Maker Pro
Maker Pro

An interesting BLDC controller concept

C

CC

Jan 1, 1970
0
Hi:

An idea has been floating in my head for a few weeks, ever since
building an analog block commutation drive for a brushless DC (BLDC) motor.

It is a way to implement a sinewave drive on a BLDC motor which lacks an
encoder. The sine drives that I have seen so far require a high
resolution encoder to provide angular information from which to
synthesize the sinewave.

My idea is to use a PLL to multiply the hall sensor (or back EMF derived
position information if a non-sensor motor) frequency by some multiple
to create a sine generation clock.

Of course, there would be no dependable angular frequency to lock on
before the motor is running. So startup would be performed with simple
block (6-step, trapezoidal, so many terms...) commutation.

Once the motor was running and the PLL could track, the drive could
morph to sinewave mode.

Acceleration and deceleration of the motor would result in small
glitches in the sinewaves, but would still probably produce much less
harmonic current than plain block commutation. Very strong acceleration
could perhaps use block commutation temporarily.

Anyone ever tried this? Obviously not applicable to position servo
applications.


I just wanted to get it out in the public domain in case it hadn't been
patented yet ;-)


Good day!
 
J

John Woodgate

Jan 1, 1970
0
dated Sun said:
I just wanted to get it out in the public domain in case it hadn't been
patented yet ;-)

What has it got to do with Lebanon? (;-)
 
CC said:
Hi:

An idea has been floating in my head for a few weeks, ever since
building an analog block commutation drive for a brushless DC (BLDC) motor.

It is a way to implement a sinewave drive on a BLDC motor which lacks an
encoder. The sine drives that I have seen so far require a high
resolution encoder to provide angular information from which to
synthesize the sinewave.

My idea is to use a PLL to multiply the hall sensor (or back EMF derived
position information if a non-sensor motor) frequency by some multiple
to create a sine generation clock.

Of course, there would be no dependable angular frequency to lock on
before the motor is running. So startup would be performed with simple
block (6-step, trapezoidal, so many terms...) commutation.

Once the motor was running and the PLL could track, the drive could
morph to sinewave mode.

Acceleration and deceleration of the motor would result in small
glitches in the sinewaves, but would still probably produce much less
harmonic current than plain block commutation. Very strong acceleration
could perhaps use block commutation temporarily.

Anyone ever tried this? Obviously not applicable to position servo
applications.


I just wanted to get it out in the public domain in case it hadn't been
patented yet ;-)

Sorry Chris. Save the champagne for the next idea.

You've just re-invented the microstepped stepper motor. If you are
careful about your acceleration and deceleration profiles, you don't
really need the encoder.
 
C

CC

Jan 1, 1970
0
Sorry Chris. Save the champagne for the next idea.

You've just re-invented the microstepped stepper motor. If you are
careful about your acceleration and deceleration profiles, you don't
really need the encoder.


Interesting.
 
J

joseph2k

Jan 1, 1970
0
CC said:
Hi:

An idea has been floating in my head for a few weeks, ever since
building an analog block commutation drive for a brushless DC (BLDC)
motor.

It is a way to implement a sinewave drive on a BLDC motor which lacks an
encoder. The sine drives that I have seen so far require a high
resolution encoder to provide angular information from which to
synthesize the sinewave.

My idea is to use a PLL to multiply the hall sensor (or back EMF derived
position information if a non-sensor motor) frequency by some multiple
to create a sine generation clock.

Of course, there would be no dependable angular frequency to lock on
before the motor is running. So startup would be performed with simple
block (6-step, trapezoidal, so many terms...) commutation.

Once the motor was running and the PLL could track, the drive could
morph to sinewave mode.

Acceleration and deceleration of the motor would result in small
glitches in the sinewaves, but would still probably produce much less
harmonic current than plain block commutation. Very strong acceleration
could perhaps use block commutation temporarily.

Anyone ever tried this? Obviously not applicable to position servo
applications.


I just wanted to get it out in the public domain in case it hadn't been
patented yet ;-)


Good day!

How much power are you talking about? Why the fascination with sine waves?
There are a variety of techniques, with working ranges from milliwatt to
megawatt. Not all techniques are suitable to all power ranges. See also
variable frequency drives (VFD) and adjustable frequency motor controls
(AFMC). Now, what is it you really want?
 
