3 phase versus 5 phase

[Pebbles]

The designs on this site are 3 phase but hugh's designs i have seen are 5 phase. Is the 5 phase more productive/efficient ? and secondly Would the extra phases have an adverse/beneficial effect on the loading / cogging ?

[Flux]

There is very little difference between any polyphase alternators, Single phase is horrible, 2 phase is inconvenient. 3 phase is normal but the 5 phase design does give some features that makes the connection simpler and Hugh tries to keep things very simple for some of the parts of the world that use his designs. It virtually dictates that you mount the rectifiers up on the alternator, this could be your deciding factor.

No air gap alternator without iron can cog, there is a lot of misunderstanding here that I have failed to dismiss after years of explaining it.

Performance between 3 and 5 phase is so close that other considerations will be your deciding factor.

I could put forward a good case for 6 phase but it never gets considered here but again the difference is slight.

Flux

[scoraigwind]

My 2005 plans used 5-phase for the 8 foot diameter turbine.  This entails putting the rectifier on the machine.  I have since decided that this is only really a good idea for 12-volt machines.  It works very well for 24 and 48 volt too but overall I prefer to put the rectifier on the ground in these cases.

My newer 'Recipe Book' designs are all 3-phase.  But the 12-volt versions mostly use an AC wire from each and every coil to the rectifier, rather like the older 5-phase design of 2005.  And the rectifier is on the tower.  I like that for 12-volts because it allows me to use thinner wire in the stator (but prevents parasitic currents between parallel coils) and I like using DC transmission (more efficient which matters with these very high currents) and the cooling is good up there for the high losses you will get in a 12-volt machine's rectifier.

5-phase was a bit of a wild idea that I did for fun and it's not really any better than 3-phase, but if you have a wire from every coil like this then you can have any number of phases.  Grateful Dead fans might for example like to try eleven phase output (depending on their age).  So long as the coil shape is suitable you can have as many phases as you like.  More phases give you a slightly higher peak voltage than 3-phase, and potentially a smoother output, but it's pretty complex to think through all of the implications.

[ghurd]

I could put forward a good case for 6 phase but it never gets considered here but again the difference is slight.

Flux

I would like to hear the basic case for 6 phase.
And the coil / magnet count.
G-

[Flux]

 6 phase is much the same as 3 phase, many motors are actually wound with the coils displaced electrically by 60 degrees, it is easier with 2 layer windings.

If you do it this way to get the 120 deg displacement for 3 phase you need to reverse the connections to the middle phase. By doing this you get the 120 deg separation.

To do it with a single layer winding you need to mirror 3 of the phases. Take the simplest case of an 8 pole axial with 6 coils, for 3 phase you have 2 coils in series per phase and 3 coils are connected to a star point. If you star all 6 coils you end up with 6 leads and a 6 phase winding. opposite coils of each phase will now  be 180 deg apart  so you have 3 phases 120 deg apart and the mirror coils giving another set 120 apart but displaced 180, You end up with 6 phases with an electrical displacement of 60 deg.

What happens when you rectify this with a 6 phase bridge rectifier ( 2, 3 phase bridges in parallel)is that instead of the peak conduction occurring at 1.73 x rms it comes out like the single phase case with the peaks directly opposite at 2 x rmsv.

You get a higher voltage at cut in than the 3 phase case but as your input voltage rises with increase speed the rectifier conduction changes and at full operating power the thing reverts back to normal 3 phase conduction pattern.

The current out with load for 3 phase into a battery is pretty much a straight line from cut in to full speed, there tends to be a bit of a toe near cut in but it is fairly small. With the 6 phase this toe is more pronounced and you get a load curve that starts at lower speed with conduction coming on more slowly and this is much more like the prop power curve, shallow near cut in and steep near full speed.

Hugh's 5 phase case shows this tendency also but instead of 2 x vrms it is nearer 1.9.

My load tests showed the actual efficiency in 6 phase to be marginally lower than 3 phase, but efficiency doesn't count for much compared with load matching.

If instead of linking all 6 leads in star, you star the 3 coils in each section you have a parallel star winding. You can convert from 3 phase parallel star to 6 phase simply by joining the 2 star points so you would have the option of trying the two methods as a comparison.

