non linear loads and how they affect neutral

[winnie]

Again, trying to keep it relevant to the OP's question:

The calculation requires that you determine the net 'triplen' harmonic content of the current drawn by your lamps, and then add up this content for all 3 phases. As Besoeker keeps stating, the Triple harmonic currents add on the neutral.

However as everyone else has been saying, you can approach this from the other side of the calculation to determine if there is likely a problem.

1) You look at the power factor of the lamp, or the specification for the total current drawn. The idea is to figure out the maximum 'not utilized' current going to the lamp in addition to that actually powering the lamp.

You have a 22W lamp operating at 277V. If it had unity power factor it would require a current of 0.079A. Say the datasheet tells you that the lamp actually draws 0.125A; this means that 0.051A of current is somehow 'circulating'. This could be standard 'displacement power factor', or it could be caused by harmonics. If you know the power factor; take 1/PF * <ideal current> to get the total current drawn.

2) Calculate the total current on each of your supply legs. In this case you have 108 lamps on each leg, for a current of 13.5A (this with my example of 0.125A per lamp; you need the actual datasheet value).

3) Calculate the total _unutilized_ current on each supply leg 0.051*108= 5.5A

4) Now you make the _assumption_ that _all_ of the 'not utilized' current is triplen harmonic content. This is _not_ a valid assumption, because there are almost certainly non-triplen harmonics such as 5th or 7th. The point is to do a _fast_ calculation that will tell you if you need to look closer or not. Again: triplen harmonics add, so in the worst case we get 3 * 5.5A = 16.5A on the neutral.

If you have 12AWG conductors this is likely _not_ a problem.

Factors which could push things in either direction:

The standard for small LED lamps permit 3rd harmonic content of up to 86% of fundamental!!!. Run the same calculation above, but with this 86% value and you've overloaded your neutral.

The specs for a lamp like the one you are considering state a PF of > 0.9 which suggests a lamp current of 0.088A; run the same calculation above and you will find very little current on the neutral. The 9 mA of 'unused' current is certainly no all triplen harmonic content, but even if it is there would not be a problem.

-Jon

 

[Besoeker]

Again, trying to keep it relevant to the OP's question:

The calculation requires that you determine the net 'triplen' harmonic content of the current drawn by your lamps, and then add up this content for all 3 phases. As Besoeker keeps stating, the Triple harmonic currents add on the neutral.

Thank you for that.

However as everyone else has been saying, you can approach this from the other side of the calculation to determine if there is likely a problem.

1) You look at the power factor of the lamp, or the specification for the total current drawn. The idea is to figure out the maximum 'not utilized' current going to the lamp in addition to that actually powering the lamp.

No, not really. It isn't a power factor issue. It's a third (triple) harmonics issue.
I did the calcs from first principles for phase controlled loads, both leading and trailing edge because I had to and I could mathematically express the wave forms and so determine all the integer harmonics.

 

[winnie]

No, not really. It isn't a power factor issue. It's a third (triple) harmonics issue.
I did the calcs from first principles for phase controlled loads, both leading and trailing edge because I had to and I could mathematically express the wave forms and so determine all the integer harmonics.

IMHO it may be considered a distortion power factor issue, because distortion power factor expresses the difference between the the fundamental current which delivers power to the load and the harmonic current which does not deliver power to the load.

The integral of harmonic current * fundamental voltage is zero, thus (if voltage distortion is low) harmonics don't deliver power to the load.

Since 'power factor' encapsulates all of the current flow that doesn't actually deliver power to the load, it of necessity includes harmonics and in particular the triplen harmonics. If one assumes that _all_ of the power factor is caused by triplen distortion, and then uses the thus calculated triplen currents to estimate neutral current, one gets a crude maximum estimate of the neutral current.

However, huge Mea culpa: I made a _huge_ error in how I was handling the power factor, simply assuming that total rms current is rms fundamental + rms harmonic. I've just demonstrated to myself that this is not correct, and in fact wrong by a huge margin. In a very simple example of fundamental + 50% third harmonic, the distortion power factor is 0.89.

