Re: I'm confused, which are better uni or bi polar ?

At 06:14 12/02/2002 +0000, you wrote:

>All else being equal, a bipolar beats a unipolar hands-down! The only
>"disadvantages" are you require a DUAL-polarity supply (i.e., instead of just
>"plus and minus", you need "plus", "Common", and "minus". Think of two
>separate supplies with the+ of one connected to the (-) of the other, and
>that tie-point is "common". THREE terminals. Clearly nearly twice as
>expensive as a single-polarity supply of same current.

That isn't true. You can drive bipolar perfectly well from a single rail
supply - take a look at the L297/L298 chipset for example. I am currently
working on a project that uses these to create a bipolar drive, and am
powering them from a simple single rail supply - no centre tap, no plus,
common and minus, simply plus and common. The bipolar drive in the L298
achieves the reversal of current through the winding by controlling which
end of the winding is at supply rail +ve and which end is effectively
grounded. Both ends grounded (or both +ve) and there is no current through
the winding. One end grounded, the other +ve, and current flows one way;
swap ends & the current flows the other.

> Second, the DRIVE is
>much more complex, as you need TWO "H-drive" circuits, one for each of two
>windings in the motor.

I would grant you that a simple unipolar drive built from discretes &
transistors uses fewer transistors than a simple bipolar drive, but if you
use one of the (many) proprietary driver chipsets around, the bipolar
circuitry is NOT any more complex. An L297, an L298 and a small bunch of
discretes and you have a 50V, 2amps/phase bipolar drive for one motor, with
a step & direction interface included. Use a couple of L298's and you can
up the current to 3 amps/phase. Some of the more modern designs make it
even easier than that - the recently released Allegro 3977 chip, for
example, appears to do what an L297 plus L298 does, and adds microstepping
as well.

Arguably, a unipolar drive is now more expensive to produce, because it
(usually) requires the use of a very high power series resistor to limit
the current, which the bipolar drives generally don't need. The extra
power dissipation in these power resistors means you need a beefier power
supply for a "simple L/R unipolar drive than for the equivalent bipolar
chopper.

So, as the saying goes, generalizations are dangerous (including this one!)

Regards,
Tony