Harmonics on branch circuits

[Dennis Alwon]

We recently had a thread where we mentioned that a neutral must be sized at 173% when harmonics are present. I understand that this is based on 1.73 that we use for 3 phase but I am curious what are the factors that go into determining what the load on that neutral is going to be.

For instance, we have a 3 phase load and all the circuits of the MWBC are loaded to 15 amps. We use the formula to calculate neutral current. In excel I would use this formula =SQRT(((A5*A5)+(B5*B5)+(C5*C5))-((A5*B5)+(B5*C5)+(A5*C5))). Now do we use 173% for the neutral on the branch circuits as well. I have never seen this done before and normal if the circuit is balanced the load on the neutral is 0. With a 16 amp load on all 3 phases the worst scenario is 16 amps on the neutral if two circuits are turned off. This would mean 27+ amps on the neutral. Is this correct or is this dependent on the ballast pf or some other factors.

Certainly with the 2008 code we have art. 210.4 that virtually will require all circuits to be deenergized at once. So is this really an issue for branch circuits? Have any of you installed #10 neutral on these circuits with #12 ungrounded conductors-- talking lighting not receptacle loads.

 

[ron]

Many modular furniture manufacturers make "super neutrals", which is an oversized neutral in the wiring harness. Desktop PC's are high in harmonics. Data center power supplies are often power factor corrected, which reduces harmonics. I don't oversize neutrals in data centers anymore unless the client requires it. Office furniture, yes.

 

[Dennis Alwon]

Many modular furniture manufacturers make "super neutrals", which is an oversized neutral in the wiring harness. Desktop PC's are high in harmonics. Data center power supplies are often power factor corrected, which reduces harmonics. I don't oversize neutrals in data centers anymore unless the client requires it. Office furniture, yes.

I was aware of the super neutrals in office furniture but my question was on the lights. BTW-- what size neutral do they use in the modular furniture-- #10?

 

[steve66]

Lighting ballasts normally list a maximum 10% THD, or 20% THD.

Maybe I'm oversimplifying this, but taking the 20% THD, and your example, it seems to me there would be 16 amps *0.2 THD from each circuit. So that's an extra 3.2 amps from harmonics as an absolute maximum possible.

Considering both phases, that would be 16+3.2+3.2 = 22.4 amps as the maximum current on the neutral.

If the phases were all balanced with 16 amps each (3 phase system), then there would be a maximum of 9.6 amps of harmonic current on the neutral.

I thought all newer computers (everything made in the last several years) complied with European standards, and had filtering to reduce the harmonics they cause. But I'm not sure how much to trust the source I got that from.

Steve

 

[ron]

Lighting ballasts normally list a maximum 10% THD, or 20% THD.

Maybe I'm oversimplifying this, but taking the 20% THD, and your example, it seems to me there would be 16 amps *0.2 THD from each circuit. So that's an extra 3.2 amps from harmonics as an absolute maximum possible.

Considering both phases, that would be 16+3.2+3.2 = 22.4 amps as the maximum current on the neutral.

If the phases were all balanced with 16 amps each (3 phase system), then there would be a maximum of 9.6 amps of harmonic current on the neutral.

I thought all newer computers (everything made in the last several years) complied with European standards, and had filtering to reduce the harmonics they cause. But I'm not sure how much to trust the source I got that from.

Steve

Steve, I heard the same thing, but I think it is mostly true of larger power supplies for servers. The desktop PC's are under the limit for the European requirement <75W. Unfortunately, also second hand information. http://www.google.com/url?sa=t&rct=...triUBQ&usg=AFQjCNGlLF_dmPWa7pK_6HwBwyP-agP6yA

 

[Dennis Alwon]

THD? Not sure what it is but you are answering what I wanted to know

 

[steve66]

Steve, I heard the same thing, but I think it is mostly true of larger power supplies for servers. The desktop PC's are under the limit for the European requirement <75W. Unfortunately, also second hand information. http://www.google.com/url?sa=t&rct=...triUBQ&usg=AFQjCNGlLF_dmPWa7pK_6HwBwyP-agP6yA

I'd trust Cooper for the most part.

But I didn't realize the European Standard was for power factor, not harmonics. I found this note:

Although pleasantly simple and robust, a passive PFC rarely achieves​

low Total Harmonic Distortion (THD).

So a computer could easily meet the power factor requirements without necessarily having low harmonics.

 

[Dennis Alwon]

Any help on THD-- what does it stands for?

 

[steve66]

THD? Not sure what it is but you are answering what I wanted to know

THD = total harmonic distortion. Mostly just another way to say harmonics. If harmonics make up 20% of the signal, its 20% THD.

I was thinking about this a little more. To need have harmonics on the neutral reach the 173% number, it seems like the THD would have to reach about 35%. (A 16A load, 16*.35 = 5.6 amps of harmonic current. If we look at your example again, thats' about 27.2 amps on the unbalanced system. And 27.2 amps divided by the standard 16 amps load = about 1.7.)

So it seems like the 1.73 number would only be needed for some really cheap ballasts, or maybe for circuits that run a lot of computers that don't try to filter out the harmonics.

 

[Dennis Alwon]

It sounds like the 173% is based on worst case scenario. Even at 20% if 2 circuits were off there would be 3.2 amps on each of the phases, as you stated-- 16+ 3.2 + 3.2= 22.4 amps. #12 is rated 25 amps so it really wouldn't be an issue and normally there isn't 16 amps on all the circuits and if there were most of the time the phases are more likely to be balanced.

