Double delta amps per winding questions. dogleg windings are half amperage?

[Demoh]

Ok, maybe I am a little rusty on electrical theory, but I am trying to wrap my head around what I am witnessing on this generator I am working on.

I was working on reconfiguring a 12 wire 10kw head and on initial startup hoping there was no smoke my amp readings for the windings aren't balanced so I am a bit confused.

10kw 1.0pf resistive load, 240v, 42 amps. The gen head is configured double delta. The gen head is rated at 10kw .8 PF and 52 amps. (MEP-803a for the curious, the work was to remove the commonly failing S8 switch.)


My connections are identical to above. 5 6 7 9 are on N, 8 and 10 are bonded, 2 and 4 are bonded. 1 and 12 is L1 and 3 and 11 are L3.
Amp readings at various points (approximate +/- 1 amp):
L1 - 41 amps
L3 - 41 amps
N - close to nothing
pigtails: I didnt write any of it down, all from memory, I think the 9-12 winding was an amp or 2 more than the 3-6 winding, enough to cause some more heat on the thermal imager but not enough for me to think something was wrong.
3 - 26 amps
12 - 26 amps
11 - 16 amps
1 - 16 amps
8 - 16 amps
2 - 16 amps

So is this normal? I get the basic theory that if you add 2 sinewaves that are 120 out they will match the third so essentially you have 2 sets in parallel. But is this the reason why the load isnt equally split between the pairs of windings? If I had a .8 load at 10kw would the 3-6 and 9-12 windings stay at 26 amps with the other windings increasing to 26 amps or would all of the windings increase proportionately?

I understand (correct me if I am wrong) that normally we de-rate by 1/3 due to circulating currents in this type of configuration. (ill do some more thorough testing today.) But Ive also read contradictory information stating that the amp draw across the single winding would be equal to the amp draw across the dogleg it is paralleled with.
Nameplate on these generators is so conservative that they routinely provide power at 125% for 8+ hours and my readings at 100% is still under the amps limit on the nameplate (52amps, so 26 per winding set) so I am not worried about harming the generator (its standby so added heat contributing to insulation degradation is nil, plus Onan marketed the same engine/genhead combo commercially as a 16kw which I also have one of those. I am just trying to A wrap my head around this and B if theres a problem with this head to know about it. (every set of windings metered at .3 ohms. I didnt meg them.)

 

[synchro]

If the windings and their associated magnetics were operating completely independently, then because they are in series the source impedance of the two windings forming an open delta would be twice the source impedance of the winding that's directly connected to the load (let's call that the main winding). If this were the case, then the current through the open delta would be half of that from the main winding, giving about a 33% / 67% split of the total load current. In other words, the 42 amps would be supplied by 14A from the open delta and 28A from the main winding, which is even more unbalanced than what you are observing.

From page 5 of the document from Cummins at the link below:

"But the Double Delta configuration does load all three phases and has been proven by all
alternator manufacturers to be the best compromise way of getting back to near balanced
conditions inside a three phase alternator when a single phase load is connected. In the
Double Delta connection, the phase windings in the greater part of the delta circuit handle
about 40% of the total load current, with the 'master' phase in I part of the delta handling some
60% of the total single-phase output current."

https://www.stamford-avk.com/sites/stamfordavk/files/AGN017_E.pdf

Such a 40/60 split would be about 17A and 25A, which is very close to your measurements. I suspect that the currents are closer than the two to one ratio mentioned above because the impedances of the windings are not completely independent. In other words, they do not just have a "self impedance" but they also have a "mutual impedance" because some of their magnetic "circuit" is in common between them. This means that some of the magnetic flux flows between the different phase windings on the stator, which could equalize the currents somewhat.

From the following text on page 5 of the document, it seems like having a 50/50 current split would not be desirable anyway because apparently there are more considerations than just the magnitude of the current when there's a very low power factor on the open delta windings ( in this case PF = cos(60°) = 0.5 ).

"But if the three phase windings are connected into a connection such as Double Delta, then at
least all three phases are carrying current and so contributing towards the single-phase output
required. However, the current through the phases that form the greater part of the delta circuit
have load originating current 'pulled' through them out-of-phase with the voltage that is being
generated in them. This leads to more complex considerations of what’s happening within the
stator and rotor components of the generator and more J notation calculations are required to
calculate the resultant effect on individual component rating and capability."

 

[Demoh]

Wow, ok. Clear, concise, answers my questions exactly. It will probably take me more than 2 reads through that document but definitely enlightening.

Damper cages, new info for me. A thought I had prior to damper cages, (maybe I need to read a 3rd time, I didnt quite grasp if the damper cage fully negates my next question/statement.) For single phase loads under double delta, What can be done to reduce the torsional vibration sent through the crank? Sounds to me like its a good way to fatigue it to failure. When circulating currents is talked about, is that because of the cage or is that unrelated and somewhere else in the alternator?

Even though Im dealing with 4 cylinder engines (I wish these were 6 cylinder) my OCD would purposely want to overfuel the cylinders that would essentially be doing the 60%. This explains why all of my single phase generators run like dog poo compared to balanced 3 phase on the same units. <-- ignore this, just fun to think about.

This opens up another question of mine. on 12 wire heads (not dedicated single phase,) are there manufacturers which have certain windings / rotor / cage geometry which help mitigate this issue? For example for the diagram above, if this alternator was to be configured double delta that they might specify that windings 3-6 and 9-12 be used for the main windings? Does doing that defeat the 12 wire purpose which causes it to act poorly when connected parallel wye? Or is this completely moot because the rotor cage takes care of that?

Now since my brain hurts, pulling a few spec sheets. Marathon has single phase derate of "approximately 50%" for both zig-zag and double delta, Power-Tronics says to derate by 1/3 with double delta and not to use zig-zag. If the nameplate on all of this generator lists "12 wire, 120v, 34.7a, 12.5kva, 10kw, .8pf" on the nameplate. Im wondering if all of the math was done so that it is pre-derated (or rated at .8pf) since the generating unit's nameplate lists 120 1p @104a, 120/240 1p @52a, 120/208 3p @34a.

(sorry, my brain is a sponge and if I was to go to school to learn all of this I wouldnt have a way to pay for it because I would be in school until death. Also I find it fascinating that searching for this answer came up fruitless and this info you have provided is more of a result of asking the question properly. And all I can say is single phase is so infuriating. 3 should be everywhere IMO. Of course I get the shaft because it seems every client has HLD or HLOD and they want a generator.... push them for a service change to wye or multiple generators, but thats above my pay grade because I just work on these things.)