Motor Master 5000 - rpm, thrust, power, etc. meter and tester - Discussions

OK, so I finally got sick of testing my motors/esc/props on make-shift test stands. I decided to make something that will be as close to finished product as possible and hopefully, if there is any interest I could make it into an official product.

Without further ado, I present Motor Master 5000 (as in 5000g max thrust). Here are the specs:

1) Max voltage - 28V (0.1V resolution)
2) Max current - 50A (0.01A resolution)
3) Max load - 5000g (1g resolution, ~175oz)
4) Max RPM - 30000 (for measuring performance without the prop)
5) Size - 50 x 35 x 10 cm (LxHxW ~ 20"x14"x4")
6) Weight - 2kg (~ 4lbs)

Right now It's a standalone product and it displays data on the color LCD and also sends data via serial connection to the computer where you can further analyze it.
It sends data in ASCII or binary format (to be interpreted in the client application e.g. Processing which I will post soon too). Here is the sample ASCII output that can be pasted in Excel and then you can do with it as you please.

MotorMaster 5000 ver 1.0.1
VCC[mV] 4935 Cell K 0.186423 Offset 2083.6311 RAM 972
Servo RPM Vb[mV] I[mA] Thr[g] Cell Eff[g/W]
1000 0 11229 0 2.49 2097 0
1000 0 11229 0 2.49 2097 0
1050 382 11200 114 4.17 2106 3.43
1050 410 11200 105 3.98 2105 3.68
1100 804 11200 168 24.49 2215 13.56
1100 813 11200 168 25.42 2220 14.11
1150 1081 11200 265 45.93 2330 15.81
1150 1082 11200 259 46.12 2331 15.91
1200 1362 11168 441 70.16 2460 14.30
1200 1369 11170 413 71.10 2465 15.45
...

I would like it to (eventually) become an open source project but for now I will keep polishing it until it is ready for "public scrutiny"

Here are some pictures (sorry for the lighting) and I will soon be posting more and some videos too.

Oh and yeah, good news is it's also an Arduino! (Pro Mini which was even cheaper than the ATMega328p chip and you can easily replace it or use elsewhere in a dire need)

I'm open for suggestions and comments so please feel free to write.

Cheers,
Misa

Great - just great.

I had such project on my endless todo - you are way ahead of me.

2 additional things i had on todo:

MPU6050 , for vibration measurement.

And then: an attemp to use optical rpm sensor +timing/math, to attempt to predict balancing position.

most of the needed code exists in br-gimbal project.

Whatever you do, please think try to have multiplatform support from the start , so Qt might be a good solution, of - in worst case, something that runs in mono, or java.

Finally, for proper thrust /vibrations measurement you will want to have some distance between the arm and the prop, the best method - is to have a arm of two or three 2mm thick aluminum/CF plates, so the airflow is as little influenced as possible.

Great project - very useful.

Replies

Developer

Jani Hirvinen March 18, 2014 at 10:10pm

That looks a bit similar what we have been using at jDrones many years. Do you also have automatic acceleration programs running on your tester?

Heres some pictures of jDrones Motorbench. You should add some safetynet on your device too :)
- Misa Rakic > Jani Hirvinen March 19, 2014 at 4:58am
  
  P.S. Love jDrones, I ordered multiple times and shipping was extremely fast each time.
- Misa Rakic > Jani Hirvinen March 19, 2014 at 4:49am
  
  Yes Jani, it looks similar which is exactly my point: "everyone" (or many to be fair) has build one because everyone (actually everyone to be fair :)) who is using electric power needs it. It is just finally the time to have the actual product so that people don't have to build themselves, rather use OTS product that would potentially bring some consistency to the data as well.
  
  Not sure about automatic accel, what do you mean? I have Auto program where it goes from SERVO_LOW to SERVO_HIGH and back and logs the data. What kind of accel test would you like to see?
  
  You are right about the net, I was thinking of having it as an add-on just still working on the design.
Emin Bu March 18, 2014 at 1:39pm

do jaja!
- Misa Rakic > Emin Bu March 18, 2014 at 5:51pm
  
  Hvala Emine :)
Andre K. March 17, 2014 at 10:59am

Great - just great.

I had such project on my endless todo - you are way ahead of me.

2 additional things i had on todo:

MPU6050 , for vibration measurement.

And then: an attemp to use optical rpm sensor +timing/math, to attempt to predict balancing position.

most of the needed code exists in br-gimbal project.

Whatever you do, please think try to have multiplatform support from the start , so Qt might be a good solution, of - in worst case, something that runs in mono, or java.

Finally, for proper thrust /vibrations measurement you will want to have some distance between the arm and the prop, the best method - is to have a arm of two or three 2mm thick aluminum/CF plates, so the airflow is as little influenced as possible.

Great project - very useful.
- James Overington > Andre K. March 18, 2014 at 1:23am
  
  Very similar rig to what I built when I was designing my quad - though I must say yours looks very much more professional. The base of mine was a rough sawn 2x8 pine... The information I learnt using mine was invaluable. I used an Arduino Nano for my project. I measured rpm using a simple circuit across ones phase of the motor the output of which was a 5v square pulse at half the frequency of the ac phase voltage. Rpm is then calculated based on the number of magnetic poles. I did have problems with switching noise from the esc but did get useful data. How fast can you sample? Including some sort of time stamp in the output might be helpful for gauging response times if you can sample fast enough. I have been meaning to do some more tests to measure the thrust response if larger vs smaller props
  - Misa Rakic > James Overington March 18, 2014 at 2:31am
    
    I decided against the phase reading because you have to enter the right number of poles for each motor, although I should probably put it on the board and have it as an option. I wanted to keep it as easy/simple as possible and IR seemed like a dummy proof method.
    Sampling is around 20-50Hz, but could be faster it just gets noisier with higher sampling. I am already averaging data to get responses more stable and similar to digital scales (very stable output but at 1-0.5 Hz).
    My previous base was also wooden plank but for load cells it was flexing too much...
    - Andre K. > Misa Rakic March 18, 2014 at 2:43am
      
      See screenshot, the narrow beam IR diode could be elevated and even closer for improved Max rpm.
      tmp_Screenshots_2014-03-18-10-39-45829001296.png
      - Misa Rakic > Andre K. March 18, 2014 at 5:36am
        
        Actually (after some thought), the error would not have to be greater (I just measure time after n pulses, n = number of prop leafs) and even the prop config "hassle" would be fairly easy to solve.
        
        However, the exact reference point would be more difficult to determine (could be either one of the propeller leafs) so that would require some solution. The tape (or paint) provides exact reference point from which I then determine the imbalance. Any thoughts on that?

of 3

This reply was deleted.