seems to be a reasonable solution, just need to have 1 winding with twice
the impedance, and one with half the impedance of the other 2 wich means
1/sqrt(2) turns, shame its not half the number of turns, as I could just put
them 4 parts of the windings in series / parellel. oh and one phase voltage
needs to be slightly higher.

I just played about with the values.

Colin =^.^=

colin wrote:

seems to be a reasonable solution, just need to have 1 winding with twice
the impedance, and one with half the impedance of the other 2 wich means
1/sqrt(2) turns, shame its not half the number of turns, as I could just put
them 4 parts of the windings in series / parellel. oh and one phase voltage
needs to be slightly higher.

I just played about with the values.

Colin =^.^=



Four phase? Where?
If you have a motor with 5 or 6 wires and have determined that each
winding has a tap (common tap in the 5 wire case), you have a stepper
motor - *not* a "four phase" motor.
BTW, theer is a way of running it with 2 phase 90 degree drive (using
a resistor and a capacitor); the secrets are usually published by the
stepper motor company.

"Robert Baer" wrote in message
link.net...
colin wrote:

seems to be a reasonable solution, just need to have 1 winding with twice
the impedance, and one with half the impedance of the other 2 wich means
1/sqrt(2) turns, shame its not half the number of turns, as I could just
put them 4 parts of the windings in series / parellel. oh and one phase
voltage needs to be slightly higher.

I just played about with the values.

Colin =^.^=
Four phase? Where?
If you have a motor with 5 or 6 wires and have determined that each
winding has a tap (common tap in the 5 wire case), you have a stepper
motor - *not* a "four phase" motor.
BTW, theer is a way of running it with 2 phase 90 degree drive (using a
resistor and a capacitor); the secrets are usually published by the
stepper motor company.

its a 4 phase motor I wound it myself, well im in the procees of doing so
....
it had 16 slots so trying to fit 3 phase windings in seemed awkward so I hit
upon 4 phase,
where each winding is shifted by one slot or 45 electical degrees on the
stator,
making 2 pole pairs.
it makes for a neat winding arangment,
after all I had 4 pwm half bridge channels,
but then I realised 4phase isnt as easy to do as 3phase.

this is not a bi phase stepper motor.

its for small variable (v high) speed reversable drive.

if you look at the plot you see there are 8 power peaks per cycle,
hence 4 true phases.
that is power plot btw not voltage,
maybe thats confusing if you dont notice its power plot,
but that was most useful to compare as the impedances arnt the same.

each winding is in 4 parts (ie 2 pole pairs)
but I since realised they are not all on the same magnetic path so putting
them in series
or parallel is not quite like doing the same with turns on a transformer.

Colin =^.^=

colin wrote:

its a 4 phase motor I wound it myself, well im in the procees of doing so
...
it had 16 slots so trying to fit 3 phase windings in seemed awkward so I hit
upon 4 phase,
where each winding is shifted by one slot or 45 electical degrees on the
stator,
making 2 pole pairs.
(snip)

each winding is in 4 parts (ie 2 pole pairs)
but I since realised they are not all on the same magnetic path so putting
them in series
or parallel is not quite like doing the same with turns on a transformer.

If you had 4 isolated AC phases, you could connect 8
windings in a phasor octagon (half of each phase winding as
parallel sides of the octagon) with the 4 phases connected
across opposite corners. But that doesn't work for 4 half
bridges, without 4 isolation transformers. I think this
might smooth out the torque pulsations, by sharing current,
phase to phase, around the circle. This method outs each
drive phase across two parallel paths, each containing 4
coils in series, so it is a good way to use low count windings.

The Y equivalent is an 8 branch star, with the same need to
drive 4 pairs of isolated phases. This arrangement does not
alter the torque pulsations, because, even if left floating
(or the star, not even connected together), symmetry forces
the center node to have no AC.

But if you just wire it up as half of that star (only one
side of each center tapped pair), you can common the
lopsided center and drive each of the 4 ends with one of
your half bridges and have the same as the last.

But if you tie the 4 'center' ends together, but float this
node, the ripple will be reduced as this node finds its
voltage as a combination of the 4 drive phases returning
through each other. instead of through the common center
node. I could be all wet, here. I admit that 4 phases
makes my head hurt.

"John Popelish" wrote in message
. ..
colin wrote:

its a 4 phase motor I wound it myself, well im in the procees of doing so
...
it had 16 slots so trying to fit 3 phase windings in seemed awkward so I
hit upon 4 phase,
where each winding is shifted by one slot or 45 electical degrees on the
stator,
making 2 pole pairs.
(snip)

each winding is in 4 parts (ie 2 pole pairs)
but I since realised they are not all on the same magnetic path so
putting them in series
or parallel is not quite like doing the same with turns on a transformer.

