Modifying the Turnigy 80-100 Brushless Outrunner – Part 1

Modified Turnigy 80-100 on testmount
Modified Turnigy 80-100 on testmount

The Turnigy 80-100 is a electric motor sold by Hobby King to replace gas engines in large RC aircrafts. To use this motor in my application two things have to be modified. Installing hall-sensors and re-winding the motor for lower speed. This motor is a BrushLess Direct Current (BLDC) outrunner. This means a couple of things:

  • The motor does not have brushes like an ordinary DC motor. Instead the commutation (switching the current direction the motor windings) is done electronically with high power semiconductors (most commonly MOSFETs)
  • It runs on DC power which is a little confusing since the motor actually is a 3-phase AC motor. But the motor+controller runs on DC.
  • Outrunner means that the motor shell is rotating while the center is static. The center of the motor contains the windings and stator core while the outer bell with permanent magnets are rotating.

Since the commutation is done electronically the motor controller must know when to switch direction of the current. This is done in one of two ways:

Sensored commutation:
The motor is fitted with 3 hall-sensors which sense changes in the magnetic field of the rotor

Sensorless commutation:
Since only 2 of 3 phases of the motor are energized the motor back-emf can be measured on the third terminal to determine rotor position.

This motor is intended to be used with a sensorless controller and has no hall-sensors. A sensorless controller need to spin the motor up to ~10% of maximum rpm in synchronous mode before the back-emf is large enough to measure. Below this speed the torque isn’t very high which works fine for a propeller drive but not on a moped where full torque is required from 0 rpm.

In general BLDC motors and controllers intended for RC toys are sensorless and use sensorless controllers while motors and controllers for E-bikes are sensored. There are exceptions from this but it’s good to know since a sensored controller will not work with a sensorless motor, the other way around could work but the low-speed performance will probably be bad and there is a risk of damaging the motor and/or controller.

To get good starting torque and be able to use an ordinary e-bike controller I mounted hall sensors on my motor. This process is well described in a thread on the Endless Sphere forum:

Adding hall sensors to outrunners

To summarize the +20 pages thread there are two ways of doing this

  • Internal sensors are mounted between the stator teeth at 120° spacing
  • External sensors are mounted on a bracket outside of the bell, this uses the magnetic flux leakage to sense the magnets on the other side of the bell.

Another problem with using the motor in it’s original configuration is the Kv value. This is the constant that determines the motor maximum speed based on the input voltage. For example a Kv value of 1000 rpm/V will result in a maximum speed of 12000 rpm with a 12 V battery. This value depend on several properties on the motor but you can say that it represents the coupling between the current and the magnets. More turns of wire around the stator and/or stronger magnets will reduce the Kv value. The Kv value is also dependent on if the motor is terminated in wye or delta. The same motor have a Kv that is sqrt(3) = ~1,73 times higher if it’s connected in delta than if it is connected in wye.

When I bought this motor it had a Kv of 180 rpm/V and I want it to be ~90 rpm/V. Each stator tooth had 6 turns of copper wire around and the motor where coupled in delta mode. By rewinding the motor with 7 turns on each stator pole and couple the motor in wye instead the resulting Kv is somewhere around

[latex]
180 * \frac{6}{7} * \frac{1}{\sqrt{3}} = 89 rpm/V
[/latex]

This is not an exact calculation since it depends on flux density, magnetic saturation of the stator iron and so forth but it will giva a hint. As i calculated in a previous post, this is enough for ~60 km/h using the same 44.4 V battery as I use on my E-MTB.

There is a thread on Endless Sphere about rewinding this motor as well
Re-wind of a Turnigy 80/100
Rewinding a motor is a tough job but the original winding is done with many parallel thin wires and in a pretty sloppy way. Instead, I used two parallel strands of 1.5 mm copper and it ended up almost as sloppy as before. I seem to have misplaced the photo of the stator with windings before the re-wind but it looked very similar to the pictures in the first post of the thread above. This is how it looked when i were done.

[Will replace with photo next time i disassemble the motor]

When mounting the sensors I choose the method of mounting them externaly. I used a CAD program to draw this mounting bracket.

Drawing of sensor bracket
Drawing of sensor bracket

Mounting the sensors 17,14° apart instead of 120° works because the motor have 14 magnet poles.

[latex]
\frac{120^\circ}{\frac{14}{2}} = 17.14^\circ
[/latex]

A nice guy on a The Swedish electronics forum helped me print two brackets on his 3D-printer and they turned out great!

Plastic sensor brackets printed on 3D Printer
Plastic sensor brackets printed on 3D Printer

With wires mounted on the sensors and the sensors temporarily glued in with heat glue (I’ll use epoxy when i know that it works).

Sensors mounted in bracket
Sensors mounted in bracket

In the pictures above the motor is mounted on a plate that i made to test this way of mounting the sensors. Just to get it running I used a Hobbyking SS Series 190-200A ESC after the rewind this controller had a tough time getting this motor running. Using a 3S LiPo battery it managed to get the motor into closed loop back-emf sensing mod about one time out of ten. With a 6S LiPo it worked perfect and had loads of power! The no-load current consumption was slightly over 1 A, which is great but mostly dependent on that I didn’t re-install the skirt bearing. This motor have been reported to have a no-load current of ~9 A with the skirt bearing and coupled in delta. I also measured the Kv constant to ~89 rpm/V exactly as calculated.

My next post on this project will be about my modified eBay cheapo e-bike controller and hopefully a video of the motor running in sensored mode.

Continue to: Modifying the Turnigy 80-100 Brushless Outrunner – Part 2

24 thoughts on “Modifying the Turnigy 80-100 Brushless Outrunner – Part 1”

  1. Grazie delle infomazioni che cercavo da tempo, per un bldc outrunner sensored
    provero e vi faro sapere.
    saluti A-R

      1. I just told you that I find your project very interesting. Wondering if you can help me, I would like to know if with the ESC that you are using is connected with 3 Sensored signals?

