Balancing during the assembly

Firmware/software/electronics/mechanics
TheFrog4u
Expert
Posts: 113
Joined: Fri Feb 08, 2013 6:59 pm
Location: Bremen, Germany

Re: Balancing during the assembly

Post by TheFrog4u » Sat May 18, 2013 11:28 am

DesTinY wrote:Maybe you can post a capture from the balancing self test without the propellers too.

Code: Select all

SYS: Crazyflie is up and running!
SYS: Build 40+:601803a36702+ (tip) MODIFIED
SYS: I am 0x872023395356854966EFF54 and I have 128KB of flash!
IMU: MPU6050 I2C connection [OK].
IMU: HMC5883 I2C connection [OK].
IMU: MS5611 I2C connection [OK].
MPU6050: Self test [OK].
MPU6050: Propeller on M1 [OK] Condition: 114 percent 
MPU6050: Propeller on M1 [OK]. X_max: 0.542, Y_max: 0.509, Z_max: 0.790 
MPU6050: Propeller on M2 [OK] Condition: 104 percent 
MPU6050: Propeller on M2 [OK]. X_max: 0.541, Y_max: 0.553, Z_max: 0.798 
MPU6050: Propeller on M3 [OK] Condition: 114 percent 
MPU6050: Propeller on M3 [OK]. X_max: 0.540, Y_max: 0.552, Z_max: 0.722 
MPU6050: Propeller on M4 [OK] Condition: 104 percent 
MPU6050: Propeller on M4 [OK]. X_max: 0.547, Y_max: 0.519, Z_max: 0.859 
MPU6050: Propeller Balancing Self Test [OK] 
HMC5883L: Self test [OK].
MS5611: Self test [OK]
after balacing:

Code: Select all

SYS: I am 0x872023395356854966EFF54 and I have 128KB of flash!
IMU: MPU6050 I2C connection [OK].
IMU: HMC5883 I2C connection [OK].
IMU: MS5611 I2C connection [OK].
MPU6050: Self test [OK].
MPU6050: Propeller on M1 [OK] Condition: 71 percent 
MPU6050: Propeller on M1 [OK]. X_max: 0.680, Y_max: 0.820, Z_max: 1.985 
MPU6050: Propeller on M2 [OK] Condition: 74 percent 
MPU6050: Propeller on M2 [OK]. X_max: 0.843, Y_max: 0.753, Z_max: 1.27 
MPU6050: Propeller on M3 [OK] Condition: 79 percent 
MPU6050: Propeller on M3 [OK]. X_max: 0.656, Y_max: 0.668, Z_max: 1.791 
MPU6050: Propeller on M4 [OK] Condition: 63 percent 
MPU6050: Propeller on M4 [OK]. X_max: 1.7, Y_max: 0.942, Z_max: 1.60 
MPU6050: Propeller Balancing Self Test [OK] 
HMC5883L: Self test [OK].
MS5611: Self test [OK].



TheFrog4u
Expert
Posts: 113
Joined: Fri Feb 08, 2013 6:59 pm
Location: Bremen, Germany

Re: Balancing during the assembly

Post by TheFrog4u » Sat May 18, 2013 12:18 pm

after a few more tries on propeller setup:

Code: Select all

SYS: Crazyflie is up and running!
SYS: Build 40+:601803a36702+ (tip) MODIFIED
SYS: I am 0x872023395356854966EFF54 and I have 128KB of flash!
IMU: MPU6050 I2C connection [OK].
IMU: HMC5883 I2C connection [OK].
IMU: MS5611 I2C connection [OK].
MPU6050: Self test [OK].
MPU6050: Propeller on M1 [OK] Condition: 92 percent 
MPU6050: Propeller on M1 [OK]. X_max: 0.629, Y_max: 0.610, Z_max: 1.119 
MPU6050: Propeller on M2 [OK] Condition: 84 percent 
MPU6050: Propeller on M2 [OK]. X_max: 0.620, Y_max: 0.736, Z_max: 1.16 
MPU6050: Propeller on M3 [OK] Condition: 81 percent 
MPU6050: Propeller on M3 [OK]. X_max: 0.646, Y_max: 0.696, Z_max: 1.480 
MPU6050: Propeller on M4 [OK] Condition: 85 percent 
MPU6050: Propeller on M4 [OK]. X_max: 0.656, Y_max: 0.745, Z_max: 0.866 
MPU6050: Propeller Balancing Self Test [OK] 
HMC5883L: Self test [OK].
MS5611: Self test [OK].
I think this is near to perfect now.. BTW: It also depends a bit on how deep one pushes the propellers in. I got the feeling that the propellers who need more force to push in tend to have less vibrations.