If you had 4 isolated AC phases, you could connect 8 windings in a phasor
octagon (half of each phase winding as parallel sides of the octagon) with
the 4 phases connected across opposite corners. But that doesn't work for
4 half bridges, without 4 isolation transformers. I think this might
smooth out the torque pulsations, by sharing current, phase to phase,
around the circle. This method outs each drive phase across two parallel
paths, each containing 4 coils in series, so it is a good way to use low
count windings.

The Y equivalent is an 8 branch star, with the same need to drive 4 pairs
of isolated phases. This arrangement does not alter the torque
pulsations, because, even if left floating (or the star, not even
connected together), symmetry forces the center node to have no AC.

But if you just wire it up as half of that star (only one side of each
center tapped pair), you can common the lopsided center and drive each of
the 4 ends with one of your half bridges and have the same as the last.

But if you tie the 4 'center' ends together, but float this node, the
ripple will be reduced as this node finds its voltage as a combination of
the 4 drive phases returning through each other. instead of through the
common center node. I could be all wet, here. I admit that 4 phases
makes my head hurt.

thanks for that in depth reply

yes it made my head hurt a bit too, heck even 3 phase is tricky,
but at least with 3 ph if you get one of the phase coils reveresed,
or swap 2 phases the worst that happens is it runs backwards,
wich my software determines at startup now anyway lol.

with 4 ph its rather nasty if you get one phase polarity wrong or swap 2.

also if you have half of the 8 star, you have to be carefull wich 4 you
chose,
no choice of 4 seems very ideal wich is where I decide to part with that
idea,
having got halfway through rewinding it.

I dont think you realy gain anything worth it with 4 phases tbh,
it was just convenience with 16 slots,
im not sure if my idea above is any better with regard to pulse ripple
although the power consumption looks fairly smooth.
it does mean however that the windings are quite well utilised,
as each segment is energised by one winding,
rather than the sum of two windings of different phase wich partially add or
subtract.

actually i did have a good idea for 6 phase where you have a delta conected
inner set of windings
and an outer star conected set
this would be ideal if i had a 24 segment rotor

for now il just run it up as per the simulation and see how good it is,
I only put 6 turns on each segment,
as its low voltage high frequency.
if its an improvement i might do a better job or perhaps find a better motor
to rewind,
or one that I dont have to. it might even do with no extra work,
although I miss calculated the number of turns forgetting to take sqrt,
so the current needed may make my pwm filters smoke some more, but theyr
used to it by now.
the fets at least are 50amp jobs.

ive done the 4 windings, ive just got to get my head round the right way to
connect them.
does my head in especialy at this hour.

Colin =^.^=

John Popelish wrote:
(snip)
But if you just wire it up as half of that star (only one side of each
center tapped pair), you can common the lopsided center and drive each
of the 4 ends with one of your half bridges and have the same as the last.

But if you tie the 4 'center' ends together, but float this node, the
ripple will be reduced as this node finds its voltage as a combination
of the 4 drive phases returning through each other. instead of through
the common center node. I could be all wet, here. I admit that 4
phases makes my head hurt.

This doesn't work. you get two large and two smaller phase
currents.

But you can tie the 'center' end of say, the 0 and 45 degree
phases, and the 'center' end of the 90 and 135 degree phases
ans two separate floating nodes, and get phase sharing
effect for these pairs that keeps all the phase currents equal.

John Popelish wrote:
John Popelish wrote:
(snip)
But if you just wire it up as half of that star (only one side of each
center tapped pair), you can common the lopsided center and drive each
of the 4 ends with one of your half bridges and have the same as the
last.

But if you tie the 4 'center' ends together, but float this node, the
ripple will be reduced as this node finds its voltage as a combination
of the 4 drive phases returning through each other. instead of through
the common center node. I could be all wet, here. I admit that 4
phases makes my head hurt.

This doesn't work. you get two large and two smaller phase currents.

But you can tie the 'center' end of say, the 0 and 45 degree phases, and
the 'center' end of the 90 and 135 degree phases ans two separate
floating nodes, and get phase sharing effect for these pairs that keeps
all the phase currents equal.

Damn. Forget that. With only two windings in series, their
currents are certainly equal, because they are identical in
all respects, including phase.

"John Popelish" wrote in message
. ..
John Popelish wrote:
John Popelish wrote:
(snip)
But if you just wire it up as half of that star (only one side of each
center tapped pair), you can common the lopsided center and drive each
of the 4 ends with one of your half bridges and have the same as the
last.

But if you tie the 4 'center' ends together, but float this node, the
ripple will be reduced as this node finds its voltage as a combination
of the 4 drive phases returning through each other. instead of through
the common center node. I could be all wet, here. I admit that 4
phases makes my head hurt.