        1. Yes, but I’ve only got the sensored mode working once. This controller detects if there is something wrong with the sensor signals and goes into sensorless mode. I’m not sure why but I guess it has something to do with the cheap hall-sensors I’m using.

          I will try to install Honeywell SS441A hall sensors (which others have used successfully) mounted internally instead.

          1. I have successfully used the ss411a externally mounted just like you have done with a smaller RC motor (I just carefully drilled holes in a plastic block– no 3d printer here, hehe). Some tuning was required– the positioning of the plastic block is important. I held the motor and block in a jig and found the position at which the motor started with the least throttle applied, then marked it with a sharpie.

            I used a normal E-Bike controller, however it does not like the motor at high throttle and cuts out. Still, the more critical low speed performance is night and day better than sensorless. The scooters I have mounted the motor to have not been able to withstand the stresses of this combo w/ lithium batteries and both broke. I have not yet debugged why the controller cuts out and probably won’t until I can bolt the motor to something more sturdy.

          2. Interesting, I’m in the process of developing my own e-bike controller. A small sensored RC motor would be perfect to use for desk-testing of this controller. What kind of motor do you use, any pictures?

            A reason for the controller cutout could be that the motor spins to fast. For my e-bike with a motor speed coefficient of ~7.1 rpm/V at 50 V will rotate at a maximum 50 * 7.1/60 = 5.9 rev/s. With 46 magnets and 6 commutations steps per phase this will result in a commutations frequency of 5.9 * 46/2 * 6 = 817 Hz at max speed. A small RC motor with 14 magnets and 300 rpm/V at the same voltage would result in a commutations frequency of 50 * 300/60 * 14/2 * 6 = 10,5 kHz. This may be more than the controller can handle.

            The rewinded Turnigy 80-100 spins at around 90 rpm/V and have 14 magnets which gives a commutation frequncy of 50 * 90/60 * 14/2 * 6 = 3150 Hz

          3. The motor is a Hyperion Z5045-18 outrunner, rated to 2150W continuous, 3kW peak. A lot of power on a tiny scooter!
            http://www.empirerc.com/z50all.htm

            KV is 168. At my pack’s 42V, that’s over 7k rpm. I forget the number of magnets inside, but it’s probably close to the Turnigy.

            On my first ebike years ago, I recall I had to replace the oscillator to “overclock” the controller to run at 36v. It would also drop out at high speed, similar to the RC motor. Another possibility is that the current limit is also kicking in.

          4. I wouldn’t suspect the current limit, if the motor is unloaded the current would be much higher at low rpm.

            It could also be that the controller tries to increase the speed over 7 krpm by using phase advance or field weakening. If you use a controller based on the infineon chip the speed can be programmed to 120% where the last 20% is achieved with phase advance / field weakening. The controller can decrease the magnetic field strength by inducing an opposing field with the stator windings. This will increase kv but because of kt [Nm/A] = 1 / kv [rad/s] the torque per amp will decrease proportional to the increase in speed. This will lead to an increase of current draw when the motor speed exceeds 7 krpm, assuming that the friction torque is the same.

  2. Ciao mr
    interessante questo suo progetto se mi risponde le farò delle proposte grazie.
    Angelo

  3. Hi
    Had a question to your calculation of the position of the hallsensors, because you said you have 14 magnets so im guessing you have 12 poles. so that would be 120/(14/2), in may case however i have 28 magnets and 24 poles, which would make my angle 120/(28/2) = 8.57 degrees, do you agree? im just not quite sure what the 2 stands for, is that for polepairs?
    would apprechate an answer because im in contact with somebody with a 3d printer and he would like to know if he can do it with 17.14 or 8.57 degress.

    regards

    -ivo

        1. No problem.

          To explain a little further:
          For each 360° electrical degrees the rotor will move “two magnets”, one north and one south. If you have 28 magnets 360° mechanical degrees will then be the same as 360° * 28 / 2 = 5040° electrical. Each electrical degree on your motor is then represented by 360° / 5040° = 0.071428° mechanical. Since you want the three sensors evenly placed on one electrical rotations you would like to place them 360° / 3 = 120° electrical degrees apart. Which would then be 120° * 0.071428° = 8.57° mechanical degrees. (or in short 120° / (28 / 2) = 8.57°)

    1. Looks like a switched reluctance motor but it’s hard to tell from your pictures. Are you sure the center part (rotor) is magnetic? If it is you could run it with a standard sensorless controller if not you need something special for switched reluctance motors.

      The power part is easy start low, if you want more speed increase voltage, if you want more torque increase current and if the motor gets hot you have increased to much.

  4. what tipe of engine it this and what power can have
    [IMG]http://i45.tinypic.com/2604eaq.jpg[/IMG]

  5. Sorry for ressurecting this thread after such a long time (if someone ever reads this 🙂 ), but I’ve been googling for a small sensored bldc (ideally outrunner) for a long time and this post comes the closest to what I need.

    Is there anywhere to buy a sensored outrunner at about 25-30A? Or do you think the sensors would fit around a case with ~50mm in diameter? Are there some more resources for selecting the hall sensors and wiring them to an esc? Thanks.

    1. I think your only option is to mount the sensors outside the rotor. If you find one with a low pole count maybe it’s possible.

  6. Great project, now i understand how to get more power from my outrunner motor used as a spindlemotor. this motor stalls to quickly at low rpm (kv250) adding closed loop will greatly enhance the usabillity: only question i have is, what type of hall sensor did you use?

Leave a Reply

Your email address will not be published. Required fields are marked *