quiz: neutral current = sqrt(3)*phase currrent
quiz: neutral current = sqrt(3)*phase currrent
(OP)
Let's say you have an ac MG set. The output of the generator (3 phases plus neutral) supplies an electronics cabinet. That electronics cabinet in turns supplies regulated dc current to an inductive load.
The true rms current on the neutral of the MG is sqrt(3) times as high as the true rms current on the phases.
Would you have any guesses whether this is realistic and what might cause this?
Assume for quiz purposes that only true rms (not waveform) is available but we know the measurement represents true rms for any type of waveform: high-freq, low-freq, dc etc.
The true rms current on the neutral of the MG is sqrt(3) times as high as the true rms current on the phases.
Would you have any guesses whether this is realistic and what might cause this?
Assume for quiz purposes that only true rms (not waveform) is available but we know the measurement represents true rms for any type of waveform: high-freq, low-freq, dc etc.
=====================================
(2B)+(2B)' ?





RE: quiz: neutral current = sqrt(3)*phase currrent
However if you lost (opened) one of the three phases, the resultant vector sum of the other two phases would have magnitude of sqrt(3) times the rms phase current.
RE: quiz: neutral current = sqrt(3)*phase currrent
=====================================
(2B)+(2B)' ?
RE: quiz: neutral current = sqrt(3)*phase currrent
To simplify, assume only third harmonic and make the RMS value of the harmonics in each phase equal to sqrt(3) times phase current. Those currents will add up to 3*sqrt(3) times phase current.
If it is a plausible situation? Yes, very, I would say.
Gunnar Englund
www.gke.org
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...
RE: quiz: neutral current = sqrt(3)*phase currrent
Another piece of data: the triplen harmonic content of the neutral current is small, on the order of 20% or less of the neutral true rms value.
Do you think it's possible ?
=====================================
(2B)+(2B)' ?
RE: quiz: neutral current = sqrt(3)*phase currrent
Gunnar Englund
www.gke.org
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...
RE: quiz: neutral current = sqrt(3)*phase currrent
=====================================
(2B)+(2B)' ?
RE: quiz: neutral current = sqrt(3)*phase currrent
1 - The output of a 500kva, 60hz, 3-phase motor-driven generator supplies an electronics cabinet, which supplies regulated dc current to an inductive load.
2 - The true rms magnitude of the neutral current is sqrt3 times the true rms of the phase current
3 - The true rms magnitude of all 3 phase currents is the same.
4 - The triplen harmonic content of the neutral current (180hz, 540hz, 900hz etc, combined) is <20% of the true rms of the neutral current.
Now a new piece of data:
5 - The 60hz component of the phase current is less than 5% of the true rms of the phase current.
Items 2 thru 5 all apply to measurements between the generator and electronics cabinet.
Again - just a quiz. If it was real troubleshooting you would be able to see the time waveform / spectrum and there would be no mystery. But that's no fun. Any more thoughts?
=====================================
(2B)+(2B)' ?
RE: quiz: neutral current = sqrt(3)*phase currrent
As you say, it would be a lot easier if you actually saw the waveform.
Gunnar Englund
www.gke.org
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...
RE: quiz: neutral current = sqrt(3)*phase currrent
We have simply a half-wave SCR controlled bridge rectifier. Each phase conducts for 120 degrees and is off for 240 degrees. The output has continuous dc current with some ripple... it flows to the inductive load and back through the neutral. The firing angle is adjusted to slowly change the level of output current (on the scale of seconds).
For simple case the output is 1 continuously and has rms of 1.
The input is 1 for 1/3 of the time and 0 for 2/3 of the time. Applying the definition of rms to the input shows it is 1/sqrt3.
The fact that we have ripple riding on top and slowly varying envelope (time scale of seconds) doesn't change this relationship.
Sorry if it was an uninteresting "quiz". A contractor submitted a proposed modification to increase the output current under the incorrect assumption that the input rms current would be 1/3 of the output rms. I have had the pleasure of trying to explain this "simple" (?) fact to 6 or 7 people in 2 companies over the last few days and it was a little exasperating. The fact that sqrt(3) is well-known in 3-phase power for other reasons seemed to muddy the waters. (more than one of them thought I must have come up with square root of 3 based on something to do with line/line vs line/neutral voltage).
=====================================
(2B)+(2B)' ?
RE: quiz: neutral current = sqrt(3)*phase currrent
For correctness, there's a lot of triplenes in that current. Not exclusively, but lots.
I don't think that this was uninteresting. We need eye-openers now and then. You quiz a few years ago (about where torque is produced in a motor - conductor or iron) was another of those eye-openers.
Gunnar Englund
www.gke.org
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...
RE: quiz: neutral current = sqrt(3)*phase currrent
If you piece together sinusoidal waveforms from the phase inputs you would picture the 180hz voltage ripple of the dc output which could approach 50% of the peak value. But the 180hz current ripple on the dc output is much smaller because it is filtered by the inductive load. I said < 20% on the output current, but the magnitude of the 180hz current ripple is actually less than 5% of the peak value.
I said the phase current fundamental was < 5% of the total - but I was probably wrong on that point. The phase current is pulses of current that last for 120 degrees and disappear for 240 degrees. So it is would certainly be rich in harmonics of 60hz... I think it must include a lot of the even harmonics, but probably also has a fundamental component much higher than 5%.
Now the more practical question – what kind of derating factor would be applied to the generator carrying that kind of load (only one phase conducts at a time and return current through the neutral).
=====================================
(2B)+(2B)' ?
RE: quiz: neutral current = sqrt(3)*phase currrent
=====================================
(2B)+(2B)' ?