C

CC

Jan 1, 1970
0
joseph2k said:
How much power are you talking about?

Doesn't matter, the technique is applicable to any power level.

Why the fascination with sine waves?

Torque ripple and harmonic currents which increase losses (heating) are
high using block commutation in BLDCs. Sinewaves are the form of the
back emf waveforms, so sinewave drive results in minimal harmonic
currents, as well as smooth torque.
There are a variety of techniques, with working ranges from milliwatt to
megawatt. Not all techniques are suitable to all power ranges. See also
variable frequency drives (VFD) and adjustable frequency motor controls
(AFMC). Now, what is it you really want?

BLDC motors are a special class of synchronous motors and by definition
operate using a form of VFD. VFD as a commercially used term are mainly
applicable to induction motors.

The point is that the market offers block commutation BLDC motor drives,
as well as sinewave drives. The sinewave drives exist for the reasons
mentioned and so are considered for applications where torque ripple
must be minimized and/or harmonic current related losses must be
minimized. However, they require extra cost and complexity of the
motor, which must be coupled to a high resolution encoder.

What I want is to explore these concepts and drives both theoretically
and experimentally for learning purposes, as well as for some present
applications with which I am involved. I have already constructed a
block commutation BLDC drive, as well as employing some commercial drives.

My next interest is in understanding harmonic currents in more detail,
and how to reduce them. This may possibly entail constructing a
sinewave drive.


Good day!
 
V

vasile

Jan 1, 1970
0
CC said:
Doesn't matter, the technique is applicable to any power level.

Why the fascination with sine waves?

Torque ripple and harmonic currents which increase losses (heating) are
high using block commutation in BLDCs. Sinewaves are the form of the
back emf waveforms, so sinewave drive results in minimal harmonic
currents, as well as smooth torque.


True, but usually because is much cheaper, those kind of motor are
using pseudo-sinewaves (trapezoidal or rectangular with sinus shape
only). That's because all motors have inductive coils and the shape is
integrated.

Using a PLL to generate a power sinusoidal signal (with a serious
load) will be a difficult task due the PLL working principle. Any spike
or load variation will track the PLL ouside the capture band and
everything will finish into a mess.
I hope I'm wrong (for your future patented methode).
:)

greetings,
Vasile
 
D

David Brown

Jan 1, 1970
0
vasile said:
True, but usually because is much cheaper, those kind of motor are
using pseudo-sinewaves (trapezoidal or rectangular with sinus shape
only). That's because all motors have inductive coils and the shape is
integrated.

There are (I believe) two main forms of windings for this sort of three
phase brushless DC motors - BLDC windings, optimized to be run by
trapezoidal waveforms, and PMSM windings, optimized for sinusoidal
waveforms. Of course, you can use either waveform for either type of
motor winding, at a slight efficiency loss, but there is no point in
striving to get a nice sinewave unless you are sure your motor matches
it, and unless you can be sure of very smooth running.

As for using a PLL based on the hall effect sensors - you do realise
that the 6 switching points from a typical motor hall effect sensor
setup are not accurately at 60 degree steps? You are going to get some
serious jitter effects if you try and multiply the step rate, unless you
just use a single step (i.e., one signal per electrical revolution).
 
C

Chris Carlen

Jan 1, 1970
0
David said:
There are (I believe) two main forms of windings for this sort of three
phase brushless DC motors - BLDC windings, optimized to be run by
trapezoidal waveforms, and PMSM windings, optimized for sinusoidal
waveforms. Of course, you can use either waveform for either type of
motor winding, at a slight efficiency loss, but there is no point in
striving to get a nice sinewave unless you are sure your motor matches
it, and unless you can be sure of very smooth running.

Are you referring to slotted vs. slotless stators? If so, which of your
designations "BLDC" and "PMSM" is slotted/slotless? Also, what does the
abbreviation PMSM mean?

I had a discussion with a sales engineer at http://www.aerotech.com
about slotless/slotted motors. It wasn't clear from this discussion
though which should be driven with trapezoid vs. sine.

There also doesn't seem to be clear guidance from motor manufacturers
about this.

