johnrobholmes
10 MW
Excellent info! More than I know what to do with :lol:
Mars Electric LLC Axial Flux Motor
- Thread starter Jeremy Harris
- Start date
very thorough & very much appreciated.
i wouldn't have thought anything of the straight lams until u pointed them out now uve got me curious about them.
i think no curvature means they would have to be individual shoes kinda like bobbins?
they're not wound individually though so it is kind of odd if it really is that way not to take advantage of it.
prior to ur last post i was considering that the bump in the laminations might be a reluctance notch for some sort of indexing but i guess that's not the case.
that separate magnet ring for the halls tho once removed sure makes a convenient place to print out & glue on a checker board code wheel for an optical resolver should u wanna play with vector control.
look forward to hearing conclusions.
i wouldn't have thought anything of the straight lams until u pointed them out now uve got me curious about them.
i think no curvature means they would have to be individual shoes kinda like bobbins?
they're not wound individually though so it is kind of odd if it really is that way not to take advantage of it.
prior to ur last post i was considering that the bump in the laminations might be a reluctance notch for some sort of indexing but i guess that's not the case.
that separate magnet ring for the halls tho once removed sure makes a convenient place to print out & glue on a checker board code wheel for an optical resolver should u wanna play with vector control.
look forward to hearing conclusions.
liveforphysics
100 TW
Bigmoose = Hero.
Jeremy Harris
100 MW
Great work, Bigmoose, this is more than enough data to work with for now.
It's interesting to see that the motor is connected in wye, as one immediate way to up the Kv would be just to reconfigure it as delta. This would increase the losses very slightly (from circulating currents) but would also reduce the effective winding resistance, increasing the safe maximum current.
The motor seems to be fairly well made, the tolerances seem pretty good to me and some of the design features, like the separate Hall trigger system and the commendably low flux on the back of the rotor seem to have been well thought through.
This motor isn't going to be a challenge to drive with available controllers, either, as it has a pretty high time constant as it stands. Pretty much any BLDC controller should drive it OK, as it's similar in characteristics to a high speed geared hub motor.
Jeremy
It's interesting to see that the motor is connected in wye, as one immediate way to up the Kv would be just to reconfigure it as delta. This would increase the losses very slightly (from circulating currents) but would also reduce the effective winding resistance, increasing the safe maximum current.
The motor seems to be fairly well made, the tolerances seem pretty good to me and some of the design features, like the separate Hall trigger system and the commendably low flux on the back of the rotor seem to have been well thought through.
This motor isn't going to be a challenge to drive with available controllers, either, as it has a pretty high time constant as it stands. Pretty much any BLDC controller should drive it OK, as it's similar in characteristics to a high speed geared hub motor.
Jeremy
Tiberius
10 kW
bigmoose said:
Bigmoose,
I posted some BEMF traces from the Turnigy 63-64 and 80-100 a while ago.
http://endless-sphere.com/forums/viewtopic.php?f=30&t=17849#p260170
They are not quite sinusoidal, but they are close.
Nick
johnrobholmes
10 MW
Mr. Harris- I did find that sensorless Xie Chang controllers would not start up the motor. Just too hard to break the initial cog force of the motor I suppose.
bigmoose
1 MW
Thanks Nick! I had missed your excellent thread on Kv.
I was dithering with the inductor charge equations this morning and plugged in the averages from the measurements that I posted yesterday. As Jeremy said, the axial motor is a kitten to drive. Very benign time constant. Note how it compares with two outrunners that just about everyone is familiar with.
Interesting is how at low time intervals the medium and large Turnigy inductor charging curves overlay each other. That is caused by their similar inductance, and being in the part of the curve that inductance dominates.
What I think is interesting when reading the CISRO motor (Coreless Axial Flux 98% efficiency wheel motor) literature, is that they place a 110 uH (12 mOhm) inductor in series with each phase lead, if I recall correctly. That makes the CISRO easy to control, as in it gives the controller guys time to sense current and turn off the FETS.
As we have learned, Low inductance, Low resistance winds are killers for the controller.
PS: To John, I agree this motor really had a "detent" at zero RPM, or as you said it really cogged.
