A bit of theory help please
- Thread starter Tek
- Start date
Ebikebert
100 W
To my knowledge there is one deciding factor and that is the windings of a motor and how many turns it has.
Thud
1 MW
#1 yesTek said:
#2 yes
below is a document cataloging rc controllers & their max potential speed they can drive a motor. this does not include any e-bike controllers but will let you see the function of erpm as it relates to pole count.
then there is the actual wire resitance & the ability to charge & uncharge coils in a real time senario.
in actuality, you will be against the motors material limitations befor seeing many of the max numbers in the list:
Lebowski
10 MW
Plus in addition its limited by battery voltage. Also then reducing pole count helps but at efficiency cost, better go reduce the amount of turns and go to fatter wiring
Tek said:
To add to thud's post, if the controller isn't a limiting factor, BEMF, or CEMF will be.
counter-electromotive force
spinningmagnets
100 TW
There is no perfect motor for every application. And, many of the upgraded features that arguably "improve" a motor will make it more expensive.
The BMC geared hub has thinner laminations and a higher quality of steel than the MACs that were made before this last month. At 48V there was no lamination performance difference between them, but...enough people wanted to run a geared hub at 72V (which results in higher RPM's), that the new MACs are reported to have upgraded lamination (at presumably only a small additional cost).
Pole-count: The MAC has twice as many poles as the BPM (both are similar geared hubs). Again, at 48V, probably no major effective difference. In a 26-inch wheel, the higher pole-count motor would theoretically perform smoother from a dead-stop using a sensorless controller, but with hall-sensored operation they should both perform the same.
And then: pole count at higher RPM's. The MAC/BPM both have no issues at their advertised 36V, but since the motor spins 5 times faster than the wheel in a geared hub (26-MPH in a 26-inch wheel is 336 wheel-RPMs, or 1,680 motor-RPM's). This is the reason for MAC upgrading the laminations. Eddy-current losses are long subject (and I still have more to learn), but...the more eddy currents you have, the more of your batteries watts are converted into waste-heat instead of work.
The BPM can run to around 2,000 motor-RPMs without eddy-current issues (IIRC), but the old MAC could only run to about 1,000-RPM's. It hasn't been very noticeable to MAC owners because they would have to spend a lot of time at the top RPMs on a long uphill that forced to controller to provide high amps for an extended time to see a problem. A high percentage of MAC kits were also being ordered to be used with 48V instead of 36V, but not necessarily using a slower winding, and the resulting higher motor-RPMs were stepping over the edge more often into eddy-current territory than before.
The most common controllers are very price sensitive, and they pinch pennies on every single component to remain competitive. The ability to run at a very high electrical frequency (a high pole-count motor compared to a low pole-count motor at the same motor-RPM's) might only cost a few cents more, but...you have to specify a certain high-RPM capability when shopping for a controller if you need that (a high pole-count motor at high volts/high motor-RPMs).
Also, the cost and complexity of external reductions to get roughly 300-400 wheel-RPMs. The is a power-to-size benefit to using a smaller motor that is run at a higher RPM, and then reduced to the desired output RPMs. Every stage of reduction that is added will also add complexity and cost, not to mention that higher RPMs at the motor have a different sound than a larger lower-RPM motor with fewer reductions. Some builders like the sound, some don't. Manufacturers make what sells, and often builders compromise on their "theoretical ideal" motor because they are limited to using existing motors.
I am going through the hassle of converting geared hubs to a shaft-drive because I can't find an existing shaft-drive motor in that size and price, with the kV selections I can have as a result of my conversions.
I must confess I still don't understand all of the factors that influence or limit the upper RPM's of our brushless DC motors...
I seem to remember that for ultra-high RPM operations, the optical timing sensors were better than magnetic-sensing halls.
The BMC geared hub has thinner laminations and a higher quality of steel than the MACs that were made before this last month. At 48V there was no lamination performance difference between them, but...enough people wanted to run a geared hub at 72V (which results in higher RPM's), that the new MACs are reported to have upgraded lamination (at presumably only a small additional cost).
Pole-count: The MAC has twice as many poles as the BPM (both are similar geared hubs). Again, at 48V, probably no major effective difference. In a 26-inch wheel, the higher pole-count motor would theoretically perform smoother from a dead-stop using a sensorless controller, but with hall-sensored operation they should both perform the same.
And then: pole count at higher RPM's. The MAC/BPM both have no issues at their advertised 36V, but since the motor spins 5 times faster than the wheel in a geared hub (26-MPH in a 26-inch wheel is 336 wheel-RPMs, or 1,680 motor-RPM's). This is the reason for MAC upgrading the laminations. Eddy-current losses are long subject (and I still have more to learn), but...the more eddy currents you have, the more of your batteries watts are converted into waste-heat instead of work.
The BPM can run to around 2,000 motor-RPMs without eddy-current issues (IIRC), but the old MAC could only run to about 1,000-RPM's. It hasn't been very noticeable to MAC owners because they would have to spend a lot of time at the top RPMs on a long uphill that forced to controller to provide high amps for an extended time to see a problem. A high percentage of MAC kits were also being ordered to be used with 48V instead of 36V, but not necessarily using a slower winding, and the resulting higher motor-RPMs were stepping over the edge more often into eddy-current territory than before.