DesTinY
Member
Posts: 97
Joined: Sat Feb 09, 2013 5:18 pm
Location: Bünde, Germany

Re: Balancing during the assembly

Post by DesTinY » Tue May 21, 2013 9:08 am

TheFrog4u wrote:I think this is near to perfect now.. BTW: It also depends a bit on how deep one pushes the propellers in. I got the feeling that the propellers who need more force to push in tend to have less vibrations.
Thanks for testing. And yes I got the same feeling that the propeller needs more force to push in tend to have less vibrations. :D

@Tobias
i tried to move the balancing self test from start up to parameter called. Bitbucket Link
The balancing test started by setting the parameter values (so far so good), but :( after the test Bitucket Link finished the CF disconnects and the Red/Green LEDs turns off. After a few sec. the red flashes normally and I can reconnect and get the last Debug message.
Is is possible that a watchdog occurs and the CF reboots?! I still have no jtag. :roll:

I'm working on a UI Tab in the Client to start the balancing self test from here (all motors and single motor as well) and get bar graphs as feedback.

tobias
Bitcraze
Posts: 1919
Joined: Mon Jan 28, 2013 7:17 pm
Location: Sweden

Re: Balancing during the assembly

Post by tobias » Thu May 23, 2013 5:26 pm

Will be fun testing you new functionality. I'll try to find some time tomorrow to look at it.

DesTinY
Member
Posts: 97
Joined: Sat Feb 09, 2013 5:18 pm
Location: Bünde, Germany

Re: Balancing during the assembly

Post by DesTinY » Wed May 29, 2013 12:15 pm

Propeller balancing self test with UI. :D :lol:
Preview:

Image

Release coming soon...

foosel
Expert
Posts: 175
Joined: Sat Feb 02, 2013 9:59 pm
Location: Obertshausen, Germany
Contact:

Re: Balancing during the assembly

Post by foosel » Wed May 29, 2013 12:32 pm

I really am missing gplus' "+1" button here for posts like DesTinY's just above....

Awesome work!
Image
Also: AR.Drone 2.0 (RC-enabled thanks to self-soldered MiruMod) and Hubsan X4 H107

tobias
Bitcraze
Posts: 1919
Joined: Mon Jan 28, 2013 7:17 pm
Location: Sweden

Re: Balancing during the assembly

Post by tobias » Wed May 29, 2013 12:50 pm

Nicely done!

So you got around the bug? Sorry I havn't had time to look at it yet.

bstott
Beginner
Posts: 25
Joined: Tue Mar 12, 2013 8:00 pm

Re: Balancing during the assembly

Post by bstott » Thu May 30, 2013 6:14 am

Very good! So, what is your balancing method? Do you start the crazyflie - check the self test results then swap out propellers until you are close? OR do you add a little scotch tape or as was suggested earlier for tops - a spot of nail polish? Then, re-test?

Anxious to use your firmware with the new D_CFclient.

DesTinY
Member
Posts: 97
Joined: Sat Feb 09, 2013 5:18 pm
Location: Bünde, Germany

Re: Balancing during the assembly

Post by DesTinY » Thu May 30, 2013 11:08 am

At the moment, the best and fastest way is to have some spare propeller, exchange and test it until you find the best setup.
For people with more patience :D you can balance the propeller it self. I prefer to find the heavy side of the propeller with the needle and magnet method and then use sandpaper to remove some material of the heavy side of the propeller. After this you can also let the propeller on the motor use the sandpaper on the heavy side and test it again. Repeat it as long as the values ​​are better.

The Balancing self test is just a tool to measure the condition of the propellers. I hope that some one find a better way to tune the propeller for less vibrations and not only use spare propellers.

For the UI release i want to add a calibration feature. With this feature you can calibrate the self test just with the spinning motors and use the balancing test with a clean start up setup.

DesTinY
Member
Posts: 97
Joined: Sat Feb 09, 2013 5:18 pm
Location: Bünde, Germany

Re: Balancing during the assembly

Post by DesTinY » Sun Jun 02, 2013 5:07 pm

tobias wrote:Nicely done!

So you got around the bug? Sorry I havn't had time to look at it yet.
Yes, now the balancing self test has its own task. And here it is!

CF FW (balancing mod):
Image for flashing: cflie_balancing_self_test_v1.0.bin
Source: Bitbucket Link CF FW

CF Client (balancing mod):
CF Client Win32installer: cfclient-win32-install-65f5baffcd9b+.exe
Source: Bitbucket Link CF Client

CF Client Balancing Tab looks like this:
Image

UI is self-explanatory, but it is recommended to calibrate the test environment before using it. For that you have to remove (Attention at removing propellers!!!) all the propellers you want to calibrate and mark they before you push the Calibrate button. At the moment the calibration will only stored in the CF and as long as the CF is on, if you switch it off and on the defaut values will be used.

Note: If you test a longer time the battery becomes weaker and you have to re-calibrate the test environment. The Tolerance should be at +-1% after repeating the test.

Happy testing! :D

Post Reply