I was dithering with the inductor charge equations this morning and plugged in the averages from the measurements that I posted yesterday. As Jeremy said, the axial motor is a kitten to drive. Very benign time constant. Note how it compares with two outrunners that just about everyone is familiar with.
Interesting is how at low time intervals the medium and large Turnigy inductor charging curves overlay each other. That is caused by their similar inductance, and being in the part of the curve that inductance dominates.
What I think is interesting when reading the CISRO motor (Coreless Axial Flux 98% efficiency wheel motor) literature, is that they place a 110 uH (12 mOhm) inductor in series with each phase lead, if I recall correctly. That makes the CISRO easy to control, as in it gives the controller guys time to sense current and turn off the FETS.
As we have learned, Low inductance, Low resistance winds are killers for the controller.
PS: To John, I agree this motor really had a "detent" at zero RPM, or as you said it really cogged.
Jeremy Harris
100 MW
johnrobholmes said:
I suspect that's down to the sensorless XieChang not having a start up algorithm in the way that RC controllers do. The sensorless XieChang is just a sensored one with the addition of a set of zero crossing detectors and filters hooked up to the phase wires, there's no change to the internal firmware to provide the "pulse some phases and see what happens" start up routine that the RC controllers use. I bet that a standard XieChang, with the Halls hooked up, would run it just fine, as it has electrical parameters that are pretty close to a geared hub motor.
Jeremy
johnrobholmes
10 MW
Yep, indeed the sensored version worked like a charm. RC controllers worked fine as well.
Jeremy Harris
100 MW
bigmoose said:
Really interesting stuff, Bigmoose, thanks a lot for taking the time to produce these plots.
If anything, this motor looks as if it may need a lower PWM frequency for high power use as it stands. Most ebike controllers work at around 15kHz, so the longest PWM on time is only going to be around 66uS, unless the controller can go into block commutation mode. The phase current seems likely to still be limited by the motor LR at this frequency, I'm guessing that it'd be hard to get a phase current of more than around 100A or so at 99% duty cycle. If the controller can switch to block commutation at maximum throttle, then the phase current could be a lot higher, but some ebike controllers don't seem to do this, they retain some PWM even at full throttle with no current limiting.
The curves for the outrunners are interesting too, especially as I'm looking to rewind a similar stator with the aim . Would you mind posting the raw data files, please?
Jeremy
bigmoose
1 MW
No Problem Jeremy. Be sure to keep me honest, and make sure I didn't make a mistake somewhere.
Jeremy Harris
100 MW
Very many thanks, Bigmoose.
Jeremy
Jeremy
jmygann
100 kW
- Joined
- May 30, 2008
- Messages
- 1,069
So ... is this a keeper in the 1-3 Kw range ?
Is any one in contact with john at Mars ?
the last e-mail I got from him ....
"Please check back in a few weeks. I should have some on the way by then, or at least know more about the schedule."
" I am waiting on some feedback before I order some for inventory. The price will be $150 for small sample quantities.
Is any one in contact with john at Mars ?
the last e-mail I got from him ....
"Please check back in a few weeks. I should have some on the way by then, or at least know more about the schedule."
" I am waiting on some feedback before I order some for inventory. The price will be $150 for small sample quantities.
Josh K.
100 W
Yes, I have been in touch with John, From Mars and he has some motors being made up right now.
Seems that a bunch are being made up for the special effects crowd in Hollywood, for on demand
wind generation with photo shoots!
We will be able to order motors soon, and I will be offering several packages including one
with a reduction and double clutch to be put through the gears on a recumbent. In addition,
I will offer one with a simple single stage gear reduction, that is a big gear on the non drive side!
We have been waiting to get the best results with an advanced controller, but may be settling
with the understanding that this is the sound of condensed power, No apologies!
I thank all that have contributed to this post, and look forward to seeing some nice machines powered
with this equipment. I will post some photos of our smaller version, with wires coming up and out of
the side, instead of out of the back...
Please let me know if you would like one of these motors, and we will get you on the list!
Here is an older video of a project that I was working on the year before last...
http://www.youtube.com/user/Joshk750w?feature=mhum#p/a/u/0/ZPUeh6QUHo0
Peace, Josh Kerson
http://www.electric-cycle.com
Seems that a bunch are being made up for the special effects crowd in Hollywood, for on demand
wind generation with photo shoots!
