No announcement yet.

Analysis of Statisical data

This is a sticky topic.
  • Filter
  • Time
  • Show
Clear All
new posts

  • Analysis of Statisical data

    BY Robert Livingston
    Global User
    Posts: 1086
    (9/16/04 9:01 am)
    Analysis of published data It might be useful to take same of the data presented so far, and spell out the variation Fate has discussed. When the motors are rank-ordered according to claimed RPM, each motorís RPM band of error extends well within the range of the next faster motorís band of error. Chances of overlap are significant. There is a significant chance that an NC-1 will develop more RPM than an NC-2, and on up the line:

    Since Fate tested a limited sample (say, one dozen motors of three types?), and it is unknown whether it was a random sample (with motors representing the full range of variation), it is likely that his data applies to only a proportion of the motors in the entire production run. Since he has already measured +/-10% variation in his sample, and the whole population necessarily varies at least that much, it is suspected that wider variation would be found in the whole population. Standard statistics suggest that it is safe to assume that Fateís measurements cover the majority of motors. Anywhere from two-thirds to nine-tenths of the motors would fall within the +/-10% variability range. The remaining 10% to 33% would fall outside the distribution he has established. Thus, 10% to 33% of the Ninco motors may vary even more widely than the data above indicates.

    Torque values may vary by 10% as well, although no experimental findings have been shown by anyone. When the motors are rank-ordered according to torque, there is notable overlap, except for the NC-1, which is significantly different:
    Ninco motors claim:------------------10% RPM:--------+/-10% RPM band:
    NC-1 15,700 RPM, 74 g cm-----------1570---------------14,130-17,270
    NC-2 18,100 RPM, 263 g cm----------1810---------------16,290-19,910
    NC 7 19,300 RPM, 265 g cm-----------1930--------------17,370-21,230
    NC-5 20,000 RPM, 290 g cm-----------2000--------------18,000-22,000
    NC-3 21,100 RPM, 326 g cm-----------2110--------------18,990-23,210
    NC-4 31,000 RPM, 226 g cm-----------3100---------------27,900-34,210

    Ninco motors claim:----------------10% torque:-------+/-10% torque band:
    NC-1 15,700,RPM 74 g cm-------------7.4-----------------66.6-81.4
    NC-4 31,000 RPM, 226 g cm----------22.6----------------203.4-248.6
    NC-2 18,100 RPM, 263 g cm----------26.3----------------236.7-289.3
    NC 7 19,300 RPM, 265 g cm----------26.5----------------238.5-291.5
    NC-5 20,000 RPM, 290 g cm----------29.0----------------261-319
    NC-3 21,100 RPM, 326 g cm----------32.6----------------293.4-358.6

    Significantly overlapping performance, as well as other factors such as poor brakes, and too-steep power curves, may have contributed to Nincoís decision to drop certain motors from production. Consider only the motors stated to be currently in production:

    -------------------------------------RPM, 10% Band:--------------Torque,10%Band:
    NC-1 15,700 RPM, 74gcm--------------14,130-17,270------------------67-81
    NC-7 19,300 RPM,265 g cm------------17,370-21,230-----------------239-292
    NC-5 20,000 RPM,290 g cm------------18,000-22,000-----------------261-319
    NC-6 23,500 RPM,350 g cm------------21150-25,850------------------315-385

    The NC-1 stands alone, while the NC-5 and NC-7 overlap. These latter two do not vary significantly. I assume the NC-6 exceeds the 5 and 7 by a significant amount. The 5 and 7 seem to be about the same as the old NC-2.
    The Ninco NC-6 has a claimed output of 23,500 rpm at 14.8V and 350 g/cm of torque.
    NC-6 exceeds the 5 and 7 by a significant amount of torque, but not RPM.

    "Significant" is a statistical concept, as I use it. A significant difference means that the difference is not attributable to chance variation. When the plus or minus 10% variability bands overlap, the differance is not statistically significant. This is not my opinion, it is a conventional statistical analysis.

    But it could be that the various NC series motors do not vary as much as claimed. We have no good evidence of that, however. The only evidence we have is that they vary quite a bit, so two motors separated by a ~2,000 rpm difference may be the same, or may be different. Buy two and see.

    The "anecdotal evidence" (the stories told on this thread by guys who have compared pairs of motors) is that the motors CAN be distinguished by the Ninco claims. The more scientific method used by Prof Fate in actually measuring a bunch of motors suggests otherwise.
    Last edited by Scaleracing; 10-12-2005, 12:44 PM.