This doesn't work. you get two large and two smaller phase currents.

But you can tie the 'center' end of say, the 0 and 45 degree phases, and
the 'center' end of the 90 and 135 degree phases ans two separate
floating nodes, and get phase sharing effect for these pairs that keeps
all the phase currents equal.

Damn. Forget that. With only two windings in series, their currents are
certainly equal, because they are identical in all respects, including
phase.

yeah sounds like the sort of thing that went on in side my head,
but while I was in the middle of winding too lol.

so its all a moot point now, as I only implmeneted 3 of the pwm half bridge
chanels anyway
for 3 phase on 16 slots, having 1 phase accros one segment and the 2 other
phases spread out lopsided over 2 segments
would probably give me a good megnetic waveform,
but rather than unwind it again and waste my wire i think il try conecting
it as my simulation.

my cats are always fascinated when I try and wind a coil,
they always try and 'help' the finer the wire the more they seem to enjoy
it,
if its so fine they can hardly see it they go nuts.

I have another possible question ... I might measure this to be sure ..

If I have 2 pole pairs, so 1 phase winding has 4 coils
say each coil is 1mh inductance,
If I put them all in series, is it just additive as if they are seperate
coils?
so 4 in series = 4mh ?
or is it increasing the turns on the same coil therefore 16mh?

or is it just each pole pair act as one coil so they add up as increasing
turns for only each pair
(this is my gues) so it would be 8mh ?
of course with the rotor in place this seriously changes the inductance and
linkage,
but the rotor has a big shorted turn on it when stationary so im dubious
about the true reading.

my simulation needs 4,8,16 mh or those proportions.
hard to get those by arranging the coils in diferent configurations.

ive ended up with 4 turns , 6 turns and 8 turns on each coil all in series.

I need to check that ive got a sensible conection now, i wonder how i can do
this,
a magnet seems to atract itself to any of the poles with 1 phase energised,
although ive only got a couple of amps,
a screwdriver seems to suggest the poles are energised as expected but
doesnt tell me the polarity.

maybe good old iron filings ...


Colin =^.^=

colin wrote:
(snip)

I have another possible question ... I might measure this to be sure ..

If I have 2 pole pairs, so 1 phase winding has 4 coils
say each coil is 1mh inductance,
If I put them all in series, is it just additive as if they are seperate
coils?
so 4 in series = 4mh ?
or is it increasing the turns on the same coil therefore 16mh?

Somewhere in between these two ideal cases, I think. Some
of the flux from each coil also passes through its
neighbors, so there is a bit more than just the sum.

or is it just each pole pair act as one coil so they add up as increasing
turns for only each pair
(this is my gues) so it would be 8mh ?
of course with the rotor in place this seriously changes the inductance and
linkage,
but the rotor has a big shorted turn on it when stationary so im dubious
about the true reading.

my simulation needs 4,8,16 mh or those proportions.
hard to get those by arranging the coils in diferent configurations.

ive ended up with 4 turns , 6 turns and 8 turns on each coil all in series.

I need to check that ive got a sensible conection now, i wonder how i can do
this,
a magnet seems to atract itself to any of the poles with 1 phase energised,
although ive only got a couple of amps,
a screwdriver seems to suggest the poles are energised as expected but
doesnt tell me the polarity.

I would sit the stator with the axis pointing vertically,
and place a small magnetic compass in the center, and see if
a slow field rotation rotated the needle smoothly. Wait.
for a 4 pole winding, you need a 4 pole compass. I think
you will have to make one of those. ;-)

I have a sheet of magnetic field viewing material that
changes color from green to dark green as it sees field. it
is sensitive enough to show the magnetic strip on the back
of a credit card. I wonder what a disc of that would do in
the center of the rotor.
http://cgi.ebay.com/Magnetic-Field-V...mZ160125398593
Much neater than iron filings.

"John Popelish" wrote in message
. ..
colin wrote:
(snip)

I would sit the stator with the axis pointing vertically, and place a
small magnetic compass in the center, and see if a slow field rotation
rotated the needle smoothly. Wait. for a 4 pole winding, you need a 4
pole compass. I think you will have to make one of those. ;-)

yes its 2 pole pairs, so i need a needle with 2 N lol,
I tried some magnets on a shaft but it wasnt very sucesfull,
no way of letting it rotate freely and yet keep it central enough.

I have a sheet of magnetic field viewing material that changes color from
green to dark green as it sees field. it is sensitive enough to show the
magnetic strip on the back of a credit card. I wonder what a disc of that
would do in the center of the rotor.
http://cgi.ebay.com/Magnetic-Field-V...mZ160125398593
Much neater than iron filings.


ooh that looks cool, i tried getting it near my crt monitor but it didnt
show up,
needs to be inside the stator obviously.