So I am guessing that #10 is never speced for this type of install. Someone asked me this today and I didn't really have an answer other than I have never seen it. Since my experience with commercial is limited I wondered if others have seen it.

 

[Dennis Alwon]

BTW, thank you for the useful info. :thumbsup:

 

[BullsnPyrs]

If you really want to get into the subject NFPA has published a report on the the subject.
http://www.nfpa.org/assets/files/PDF/Research/RFEvaluationOfNonLinearPowerWiringRequirements.pdf

 

[Rick Christopherson]

Unfortunately, the NEC has used the term "harmonics" as a blanket term to include phase-shifted signals too. This is what can make this a little misleading in terms of current in the neutral. There are common cases where single-phase loads with differing power factors that will exceed the rating of the neutral by a large amount. As a matter of fact, it can get really tricky, because simply reversing 2 of the phase conductors can make this current increase or decrease above the limit of the neutral rating. Having one motor load and one resistive load is enough to do this.

 

[Besoeker]

We recently had a thread where we mentioned that a neutral must be sized at 173% when harmonics are present. I understand that this is based on 1.73 that we use for 3 phase but I am curious what are the factors that go into determining what the load on that neutral is going to be.

I don't know what other do but I've generated spreadsheets for various types of load I come across. This is part of one to illustrate that:

I(h) is given by SQRT(2)*$L$5*SIN(HN*$P8+ATAN)
HN being the harmonic number.
An and Bn are sine and cosine components of the the harmonic in turn derived from the instantaneous values of the waveform being considered.

Describing it in words makes it sound more complicated than it is.
Anyway, that's how I do it for each harmonic.

If I save a waveform from the scope I generally use Fourier to extract the harmonic components. I gave an example of that in another thread....

For instance, we have a 3 phase load and all the circuits of the MWBC are loaded to 15 amps. We use the formula to calculate neutral current. In excel I would use this formula =SQRT(((A5*A5)+(B5*B5)+(C5*C5))-((A5*B5)+(B5*C5)+(A5*C5))).

I think that's valid only if all three phases operate at unity or at the same power factor.
This came up a while back.

 

[T.M.Haja Sahib]

BTW, thank you for the useful info. :thumbsup:

One more information I want to add is that in a MWBC (single phase),the currents from outer conductors flow in the opposite direction in the neutral with the result that the third harmonic currents cancel out.So I do not think it is necessary to oversize the neutral conductor beyond the size of the phase conductor.
It is a different matter for a MWBC (Three phase).............

 

[iMuse97]

One more information I want to add is that in a MWBC (single phase),the currents from outer conductors flow in the opposite direction in the neutral with the result that the third harmonic currents cancel out.So I do not think it is necessary to oversize the neutral conductor beyond the size of the phase conductor.
It is a different matter for a MWBC (Three phase).............

What seems to be the issue is that phase shifts, leading or lagging, and harmonics can require an oversize conductor even when the 120 (single phase) loads would normally cancel eachother on the neutral.

 

[T.M.Haja Sahib]

What seems to be the issue is that phase shifts, leading or lagging, and harmonics can require an oversize conductor even when the 120 (single phase) loads would normally cancel eachother on the neutral.

For purely non-harmonic 120V loads with 90 degree leading on one side and 90 degree lagging on the other side,current in the neutral is two times the phase current.

But I believe for non-linear 120V loads on the two sides,the triplen harmonics still cancel in the neutral..................

 

[kingpb]

On three phase systems, the three phases of the power system are 120? out of phase. If we have balanced 60 Hertz currents on our three phase conductors, our neutral current will be zero. It can be shown mathematically that the neutral current (assuming only 60 Hertz is present) will never exceed the highest loaded phase conductor. Thus, overcurrent protection on phase conductors also protects the neutral conductor, even though there is not any overcurrent protective device in the neutral conductor.

When these harmonic currents come together on the neutral, rather than cancel, they actually add and the third harmonic of each of the three phase conductors is exactly in phase. Therefore, there can be more current on the neutral conductor than on the phase conductors, and the neutral conductors are no longer protected by mathematics.

These harmonic currents create heat. This heat over a period of time, will raise the temperature of the neutral conductor. This rise in temperature can overheat the surrounding conductors and cause insulation failure. These currents also will overheat the transformer sources which supply the power system. This is the most obvious symptom of harmonics problems; overheating neutral conductors and transformers. Other symptoms include nuisance tripping of circuit breakers, malfunction of UPS systems and generator systems, metering problems, computer malfunctions, and over-voltage problems.

In three phase circuits with shared neutrals, it is common to oversize the neutral conductor up to 200% when the load served consists of non-linear loads. Most manufacturers of system furniture provide a #10 AWG conductor with 35 amp terminations for a neutral shared with the three #12 AWG phase conductors. In feeders that have a large amount of non-linear load, the feeder neutral conductor and panelboard bus bar should also be oversized.

On three phase branch circuits, another philosophy is to not combine neutrals, however the panelboard neutral bus and feeder neutral conductor still must be oversized. You can use K-rated transformers that have more copper and can take more heat, or Delta-wye transformers trap the triplen harmonics in the delta winding, but require more copper (oversized) to take the additional heat.

You can use special filters to mitigate but they can be expensive, and metering must be True RMS to measure the harmonic currents.