You can do this trick with any 3 phase single layer winding with coils diametrically opposite. It doesn't work for 12 pole 9 coil but does work for 16 pole 12 coil etc.

[ghurd]

Thanks Flux.

I'll have to think about all that.
Especially "input voltage rises with increase speed the rectifier conduction changes and at full operating power the thing reverts back to normal 3 phase conduction pattern."

Pretty sure I have done or seen it with motor conversions. (2 x rmsv)
G-

[Pebbles]

Thanks for the replies guys. appreciate the clarification
Pebbles

[wiredup]

hey flux is the 6 phase you are talking about the same as what some call y2 connection where you connect 456 101112 in y and connect 1 and 7,,2and 8,,3and9 as primaries?? your friend and a reader on this site since 03....... thanks

[Flux]

No the various motor connections are all 3 phase to suit the normal 3 phase distribution system. The various connections let one motor suit various voltages.

The various combinations are all series and parallel forms of star or delta.

There is no commonly available 6 phase supply so you will not find it in normal use. Six and 12 phase were common with mercury arc rectifiers where bridge connections were difficult. They were used as it improved the smoothness of the dc and had other advantages. Six and 12 phase can be constructed with transformers from a normal 3 phase supply by mirroring the secondary windings as I described for the alternator and you get the extra shift for 12 phase by using the 30 deg relationship between star and delta. In addition zigzag windings were often used on the transformers to minimise the effects of the dc component in the secondary legs.

For normal wind power you would use a 6 phase bridge rectifier which is easy with silicon diodes. With mercury arc the arc pool was common to all the anodes and anything other than half wave rectifiers was difficult.

This has gone way off topic but you may pick up more information on 6 phase if you look at early rectifier stuff.

Flux

[wiredup]

thanks flux, i guess i need to read alot more.

[joestue]

Flux I think what you got going on with the 6 phase stuff is you got the 2x rms of the coil.
But as the current increases and thus the voltage drop across the coils increase, then the other phases start taking their share of the current.
At any given point you got two coils out of the 6 that are pushing more than their "fair share" of the current, and this explains the drop in efficiency.
the rms voltage drop across 2 out of the 6 at any given point in time would have to drop a minimum of 15% before the other four take up any current.
sounds rather inefficient to me...

[DamonHD]

Hmm, flux is almost always right to the point that if he told me that black was white I'd have to avoid pedestrian crossings until I'd thought about it a while.   B^>

Rgds

Damon

[ghurd]

Hmm, flux is almost always right to the point that if he told me that black was white I'd have to avoid pedestrian crossings until I'd thought about it a while.   B^>

Rgds

Damon

Every time I thought he was wrong, it was because I did not understand what he meant!
G-

[Flux]

Yes I am sure Joestue is right, the conduction starts as the single phase, with the thing running as 3 biphase rectifiers with dc peak being that of the single phase case.

When running at full load the figure is pretty much as for the 3 phase bridge. There must be points where more than one phase is conducting just as for a normal 3 phase rectifier with overlap.

The measured drop in efficiency was very slight and in all cases I measured the 6 phase was better than 3 phase delta. With these machines the internal losses are way higher than in the case of 3 phase machines used for normal loading.

I am sure that the better matching to the prop will be way more than the slight loss measured on a bench test. Even the measured difference between star and delta is not normally seen under wind conditions as we just don't have the consistency of wind to measure things this closely. The delta case does show at cut in whereas the 6 phase is working well at these power levels where you want the best matching and efficiency.

In the industrial case of 6 or 12 phase rectification the load is always inductive and the rectifiers always work with overlap. Battery charging often still results in some overlap with alternators with significant reactance but air gap alternators don't show it in the 3 phase case. I have not looked in detail at 6 phase but it must exist.

I suspect something similar must happen with IRP ( Jerry connection). At normal power levels it behaves much as delta, with the circulating loss below cut in removed. Jerry actually claimed things that I have not been able to reproduce but I think that if you overload things badly enough more coils share the current and his claim of less heat than star could well be true in this very overloaded case.

With the exception of car alternators, no commercial machines operate with rectifiers in this way so common industrial practice doesn't tell us everything. Although industrial alternators have exciters feeding rectifiers the load is always a field magnet with huge inductance.

More room for experiment I am sure.

Flux