I sat down with a strong cup of coffee and worked out where all the dots on
each winding should get connected to.
I also put the dots on my simulation too, then windows decided it was time
to reboot my pc after an update while I was making the cup of coffee and I
lost that .

ive wired it up now, il just suck it and see.

Colin =^.^=

"John Popelish" wrote in message
. ..
colin wrote:
(snip)

I have a sheet of magnetic field viewing material that changes color from
green to dark green as it sees field. it is sensitive enough to show the
magnetic strip on the back of a credit card. I wonder what a disc of that
would do in the center of the rotor.
http://cgi.ebay.com/Magnetic-Field-V...mZ160125398593
Much neater than iron filings.

can you cut that stuff ?
not sure if its got tiny bubbles with fillings in or the whole sheet is one
space.

I have tried iron fillings on a plastic tray wich just fitted inside the
stator
however the iron filings congregate between the segment poles and stay
there,
just dancing around a bit. i was exciting it at 12rpm.

gues i need more iron fillings, i just collected a few off the dril stand
and around my vice with a magnet lol.

Ive tested it with the induction rotor, and it does go round,
however ive not been able to get it upto any speed,
with only 12 in each slot im not toooo surprised though,
although this was with 10 amps.
my gues each slot originally had 1000 turns for 250v and was rated at 0.2
amps (6w)

however I was hoping to ascertain if this would work before I spent a few
days
winding about 30 turns onto each coil.

I wound them coils on a former first then squezed them into the slots.

are there any other significant problems associated with running an
induction motor at high speed ?

Colin =^.^=

colin wrote:
"John Popelish" wrote in message
. ..
colin wrote:
(snip)

I have a sheet of magnetic field viewing material that changes color from
green to dark green as it sees field. it is sensitive enough to show the
magnetic strip on the back of a credit card. I wonder what a disc of that
would do in the center of the rotor.
http://cgi.ebay.com/Magnetic-Field-V...mZ160125398593
Much neater than iron filings.

can you cut that stuff ?

I haven't cut mine, but the piece I have looks like it has a
cut edge. It contains some sort of micro capsules, I think.
I've misplaced it at the moment, or I would examine it
more closely.

not sure if its got tiny bubbles with fillings in or the whole sheet is one
space.

It didn't appear to have any sealed edge.

I have tried iron fillings on a plastic tray wich just fitted inside the
stator
however the iron filings congregate between the segment poles and stay
there,
just dancing around a bit. i was exciting it at 12rpm.

gues i need more iron fillings, i just collected a few off the dril stand
and around my vice with a magnet lol.

Ive tested it with the induction rotor, and it does go round,
however ive not been able to get it upto any speed,
with only 12 in each slot im not toooo surprised though,
although this was with 10 amps.
my gues each slot originally had 1000 turns for 250v and was rated at 0.2
amps (6w)

however I was hoping to ascertain if this would work before I spent a few
days
winding about 30 turns onto each coil.

I wound them coils on a former first then squezed them into the slots.

are there any other significant problems associated with running an
induction motor at high speed ?

I think eddy current and hysteresis heating of the stator
laminations are the biggest limitations, after you get past
mechanical resonances it the rotor that make the shaft bend.
I have seen some pretty high speed 2 pole induction motors.

http://scitation.aip.org/getabs/serv...cvips&gifs=yes
http://adsabs.harvard.edu/abs/1991crin.rept.....M
http://stinet.dtic.mil/oai/oai?&verb...er =AD0601968

"John Popelish" wrote in message
. ..
colin wrote:
(snip)

are there any other significant problems associated with running an
induction motor at high speed ?

I think eddy current and hysteresis heating of the stator laminations are
the biggest limitations, after you get past mechanical resonances it the
rotor that make the shaft bend. I have seen some pretty high speed 2 pole
induction motors.

http://scitation.aip.org/getabs/serv...cvips&gifs=yes
http://adsabs.harvard.edu/abs/1991crin.rept.....M
http://stinet.dtic.mil/oai/oai?&verb...er =AD0601968

yeah I gues iron losses go up, im running about 10-20x normal speed wich is
stil less than 1khz,
dont losses go up linearly with frequency ?
if iron losses are similar to copper losses I gues il have to derate it
rather a lot.

at least the rotor should be electrically similar as it only sees the slip
frequency.

those motors look quite cool, although I only see the abstract.

im gona get some of that material, i got some magnets from the same place
too.

im also gona see if i can re magnetise the rotor of my current motor to see
if it can be made to be more of a sinewave so have less torque ripple. might
be easier, although might take a few tries.

Colin =^.^=