More Phase Questions

[Shawncfer]

The neutral needs to be oversized because dimmers are an inductive load. What that means is the the dimmers can 'Shift' the time of the returning electrons a bit. ( This is one of the things the choke does in a dimmer - also the filament in the lamp ). So it is possible in a system with inductive load to have more electrons going through the neutral than a single phase. We handle this by oversizing - IE we put in a wire which is physically bigger and can handle more electrons without heating up for the neutral wire.

I Understood you perfectly up until that...

 

[derekleffew]

Just roll with it, Shawncfer. And take comfort in the fact that you are not alone.

Here is some real Neutral History in the attached file. I did this around 1981. We were having weekly Code revision meetings at See Factor (boy, were we young and full of piss-and-vinegar!). Mitch Hefter said to me: "SCR dimmers can generate neutral currents in excess of the maximum phase current." "Impossible!" said I, and then went back to my shop to prove it. Oops! I proved Mitch's statement, and the gory details are attached.

This was the start of the work we did in the 1987 Code to require upsized neutrals on phase-control dimmers. I don't think many people in the electrical community were even aware of "harmonics" or "non-linear loads" at that point. ...

 

[STEVETERRY]

Pretty darn close! As the waveform crosses the 0 volt point, the ZVD produces a logic pulse and the CEM starts counting. It is the CEM that calculates how to translate various dimmer curves. The CEM outputs a firing signal which goes through an opto-isolator to turn on the SSR, so yes, there is not much going on in the power module outside of filtering and protection.

As Steve said, there are three ZVD circuits, one for each phase leg. The CEM also takes care of mapping which ZVD pulse should be used on which virtual dimmer counter. Everything between the ZVD pulse and the firing pulse is more or less software programed into the CEM.

In the case of ETC, line voltage is also inputed as data to the CEM so that pulse timing can be altered to compensate for fluctuation in the line voltage and keep the output at the same level.

JD is doing fine with this, and I'm in England at the ITEAC conference, so I'll let him keep going!

ST

 

[Shawncfer]

Okay so I reread it a couple times and was able to grasp what you were saying and I get it now.

However,

There are 120 volts between phase A and neutral, and 120 volts between phase B and neutral. (And 240 volts between Phase A and Phase B - but we don't care about that now)

So even though there are three legs, they're all connected to one nuetral, which is the same as them all being connected to their own nuetral. They just share one. Therefore each 1/3 of the rack is 120v because phase to nuetral is 120v. Right?

 

[JChenault]

Okay so I reread it a couple times and was able to grasp what you were saying and I get it now.

However,


So even though there are three legs, they're all connected to one nuetral, which is the same as them all being connected to their own nuetral. They just share one. Therefore each 1/3 of the rack is 120v because phase to nuetral is 120v. Right?

(This post may be confusing as I could not get my original quote in there. )

The section you quoted is talking about single phase three wire. In this type of feed there are only two hot legs, not three. But your understanding of a common neutral is correct.

 

[JChenault]

So there seems to be some confusion as to the 'why' of the neutral needing to be oversized. Here is my understanding.

It's all Physics, magnetism, and the relationship between electricity and magnetism.

If you have current ( electrons) moving through a wire and the current is changing direction, the current generates a magnetic field. Similarly if you have a piece of wire moving through a magnetic field, it generates a current flow. ( This is how generators work).

In an SCR dimmer, the wave form of the output is pretty abrupt when you turn the dimmer on. The sharp start of current can make lamps sing and vibrate. ( A filament is a coiled piece of wire. When the power is applied, it generates a magnetic field which makes the filament vibrate ).

The way we get around this is with a choke. The ETC website talks about it a bit here. (Note that I believe the page is incorrect as to the cause of the sing, my understanding is that it is not heat, but magnetism.

The choke is a ferrite (Iron) core with a wire running around it. The output from the SCR passes around and around and ... the ferrite core. When the power turns on from the dimmer, the electrons generate a magnetic field in the choke. This field slows up the movement of the electrons and smooths out the waveform. This delays the electrons moving in the circuit, and this delay is greatest when the dimmer is at about 50%.

Remember that the reason that the neutral ( in a perfect, evenly loaded world) would not be carrying any electrons is that the positive and negative currents canceled out. If you have one phase at full ( minimal delay from the choke) and the other phase at 50% ( maximal delay at choke ) you can have more electrons passing through the neutral than through any phase.

 

[JD]

Regarding the Neutral;
When your dimmers are at 100% or at 0%, the neutral is caring very little current. However---- Think about how dimmers work. They work by chopping the waveform going to the load. The SSR in the dimmer is either in conduction or open circuit. Nothing between! At 100%, it is in conduction through almost the entire waveform. At 50%, it turns on halfway through the waveform. Now, think carefully about what is happening if all your dimmers are set to 33% on a three phase source. (we will assume it is a liner curve.)

While the SSRs on phase leg A are on, the SSRs on B and C are open circuit.... No Balancing is occurring!

While the dimmers on leg B are on, A and C are open circuit.

Likewise, while the dimmers on leg C are on, A and B are open circuit.

In other words, because there is no offset balance, the Neutral is conducting 300% more current then any of the phase legs. It shouldn't be a problem as all the dimmers are only at 33%, right? Unfortunately, tungsten lamps do not have a liner relationship between voltage and current draw. Because of this, the current on the Neutral actually exceeds the rated capacity. Between 30% and 50% is a real bad area, even though there is some overlap on phases past 33%. The current ripple of each phase on the neutral is occurring at a different time for each phase leg. Because of this, the ripple current is three times the frequency of the line (in America, that's 180 cps) thus the term "Triplen Harmonics." (google- Triplen Harmonics three phase wye)

To make life real interesting, some transformers now come with Triplen Filters. Good idea? Nope! Long story short, dimmers need a solid neutral so the filter has the same effect as a loss of neutral! You won't have anything blow up, but the zero voltage cross detectors will be thrown for a loop causing all sorts of crazy problems!

EDIT: One last little sting- Since the peak power draw is occurring during the last 1/3 of the waveform, the "Power Factor" is considered to be very low/bad (dependent on the actual dimmer setting.) So, when power companies start billing based on Power Factor Compliance (see "cap & trade"), theaters may be especially hard hit on the billing end.