One thing is certain, my Maxon motors produce perfect sinewave back EMF
waveforms. I wonder if a slotted motor would produce something more
"trapezoidal?"
As for using a PLL based on the hall effect sensors - you do realise
that the 6 switching points from a typical motor hall effect sensor
setup are not accurately at 60 degree steps? You are going to get some
serious jitter effects if you try and multiply the step rate, unless you
just use a single step (i.e., one signal per electrical revolution).

Yes, they are not perfectly accurate. I think I was inclining just
toward looking at a single pulse/rev. Though now it brings up an
interesting refinement:

A controller could calibrate itself to the hall misalignment by running
the motor in a PLL and measuring the hall timings. Now I have to think
about how this information could be used later to correct the hall
timings...but there must be a way. Probably would have to be a digital
PLL implementation of some sort.



--
Good day!

________________________________________
Christopher R. Carlen
Principal Laser&Electronics Technologist
Sandia National Laboratories CA USA
[email protected]
NOTE, delete texts: "RemoveThis" and
"BOGUS" from email address to reply.
 
J

John

Jan 1, 1970
0
Hi Chris,

One thing is certain, my Maxon motors produce perfect sinewave back EMF
waveforms. I wonder if a slotted motor would produce something more
"trapezoidal?"

ST has some interesting app notes using sine-wave drive on BLDC motors.
Here's one:

http://www.st.com/stonline/products/literature/an/12375.pdf

look at some of their others:

http://mcu.st.com/mcu/modules.php?name=mcu&file=familiesdocs&FAM=15
#Application%20Note

Are you trying to build flexible chopper wheels for your LASER systems?
I've built a phase-lockable motor controller that lets you go up/down in
speed and allows for nearly arbitrary phase offsets for such systems (DC
motor though, not BLDC).

Regards.
R.
 
D

David Brown

Jan 1, 1970
0
Chris said:
Are you referring to slotted vs. slotless stators? If so, which of your
designations "BLDC" and "PMSM" is slotted/slotless? Also, what does the
abbreviation PMSM mean?

I haven't heard the terms "slotted" and "slottless", so I'm afraid I
can't help you.

"PMSM" is "permanent magnet synchronous motor".
I had a discussion with a sales engineer at http://www.aerotech.com
about slotless/slotted motors. It wasn't clear from this discussion
though which should be driven with trapezoid vs. sine.

There also doesn't seem to be clear guidance from motor manufacturers
about this.

That's certainly true, at least for all the motors I've seen.
One thing is certain, my Maxon motors produce perfect sinewave back EMF
waveforms. I wonder if a slotted motor would produce something more
"trapezoidal?"

Actually Maxon motors can be driven well with either form, and it is not
clear from their documentation (that I've read) which is optimal. I
think they work well enough with either kind of driving.
 
C

Chris Carlen

Jan 1, 1970
0
David said:
I haven't heard the terms "slotted" and "slottless", so I'm afraid I
can't help you.

Slotted motors have laminated stators with actual slots to hold the
windings. Slotless motors have no such slots. How the winding is done
in a slotless motor looks something like Maxon's ironless rotor design
for permanent magnet DC motors. Pictures of this and the slotless BLDC
construction appear in their catalog.
"PMSM" is "permanent magnet synchronous motor".

Yes, but what is the physical difference between BLDC and PMSM?



--
Good day!

________________________________________
Christopher R. Carlen
Principal Laser&Electronics Technologist
Sandia National Laboratories CA USA
[email protected]
NOTE, delete texts: "RemoveThis" and
"BOGUS" from email address to reply.
 
D

David Brown

Jan 1, 1970
0
Chris said:
Slotted motors have laminated stators with actual slots to hold the
windings. Slotless motors have no such slots. How the winding is done
in a slotless motor looks something like Maxon's ironless rotor design
for permanent magnet DC motors. Pictures of this and the slotless BLDC
construction appear in their catalog.


Yes, but what is the physical difference between BLDC and PMSM?

Freescale have some quite good overviews:
http://www.freescale.com/webapp/sps/site/homepage.jsp?nodeId=02nQXG

The physical difference between the BLDC and the PMSM is a minor change
in the winding topology to make the former most efficient with
trapezoidal waveforms, and the later most efficient with sinusoidal
waveforms.
 
Top