Maker Pro
Maker Pro

searching a brushless DC startup procedure

E

endrophie

Jan 1, 1970
0
Hello,

if you have a look to a modern brushless controller for radio
controlled flying models you can see

- that the motor has a very short startup time until the right moving
(about 100-200ms)

- that no initial alignment for a known position is done

- that a backward rotation only about 30 - 60 ms after starting is
recogniced and the motor is running in the right direction

- that because of the used 8 bit Microcontrollers there is no extrem
complex algorithm possible

- that the controller works probably in voltage controll

Does anywhere have information about the used startprocedure or a
webside/usegroup where mor information is available?

Best regards

Endrophie
 
A

Andrew Holme

Jan 1, 1970
0
endrophie said:
Hello,

if you have a look to a modern brushless controller for radio
controlled flying models you can see

- that the motor has a very short startup time until the right moving
(about 100-200ms)

- that no initial alignment for a known position is done

- that a backward rotation only about 30 - 60 ms after starting is
recogniced and the motor is running in the right direction

- that because of the used 8 bit Microcontrollers there is no extrem
complex algorithm possible

- that the controller works probably in voltage controll

Does anywhere have information about the used startprocedure or a
webside/usegroup where mor information is available?

Best regards

Endrophie

You can use a technique known as "six stepping" to get the motor up and
running. Once it has some momentum, the switching transistors are
synchronised to the shaft position either using Hall Effect sensors or by
monitoring back emf e.g. using a microcontroller with built in A-D
connverters. This is electronic commutation.

I recently built a very simple six-stepping controller for a 3-phase
brushless DC motor taken from an old hard drive using a 74HC164 shift
register and 74HC132 quad schmitt nand. One of the nand gates was a
variable frequency RC oscillator acting as the speed control.

The pull-up / pull-down transistors must be driven as follows :

pull ups (1 = on)
Phase1 = 110000
Phase2 = 001100
Phase3 = 000011

pull downs (1 = on)
Phase1 = 000110
Phase2 = 100001
Phase3 = 011000

The motor will only start from stationary if the speed control is set to
minimum. The speed can then be increased fairly rapidly but not too quickly
or else the motor will stall. I haven't tried the back-emf stuff but I know
that's how it's done e.g. on http://www.speedy-bl.com/speedybl-e.htm
 
E

endrophie

Jan 1, 1970
0
Andrew Holme said:
You can use a technique known as "six stepping" to get the motor up and
running. Once it has some momentum, the switching transistors are
synchronised to the shaft position either using Hall Effect sensors or by
monitoring back emf e.g. using a microcontroller with built in A-D
connverters. This is electronic commutation.

I recently built a very simple six-stepping controller for a 3-phase
brushless DC motor taken from an old hard drive using a 74HC164 shift
register and 74HC132 quad schmitt nand. One of the nand gates was a
variable frequency RC oscillator acting as the speed control.

The pull-up / pull-down transistors must be driven as follows :

pull ups (1 = on)
Phase1 = 110000
Phase2 = 001100
Phase3 = 000011

pull downs (1 = on)
Phase1 = 000110
Phase2 = 100001
Phase3 = 011000

The motor will only start from stationary if the speed control is set to
minimum. The speed can then be increased fairly rapidly but not too quickly
or else the motor will stall. I haven't tried the back-emf stuff but I know
that's how it's done e.g. on http://www.speedy-bl.com/speedybl-e.htm


Hello,

ste controller must work in sensorless mode. The problem is, that when
you have a very slow rotation speer, the BEMF has a littly amplitude.
And if you a starting sequenze to the motor, it should rotate reliable
in the right direction. So it must be possible to detect near zero
speed the direction and the rotor position (but without any strategies
who need the current).

Best regards

Endrophie
 
endrophie said:
ste controller must work in sensorless mode. The problem is, that when
you have a very slow rotation speer, the BEMF has a littly amplitude.
And if you a starting sequenze to the motor, it should rotate reliable
in the right direction.

Six-stepping works reliably at low speed without feedback.
The six-step sequence dictates the direction of rotation.
So it must be possible to detect near zero
speed the direction and the rotor position (but without any strategies
who need the current).

What else can you measure apart from voltage and current?
You would need to mount a sensor on the shaft.
 
Top