  • #2
    NINCO motor tach test findings

    NINCO motor RPM test findings.

    I measured no-load RPM of several NINCO motors to see how close they are to the NINCO claims, and to provide baseline data for technical inspection of the RAA '06 cars.
    All RPM is reported at 14.8 volts (motors marked with an asterisk* courtesy of Ninco Bob). Test instrument calibrated within 1 RPM accuracy.

    NC6 23,500 RPM, actually measured 23,476 (new motor*)
    NC5 20,000 RPM, actually measured 20,930 (new motor*)
    NC5 20,000 RPM, actually measured 21,638 (broken-in motor)
    NC2 18,100 RPM, actually measured 20,868 (broken-in motor*)
    NC2 18,100 RPM, actually measured 20,594 (broken-in motor)
    NC1 15,700 RPM, actually measured 16,708 (new motor*)

    It is interesting that all the tested motors meet or exceed the claims made by NINCO. The NC2 has nearly the same unloaded, maximum speed as the NC5. The NC2 is 2 grams heavier, due to the lack of holes in the case.
    All of the motors ran smoothly under test. I broke in one of the NC5's at 6v for one hour, resulting in a gain of some 1,500 RPM over the as-new RPM. During the break-in period, this motor barely grew warm to the touch.

    The low-RPM, high torque characteristics of these motors should provide endurance racers with low-vibration power, little heat, and long motor life.
    Next, I will be measuring the torque output.

    I also observed that the brushes are set at an angle across the end of the motor case, advanced about 20 degrees. Under no-load conditions, this results in only about 200 RPM difference in speed. The motors are timed to be fastest with the crown gear to the right of the pinion (looking forward). When I do the torque readings I will report them forward and backward.

    In light of earlier statements about a 10% variabilty band, it is interesting that the NC2 exceeds its rated 18,100 RPM by 2,768 RPM, or 15%.


    • #3
      RL do you find a new motor's rpm can fluctuate greatly?
      Thank You
      Visit my web site for your track timing options.


      • #4
        Yes, a new motor can vary within a 1,000 RPM band. They tend to run faster the longer you run them, as they are breaking in. Sometimes the RPM varies as little particles of brush material get wedged in the commutator slots and are thrown free. If you run in a new motor in the dark, it sometimes throws sparks from the commutator, or the commutor is encircled in a ring of fire. Most entertaining. After a while they settle down and run more steadily. But some motors are steady from the beginning.


        • #5
          Hi Robert

          You wrote

          "NC-1 15,700 RPM, 74 g cm-----------1570---------------16290-17,270"

          Not sure I understand, is this correct the +/- 10 % seems to be all plus on this motor.

          Let me know and we can edit, or of course more likely i just don't understand.


          Alan Smith
          SCI Owner.



          • #6
            That is an error. If the NC1 does 15,700, and 10% of that is 1,570, the range of variation would be 14,130-17,270.

            Thanks for pointing that out. I hate proof reading numbers.

            I fixed it Robert,

            Thanks for all your efforts here.


            Last edited by Scaleracing; 10-13-2005, 09:48 AM.


            • #7
              This is the device that I'm using to measure rpm, gauss and winding resistance.
              Some new motors do not give a good reading but usually after some break in things settle out. I borrowed the device from a gentleman that I just used to run blob cars and glue. He has a wood track about 5 minutes from me. He and his buddies stopped racing for more than a few years but once they ran across my web site and took a few of my cars for a ride they are back in the game but I must say they are going glueless! I look forward to their company and expertise!

              A new Little Ripper motor on the test block.

              At 15vdc the motor is humming right along. I have to multiply the out from the S-Tron by 60 to get the RPM.
              Last edited by MrBugs; 10-15-2005, 03:37 PM.
              Thank You
              Visit my web site for your track timing options.


              • #8
                I am finding that a group of around 30 stock, unaltered motors, as tested for the Race Across America Proxy series (RAA), vary up to 17% more RPM over manufacturer's rated speed at 12 volts; and down to 13% below rated speed. Most are close to the manuf. rating.

                The motors that exceed the spec by the greatest amount don't vary much from one motor to another, and are of the same make and model; it seems the error is in the published rating.

                More details will be posted after all the cars have been received for tech inspection. I am still in the middle of Tech Inspection, which is why all these motors are being tested.
                Last edited by Robert Livingston; 12-10-2005, 08:44 AM.