ECN Forum Forums General Topics of the Trade General Discussion Area Calculating neutral loads.

One thing that you must keep in mind is that many 240 volt appliances also require 120 volts. If the neutral weren't there, you would have to derive one anyway. It is common to drive heating elemebts with 240 and timers, lights, fans, mini-brains, Et,c, with 120. So the return current will often be in the mA range while the 240 volt load will be in the 10's of amps.
Joe
Of course, in reality, it wouldn't be a derived neutral. It would be a derived equal voltage point within the load appliance. This would add alot of complexity and could only be useful as a college senior project or to impress the local MENSA babe.

[This message has been edited by JoeTestingEngr (edited 07-30-2006).]

you throw harmonics into the picture, and ALL rational computations just fly out the window! With cheap switched power supplies becoming more and more common, I don't think it will be too many more revisions of NEC before we start having to put OCP on shared neutrals, too. At least for certain types of loads.

[This message has been edited by SteveFehr (edited 07-30-2006).]

If it's a Y system, all loads touching the neutral, the neutral current, if the harmonics are so bad that the currents never overlap will just add up as (phase A current)+(phase B current)+(phase C current)= current on the neutral. Makes you ask "why bother with neutrals anymore?" Or 3 phase power.... You don't get this problem on a single phase 240V centertapped system, current spikes on both sides happen at the same time.

wa2ise, that's true for half-rectified power supplies, but full rectified power supplies will also have reverse current that largely cancells out the forward current in the neutral, leaving just the harmonics as the difference between the two.

Steve, I believe the half wave rectifiers will give you a current draw only during the positive or negative half of the alternating current. The full wave rectifiers will give a current draw during both halfs of the alternating current.

LarryC

edit trying to figure out this @#$%@#$!!! so I can post an ASCII drawing.


[This message has been edited by LarryC (edited 07-31-2006).]

Actually half wave really refers to the half of the cycle where current is flowing in a given direction. That is from peak to peak, not when it is above or below zero. You either get all the "uphill" or all the "downhill". Full wave is like your dad's walk to school. The arrangement of diodes in the bridge makes it uphill both ways.

Half-wave rectified usually means 1 diode per phase, and only works when the voltage is positive. EG, current travelling from the hot wire to the neutral. Full wave uses a 2nd diode to rectify power when the voltage is negative- EG, current travelling from the neutral to the hot. In ideal balanced 3 phase power, the 3 phases will always balance perfectly. In a 6-pulse (simple 3-phase/6-diode) power supply, even on a perfect ideal system, the switching drops a good chunk of the sine wave, making the neutral current cancellation imperfect, setting up semi-sinusoudal current at a higher frequency on the neutral in, IIRC, the 3rd and 5th harmonics.

Edit: er, I might be thinking of delta for 3rd and 5th harmonics. I'd have to hit the books back at the office to check- either way, the exact frequency of the harmonics isn't really important to this discussion!

[This message has been edited by SteveFehr (edited 07-31-2006).]

True, but this doesn't help the neutral any. Assume for the moment that switching power supplies in computers in a server farm draw large spikes of current that last 10% of the 60 Hz cycle time. These supplies use a bridge rectifier or voltage doubler circuit, both would draw current spikes at the positive peak and the negative peak of the phase that they are connected to. Think in terms of time: computers loading phase A will draw a spike of current at the top (positive) of phase A, and return it to the neutral. Call the start of this spike time zero. a full cycle of 60Hz lasts 16.7 msec. The spike lasts for 5% of that time, 0.83msec. Computers on phase C will see the negative peak of that phase 2.77msec later, and their current spike will last 0.83msec. ANd so on thru the rest of the positive and negative peaks, total of 6 over one 60Hz cycle. None of the spikes returned on the neutral overlap in terms of time. Thus the neutral sees 6 current spikes, and any one hot phase wire only sees 2 spikes in one 60Hz cycle time. Thus: neutral current = phase A current + phase B current + phase C current in the worst case. If this were a system feeding pure resistors, each load would be drawing overlapping currents, and the neutral would see no current at the panel (assume all 3 phases equally loaded here).

Oh, but it does help! Peaks in the real world are not square waves, and the phase angles of 3-phase power does help us. Here is an o-scope metering of 3 typical PC supplies, from the high end to the cheapest.


On the cheapest of the cheap (which we don't see in places that are going to be using 3-phase power), you'll see a peaked sinusoudal wave at about 180Hz on the neutral with short pauses of 0 current between. As PFC is added to the PSUs, the spikes widen, and current flowing in opposide direction from the different phases begin to cancel. A large harmonic is left, but it's not as bad as our super-cheap PSU. On the best PSUs, the current on the neutral cancels almost completely. In all cases you'll see harmonics to some extent, though.

[This message has been edited by SteveFehr (edited 08-01-2006).]