

-"f) Rotate fan to approximately 45 degrees. Direction is not important as far as you rotate in the same direction each run."

Am I rotating the EDF fan shroud 45 degrees from a marked top and then doing another test or am I rotating the EDF rotor 45 degrees then doing another test?

TIA for any input and again thx,

Also, thx for the windows mobile sensor app...helps me wait a lil bit till google comes out with droid 2.x somethin at the end of the year
Last edited by erh7771; Jul 21, 2010 at 03:13 PM.
Jul 21, 2010, 03:03 PM
Ugly planes do not fly
Quote:
 Originally Posted by shhhhh! Sergey....would this vibration meter work for balancing a fan unit? looks like it is wired for the y-axis? http://www.dimensionengineering.com/...tion_meter.htm -Dave
I think it would. The only problem is how to collect vibration data for analysis? I think it's possible to come up with a method to measure vibration at various RPM levels, enter numbers manually and go from there?
Jul 21, 2010, 03:07 PM
Ugly planes do not fly
Quote:
 Originally Posted by erh7771 - The Excel spread sheet, did that come with your software or did you do the calculations and the spider chart yourself? If you did them yourself are you willing to sell the sheet? 5 bucks?
Haha! I did it myself and it's free. Enjoy!
Last edited by _Sergey_; Jul 21, 2010 at 03:32 PM.