The most common controllers are very price sensitive, and they pinch pennies on every single component to remain competitive. The ability to run at a very high electrical frequency (a high pole-count motor compared to a low pole-count motor at the same motor-RPM's) might only cost a few cents more, but...you have to specify a certain high-RPM capability when shopping for a controller if you need that (a high pole-count motor at high volts/high motor-RPMs).
Also, the cost and complexity of external reductions to get roughly 300-400 wheel-RPMs. The is a power-to-size benefit to using a smaller motor that is run at a higher RPM, and then reduced to the desired output RPMs. Every stage of reduction that is added will also add complexity and cost, not to mention that higher RPMs at the motor have a different sound than a larger lower-RPM motor with fewer reductions. Some builders like the sound, some don't. Manufacturers make what sells, and often builders compromise on their "theoretical ideal" motor because they are limited to using existing motors.
I am going through the hassle of converting geared hubs to a shaft-drive because I can't find an existing shaft-drive motor in that size and price, with the kV selections I can have as a result of my conversions.
I must confess I still don't understand all of the factors that influence or limit the upper RPM's of our brushless DC motors...
I seem to remember that for ultra-high RPM operations, the optical timing sensors were better than magnetic-sensing halls.
Ok. I read your post and looked at the list of hobby controller RPM limits. Can anyone explain why all the motorcycle sized BLDC that I can find are usually rated at a max of around 4000-6000rpm? I mean 'inrunners' making over 5kW and suitable for a fullsized scooter or motorcycle. I understand that they aren't going to spin over 20,000rpm and I wouldn't really want or expect more than 12,000rpm safely.
Somebody explain...
(I have little interest in hub motors, sorry)
Somebody explain...
(I have little interest in hub motors, sorry)
Thud
1 MW
I'll take a stab:
In industry (where the bulk of the motors in these sizes are utilized) there is little need for the higher rpm rattings as a standard...most of the larger BLDC motors are constrained into the 8k RPM range anyway.
the only controller supplier that is raising to the challange is Kelly (offering upgrades to the ERPM limits of the controllers) sevcon & others have no reason to fill this miniscule niche in the market....sevcon sales to the crane industry alone dwarfs all the e-v conversions on the planel....add in lift trucks & hydrolic systems....& we are invisable.
In industry (where the bulk of the motors in these sizes are utilized) there is little need for the higher rpm rattings as a standard...most of the larger BLDC motors are constrained into the 8k RPM range anyway.
the only controller supplier that is raising to the challange is Kelly (offering upgrades to the ERPM limits of the controllers) sevcon & others have no reason to fill this miniscule niche in the market....sevcon sales to the crane industry alone dwarfs all the e-v conversions on the planel....add in lift trucks & hydrolic systems....& we are invisable.
Of course we can't run 50,000rpm on a 15cm rotor, but maybe 15-20 yeah? If piston ICE engines are running 15,000rpm and throwing pistons back and forth at the speed of sound, it would stand to reason.
Has anyone made a controller that uses the FET driver from a hobby unit and made their own 'power stage'?
Has anyone made a controller that uses the FET driver from a hobby unit and made their own 'power stage'?
You should check out shane colton's blog. Lately he's been blogging about kerbals in space - but look through his history and he has lots of info about RC motors.
On the controller side, I think there are mainly three limiting factors to high RPM. Control algorithm, algorithm speed, and PWM frequency,. If we assume that the control algorithm is both good and fast, the main limiting factor is the PWM frequency. To be able to create smooth and well timed waves into the motor, the ratio between the PWM frequency and the motors electrical frequency, has to be high enough. That ratio depends on the control algorithm. A true sine wave controller should be able to operate with a lower ratio between PWM and motor frequency, than a "6-step".
Most controllers of today operates below 20kHz PWM. Depending on how smooth you want the waves to be, that will create different limits on the electrical frequency. Lets say that the minimum to create a 360° wave is 20 PWM cycles. That would limit the motor frequency to below 1000Hz. A motor with 7 pole pairs, like the 80100, is then limited to 142 rps / 8500 RPM. A 12 pole pair motor, like the 12080 and new Colossus 12kW, is limited to 5000 rpm.
Another thing at play is the inductance of the motor. With too low inductance, the PWM is not smoothed out, so the waves will get rough.
On the motor side, the pole count will limit RPM because of eddy currents in it's iron core.
Most controllers of today operates below 20kHz PWM. Depending on how smooth you want the waves to be, that will create different limits on the electrical frequency. Lets say that the minimum to create a 360° wave is 20 PWM cycles. That would limit the motor frequency to below 1000Hz. A motor with 7 pole pairs, like the 80100, is then limited to 142 rps / 8500 RPM. A 12 pole pair motor, like the 12080 and new Colossus 12kW, is limited to 5000 rpm.
Another thing at play is the inductance of the motor. With too low inductance, the PWM is not smoothed out, so the waves will get rough.
On the motor side, the pole count will limit RPM because of eddy currents in it's iron core.
