| On 30 Dec 2003 09:21:11 -0800, the renowned
[email protected] (Alan
| Horowitz) wrote:
|
|>what is the attraction of three-phase power? Why not 9 0r 317
|>phases? Why not plain ole hot & neutral?
|
| It better uses transformer iron and makes better/cheaper induction
| motors. For making DC it requires less in the way of filtering. It
| would be nice if we had something like 6 phase power, we could get rid
| of the input filter caps in most equipment, the only part that we
| can't make smaller by increasing efficiency and frequency of SM power
| supplies.
Actually, you can get 6 and even 12 phases out of 3 phases fairly easily.
If you put a normal single primary to center tapped secondary (120/240 in
North America) on each of the three phases (tie all the neutrals together),
you really have 6 phases.
Y3 * * X2 X0 is in the center, and legs are X1, Y3, X2, Y1, X3, Y2
\ / X[123] to X0 gets 120 volts
\ / Y[123] to X0 gets 120 volts
X1 *----*----* Y1 Xn to Yn gets 240 volts
/ \ X1 to X2, or X2 to X3, or X3 to X1 gets 208 volts
/ \ Y1 to Y2, or Y2 to Y3, or Y3 to Y1 gets 208 volts
Y2 * * X3
I've never seen a 3 phase transformer with this configuration, probably
because there is virtually no need for it. If you really do want it,
you'd have to build it from 3 separate 1 phase transformers.
If you were building a large computer machine room with lots of switching
power supplies, and your power feed is three phase, this might actually be
useful to minimize the 180 Hz component on the neutral. If you used a
standard 208Y/120 feed and ran computers on everything, you could see as
much as 2-3 (depending on how wide your current pulse is) times the current
on the neutral because of the timing offsets of the switching power supply
currents. 1 phase power doesn't have this problem as long as the computers
have similar current waveforms (they cancel out on each side). With the
above 6 phase configuration, you'd get the advantage of 1 phase balancing
while keeping the load balanced on your 3 phase feed.
But the payoff might not be there since the above requires 7 wires, whereas
you can compensate for high neutral currents with larger neutral wires and
stick with 4 wires (one fat, or doubled or tripled). Also, you can get
UPSes that use 3 phase power, and that should take care of the problem.
You'd want to use UPSes, right?
Motors and transformer power supplies can work on 3 phase power as if they
were 6 or 12 phase. Normal power feeds have a wye secondary, so you get
either 208Y/120 or 480Y/277 or 600Y/346 depending on your service category
(the latter 2 use an additional transformer to get the 120v level). These
wires are labeled X0 for neutral, X1, X2, and X3 for the three hot legs.
Lets say you have a very high power motor that uses 480Y/277 (typical for
high power industrial motors) and does this with 12 windings. 6 of them
would be designed for 480 volts, and the other 6 for 277 volts. Then the
connections would be, in order:
12 o'clock X0-X1 277 volts, 0 degrees
1 o'clock X3-X1 480 volts, 30 degrees
2 o'clock X3-X0 277 volts, 60 degrees
3 o'clock X3-X2 480 volts, 90 degrees
4 o'clock X0-X2 277 volts, 120 degrees
5 o'clock X1-X2 480 volts, 150 degrees
6 o'clock X1-X0 277 volts, 180 degrees
7 o'clock X1-X3 480 volts, 210 degrees
8 o'clock X0-X3 277 volts, 240 degrees
9 o'clock X2-X3 480 volts, 270 degrees
10 o'clock X2-X0 277 volts, 300 degrees
11 o'clock X2-X1 480 volts, 330 degrees
An AC/DC power supply could use 6 transformers wired for each of the 6 phases
(but not the opposites), and feed full wave bridges to get nearly DC with a
very low 720 Hz ripple (the secondaries would have the 12 phases above). For
very large power supplies this can become practical since you don't need as
much of a capacitor bank to smooth it out (or you can get a much smoother DC
if that's your goal). 3 of the transformers would have one voltage for the
primary and the other 3 would have a different voltage (208 and 120 for the
most common three phase power in the US).
If you didn't want to go to that extreme, you could still do a 3 transformer
version that gets you a 360 Hz ripple. These would all have the same primary
voltage, and could be attached either between legs and neutral, or between
each hot leg, depending on voltage, or even work on a delta feed.
With more transformers strategically wired, you can get 24 phase, 96 phase,
and so on. At some point you might not need any filter caps