Hej Sean H. Breheny. Tack f=F6r ditt meddelande 02:47 2002-08-05 enligt=
 nedan:
>Hi Morgan,
>
>Thanks for the quick reply. As for bipolarity, I think that this is
>completely equivalent to a bipolar drive (just like feeding a brushed motor
>with an H bridge and flipping the polarity around). Of course you are
>right, it is probably more efficient if you include more points of the sine
>wave (you wouldn't really want off, you would want 0 volts, which is
>different),

I believe binding to zero only cause power losses.
Not binding it to zero volts will let the magnetic play more freely.  I=
 think freewheeling is better regarding efficiency, heat up, and putput=
 power.
Freewheeling between + and - (and PWM, current sensing, lots of=
 calculations...) is how Variable frequency drives drive 3-phase async=
 motors.

> but I don't think that a square wave would cause 7 or 40 times
>more current.

Right.

Either commutation sequence, levels, or the phasing with sensors to=
 commutation are wrong, or both wrong...


I wrote

>L1: +++000---000+++000---000
>L2: +000---000+++000---000++
>L3: 00---000+++000---000+++0
>
>Not seeing it clear though, it is very late here...

Above is 60=B0 phase shift...
Um, i think it is better to use 120=B0 phase shift ;)
  =20
L1: +++000---000+++000---000
L2: 00---000+++000---000+++0
L3: -000+++000---000+++000--

As a test could you program a pic to output that sequence, and feed the pic=
 with a variable oscillator (for easy stepless adjustment), and the drivers=
 with variable voltage, and see if you can get the motor freerunning fast=
 with low power?  (Not using hall sensors, just see of we can push the rotor=
 correctly)  Then we probably have the correct commutation :)

Next you compare the sensor signals with the transistor drive using a=
 oscilloscope, while running, and revise the PIC code.

/Morgan

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads