Anyone else working on 5 phase unipolar BL ESC?

[INewton]

I've standardized on the voltage for 10S lipo or 35 1.2v cells, which will also be the standard for the auto industry, 42 volts. It's safe, yet removes much of the stress on our wiring and batteries (and associated extra weight) for higher power motors in .40, .60 and .90 conversions. And... ten 2200 lipos (10C) is much lower cost than the 4S6000 batteries.

The problem is lack of controllers in this segment of the market and high cost of those that are available.

It turns out 5-phase unipolar BLPM motors have about the same performance as 3-phase motors and even a bit less current ripple. A 5-phase controller has only n-channel FETs to ground, which has 1/2 the heating of full bridge 3-phase ESCs. In addition, the zero phase sense is much easier since there's virtually no noise and/or phase shift for zero crossing detection. In fact, I don't even think a microprocessor is required since the logic is so simple (including open loop startup).

Naturally the motor needs 10, 20, or 30... slots in the stator.

Is anyone else working on this concept?

[Hell-e-Guy]

Some questions....
Where can one find a 5 phase BL Motor for R/C?
How is it possible to invert DC to AC with only n-CH FETs to ground?
A link to a document about 5 phase BL motors will be appreciated !

Thanks.

[z-matrix]

I've standardized on the voltage for 10S lipo or 35 1.2v cells, which will also be the standard for the auto industry, 42 volts. It's safe, yet removes much of the stress on our wiring and batteries (and associated extra weight) for higher power motors in .40, .60 and .90 conversions. And... ten 2200 lipos (10C) is much lower cost than the 4S6000 batteries.

The problem is lack of controllers in this segment of the market and high cost of those that are available.

It turns out 5-phase unipolar BLPM motors have about the same performance as 3-phase motors and even a bit less current ripple. A 5-phase controller has only n-channel FETs to ground, which has 1/2 the heating of full bridge 3-phase ESCs. In addition, the zero phase sense is much easier since there's virtually no noise and/or phase shift for zero crossing detection. In fact, I don't even think a microprocessor is required since the logic is so simple (including open loop startup).

Naturally the motor needs 10, 20, or 30... slots in the stator.

Is anyone else working on this concept?

I disagree.
And note that half bridges can be synched to damp the inductive energy.
Only very basic, low power esc uses open drain drivers.


In addition, the zero phase sense is much easier since there's virtually no noise and/or phase shift for zero crossing detection. In fact, I don't even think a microprocessor is required since the logic is so simple (including open loop startup).

well...
have you made a speed controller for 3 phase?
Its not that easy, and i doubt you can make a better 5 phase driver.
In a 3 phase sensorless bldc esc, the unused phase is used as emf,
think about which phase you are going to sense in a 5 phase motor :D
I whould guess why are those 5 phase controllers expensive.
No phase shift?
3 phase : 120 deg
5 phase : 72 deg

I'm desingning a 3 phase esc now, and did not even think about making 5 phase one, maeby one day...

[INewton]

I disagree.
And note that half bridges can be synched to damp the inductive energy.
Only very basic, low power esc uses open drain drivers.


I have flyback diodes that are between the inductive load and Vcc. Also I was thinking about using a dump cap instead of Vcc to speed the the energy discharge of the motor slot inductance and reclaim that energy with a small buck regulator.


well...
have you made a speed controller for 3 phase?
Its not that easy, and i doubt you can make a better 5 phase driver.
In a 3 phase sensorless bldc esc, the unused phase is used as emf,
think about which phase you are going to sense in a 5 phase motor :D
I whould guess why are those 5 phase controllers expensive.
No phase shift?
3 phase : 120 deg
5 phase : 72 deg

I'm desingning a 3 phase esc now, and did not even think about making 5 phase one, maeby one day...


That's what really nice about 5 phase. The two off-phases provide a beautiful differential input for back-emf (zero crossing)phase detection with virtually no noise...

[z-matrix]

I have flyback diodes that are between the inductive load and Vcc. Also I was thinking about using a dump cap instead of Vcc to speed the the energy discharge of the motor slot inductance and reclaim that energy with a small buck regulator.