We will be able to order motors soon, and I will be offering several packages including one
with a reduction and double clutch to be put through the gears on a recumbent. In addition,
I will offer one with a simple single stage gear reduction, that is a big gear on the non drive side!
We have been waiting to get the best results with an advanced controller, but may be settling
with the understanding that this is the sound of condensed power, No apologies!
I thank all that have contributed to this post, and look forward to seeing some nice machines powered
with this equipment. I will post some photos of our smaller version, with wires coming up and out of
the side, instead of out of the back...
Please let me know if you would like one of these motors, and we will get you on the list!
Here is an older video of a project that I was working on the year before last...
http://www.youtube.com/user/Joshk750w?feature=mhum#p/a/u/0/ZPUeh6QUHo0
Peace, Josh Kerson
http://www.electric-cycle.com
johnrobholmes
10 MW
You have an even smaller version?
liveforphysics
100 TW
Wow! That's an extremely controller friendly motor to drive. You could set your carrier freq way down and have no trouble at all.
Josh K.
100 W
Well, When we were looking at this motor, a couple of years ago...
We had another version made that was narrowed up a little.
Mr. Fiorenza from Mars, went to China, and got the ends made up to match up to the Cyclone
gear box for us, and re routed the wires out the back.
This helped us get the over all motor thickness down to 3 inches. And, with the wires going up,
instead of out the back, we got almost another half an inch, where our feet can fly by without
hitting the wires...
Additionally, the cover was made up to keep the rocks out, that is without holes for cooling!
Josh k.
We had another version made that was narrowed up a little.
Mr. Fiorenza from Mars, went to China, and got the ends made up to match up to the Cyclone
gear box for us, and re routed the wires out the back.
This helped us get the over all motor thickness down to 3 inches. And, with the wires going up,
instead of out the back, we got almost another half an inch, where our feet can fly by without
hitting the wires...
Additionally, the cover was made up to keep the rocks out, that is without holes for cooling!
Josh k.
Samba
100 W
Josh K. said:
If you are getting cyclone gearboxes made, could you have them use tougher planetary gears or order spare planetaries?
Sounds interesting but some power, rpm, decibal info would be great!
MitchJi
10 MW
Hi Josh,
Thanks!
Its fine if you decide to settle for selling a noisy version of the motor rather than fix the problem but please don't say its a feature.Josh K. said:
Thanks!
bigmoose
1 MW
I have been pondering the noise signature since I had this motor apart last week. I am thinking that axials put a lot of stress on the rear cover plate that the stator attaches to. Think of it as a drum with the rotor exciting the drumhead right at the center with an axial force that varies with phase excitation. I think the noise may be coming from the rear cover and not the rotor. I do not have the instruments to determine for sure. This may be difficult to engineer a light and effective solution for. The cover plate is going to have to be extremely stiff.
johnrobholmes
10 MW
I have contacted John at Mars and we are seeing about skewing the stator and using 20 round magnets to help reduce the detent force. I think this will greatly reduce the reverberations of the motor, even if it doesn't totally solve it. I would love to build a bike around the motor, but unless they run quiet on 6 step commutation it just won't find a long term place in my stable.
We can get the noise issue solved one way or another. The standard design will still be available in the mean time for people that don't mind the noise. If you are wanting to build something with this motor, I would say go ahead and get one.
We can get the noise issue solved one way or another. The standard design will still be available in the mean time for people that don't mind the noise. If you are wanting to build something with this motor, I would say go ahead and get one.
liveforphysics
100 TW
Please dont skew the rotor teeth....
It will only help with the torque ripple at the cost of efficiency, and mimimaly effect the noise.
Remember, the cogging isnt what causes the noise, its the alternating push-pull of the torque producing magnetic field flexing the rotor back and forth a whisker, much like a speaker.
A little acustic dampening material on the back of the rotor, or altering the rotor design would be better options IMHO.
It will only help with the torque ripple at the cost of efficiency, and mimimaly effect the noise.
Remember, the cogging isnt what causes the noise, its the alternating push-pull of the torque producing magnetic field flexing the rotor back and forth a whisker, much like a speaker.
A little acustic dampening material on the back of the rotor, or altering the rotor design would be better options IMHO.
Similar threads
