Re: [CAD_CAM_EDM_DRO] RE: Stepper questions

Roy J. Tellason wrote:

>On Wednesday 03 May 2006 01:05 pm, Jon Elson wrote:
>
>
>>The two halves of a stepper phase coil (on 6- and 8-wire motors) DO
>>interact. They are usually wound on the same coil bobbin, and act as two
>>windings of a transformer. If you have twice the turns, then the inductance
>>is squared. Because you only need half the current in full winding mode, it
>>only takes twice the voltage to turn the current around in the same time.
>>But, at the same voltage, it takes 4 times longer to turn the current
>>around. That's why the half-winding scheme is often used on a 6-wire motor
>>with a bipolar drive.
>>
>>
>
>I really wasn't clear on the reasoning behind that until you posted
>this... :-)
>
><...>
>
>
>
>>It depends entirely on how the drive manages the collapse of the field in
>>the winding that is being turned off. Providing an effective scheme to
>>dump the inductive energy quickly can greatly improve the running
>>performance. But, that adds a lot of parts. The most common technique is to
>>place diodes to return the current to the power supply, but this basically
>>leaves the coil short-circuited. Putting some resistance in series with
>>those diodes will increase the voltage on the transistors, but will make the
>>field collapse faster.
>>
>>
>
>Not disagreeing with what you're saying here, but that sure is
>counterintuitive, that adding resistance in that context would make the
>field collapse faster...
>
>

Sure, if you short an ideal inductor, the current never decreases. This
is exactly how the
superconducting magnets in MRI scanners work, they have a coil of wire
with no
resistance, and after building up the current, they short it, and the
current flows as
long as it stays cryogenically cold. If you want to shut it off, you
put some resistance
in series, and the current decreases.

>
>
>>But, then the resistors will get hot. Especially in a chopper-type drive,
>>they will be carrying the circulating current nearly all the time while the
>>motor is holding position, and get REAL hot. So, you need more transistors
>>to shunt the resistors except when moving above some speed. Eventually, it
>>becomes obvious why a bipolar drive is the easiest way to get a simple, high
>>performance drive.
>>
>>
>
>Is the above stuff referring to a unipolar setup then?
>
>

Everything in the comment I made was referring to unipolar. The
topology difference
between unipolar and bipolar makes a great difference in how this is
handled.

>
>
>>And, as for microstepping, if it doesn't do microstepping, it isn't worth
>>having! (My opinion, of course! But, if you've worked with a Gecko G201
>>drive, not much else even comes close! And, a full-step, or even half-step
>>unipolar, L/R drive is so FAR from a Gecko that it is just a really bad
>>joke.)
>>
>>
>
>Maybe I'll get there, eventually. But for now I've gotta work with what I've
>got, which is a pile of TTL (and some CMOS, though not nearly as much as
>I'd like), some bipolar parts, a few darlingtons, and even a small number
>of FETs...
>
>So I'll get something going that'll at least let me move things, to start
>with. Which is okay, as I'm in no particular hurry here, having a pretty
>full plate as it is.
>
>
>

Unipolar L/R drives are abominable, unless the supply voltage is high,
then the power
losses are awful. Very simple unipolar chopper drives are a good deal
better, but
they still have this limitation on getting rid of the inductive energy.
The simple
bipolar drives get rid of the inductive energy in the most simple
manner, pumping
some of it back to the power supply when the motor is moving. Some is
always
lost as heat in the motor, of course.

Jon