That's what really nice about 5 phase. The two off-phases provide a beautiful differential input for back-emf (zero crossing)phase detection with virtually no noise...

i'm a bit interested,
could you post your idea on how whould you like to drive the 5 phase motor?
just a basic, like in 3 phase
1 +-x
2 x+-
.
.
.

will you wind it lrk ?
you said you wil use the differential of the 2 remaining phases - so you will use the other 3 ?

[INewton]

i'm a bit interested,
could you post your idea on how whould you like to drive the 5 phase motor?
just a basic, like in 3 phase
1 +-x
2 x+-
.
.
.

will you wind it lrk ?
you said you will use the differential of the 2 remaining phases - so you will use the other 3 ?


z,

Here's an excellent paper on the 5 phase unipolar motor (http://www.ece.tamu.edu/~empelab/papers/ias'99-tilak.pdf), but not the controller specifically. I believe it answer's all your questions. :)

The best slot/pole combos are 10/4, 20/8 and 30/12. They are wound very similar to lrk. Two or Three phases are always active in a FIFO drive pattern.

[jeffs555]

Here is a good article on multiphase brushless motors. It discusses 2,3,4, and 5 phase motors. http://www.powerpulse.net/features/techpaper.php?paperID=39

[z-matrix]

z,

Here's an excellent paper on the 5 phase unipolar motor (http://www.ece.tamu.edu/~empelab/papers/ias'99-tilak.pdf), but not the controller specifically. I believe it answer's all your questions. :)

The best slot/pole combos are 10/4, 20/8 and 30/12. They are wound very similar to lrk. Two or Three phases are always active in a FIFO drive pattern.

oh cool
http://www.powerpulse.net/features/archive/image/aa_102599d_fig1_large.gif
well the infos you want to use say,
in your preferred open drain configuration, 1 phase is active at a time, so i think you will have 4 phase floating with a 5 phase motor, not 2 ...

but go on, make it, if it works then all right.

[INewton]

oh cool
http://www.powerpulse.net/features/archive/image/aa_102599d_fig1_large.gif
well the infos you want to use say,
in your preferred open drain configuration, 1 phase is active at a time, so i think you will have 4 phase floating with a 5 phase motor, not 2 ...

but go on, make it, if it works then all right.


z,

Bottom right (2 or 3 on)... not the bottom left (only one on) in your reference. So 2 or 3 float in the 'preferred configuration' for zero crossing detection. :D

[z-matrix]

z,

Bottom right (2 or 3 on)... not the bottom left (only one on) in your reference. So 2 or 3 float in the 'preferred configuration' for zero crossing detection. :D


oh, ok, i thought you were doing the lower left...

have you got some success with it already?
what MCU do you prefer?

[INewton]

oh, ok, i thought you were doing the lower left...

have you got some success with it already?
what MCU do you prefer?


I'm using a logic array (FPGA)... Just to be a bit different (Actually to control the communation better at higher RPM). I'm using maximum current sense on windings for PWM.

Success? Well, yes. I started building.

[pbp_et]

Magnificent idea, BUT you will have a smaller effeciency with unipolar drive referred to bipolar. You will only use half the magnetizing curve of the stator teeth = one quarter of the full power possible, but you will spend half the electric power possible.

[pbp_et]

The idea of using 42 Volts instead of 12 is very smart. Ohmic losses in cables and MOSfets wil be down 12 times for the same output in Watts

[INewton]

Magnificent idea, BUT you will have a smaller effeciency with unipolar drive referred to bipolar. You will only use half the magnetizing curve of the stator teeth = one quarter of the full power possible, but you will spend half the electric power possible.

For a 3 phase bipolar wye wound motor, 2 of 3 windings are always active. For a 5 phase uniploar motor, 3 of 5 windings are always active. That's .66 vs .6, so magnetic utilization is almost the same. However the ripple in the 5 phase is 7% vs 14% for the 3 phase wye. So the maximum power turns out to be only about 5% less for a 5 phase unipolar motor and the efficiency a bit higher for the 5 phase (less ripple).

My goal is to eliminate the need for more expensive, less efficient (2 FETs in series), slower rise-time bipolar drive.

I also looking at 3 phase SRM (switched reluctance motors) that are inherently unipolar and have a better power to weight ratio than BLPM motors.