I retract the statement that a power factor > 0.9 implies low harmonic currents. In fact the harmonic currents can be quite high even with a very good power factor.

-Jon

 

[GoldDigger]

Not trying to quibble, but the idea that a power factor of 1 means no harmonic content is not entirely true. There is the assumption that the supply voltage is a sine wave with no harmonics.
If you look at a square wave applied voltage, you can have a supply current which is also a square wave, with a resistive load. The PF will be 1.
And the harmonic content will be that of the supply waveform, in this case quite large.

If the utility power has serious voltage distortion, that can cause additive current in the neutral even with balanced loads of perfect PF.

 

[Besoeker]

IMHO it may be considered a distortion power factor issue, because distortion power factor expresses the difference between the the fundamental current which delivers power to the load and the harmonic current which does not deliver power to the load.

No disagreement with that.
But what this topic is about non-linear loads and neutral currents. Specific harmonics.

 

[Smart $]

With all the rhetoric in this thread, I'm going back to suggesting the test setup with measurement of actual neutral current.

 

[Besoeker]

With all the rhetoric in this thread, I'm going back to suggesting the test setup with measurement of actual neutral current.

I've done that in real life.The Chicago Beach Resort Development (CBRD)No rhetoric required.

 

[Open Neutral]

Reading this thread, I ask a somewhat loaded question....

vs. 25 year ago, what loads are there that are linear? [residence, commericial, industrial...]

Well, resistive heaters, incandescent lighting, across the line [no VFD] motors, and ........
And non-linear? Switcher suppliers in everything from iPhone chargers to microwave ovens, VFD's, LED/CF lighting, ranges [I bet], and......

It's a wonder the RFI and harmonic current situation is not worse than it is....

(We should have listened to Edison.)

 

[Besoeker]

Reading this thread, I ask a somewhat loaded question....

vs. 25 year ago, what loads are there that are linear? [residence, commericial, industrial...]

Well, resistive heaters, incandescent lighting, across the line [no VFD] motors, and ........
And non-linear? Switcher suppliers in everything from iPhone chargers to microwave ovens, VFD's, LED/CF lighting, ranges [I bet], and......

It's a wonder the RFI and harmonic current situation is not worse than it is....

(We should have listened to Edison.)

Valid points. The tests I did 20 years ago at that pumping station were an indication of worse to come. The voltage waveform at our home has the signature flattened top of third harmonic. The current is anything but sinusoidal.........
Those non-sinusoidal domestic and commercial loads are small. But the make up for that in sheer numbers.

 

[winnie]

No disagreement with that.
But what this topic is about non-linear loads and neutral currents. Specific harmonics.

Agreed, the topic is about harmonics.

I am simply stating that you can calculate the maximum possible value of those specific harmonics by using power factor. If the only information that is easily available is power and power factor, then you can calculate a maximum bound for the triplen harmonics, run your calculation to see if there is a problem, and if there isn't a problem then you can move on. This maximum bound comes by assuming that _all_ of the non-unity power factor is caused by triplen harmonics, which is a conservative assumption.

If you are calculating a 3 MW system, you probably need a more accurate calculation , but for 10A system probably wired with 12AWG wire an upper bound calculation of neutral current is probably sufficient.

-Jon

 

[junkhound]

This thread brings up a question:

What percentage of electricians have a spectrum analyzer or scope with FFT function?

Advent of scopemeters about 30 years ago probably caused that number to jump significantly, eh?

Back in the 1950s when I started if you wanted a quantitative number for harmonics it meant a laborious plotting of tuned receiver measurements.
Or a very well trained eye to do 'eyeball Fourier' from a flickering CRT trace.

 

[Open Neutral]

Valid points. The tests I did 20 years ago at that pumping station were an indication of worse to come.

I deliberately buried my lede:

Well, resistive heaters, incandescent lighting, across the line [no VFD] motors, and ........

My silent point was these are all disappearing from the grid.....leaving what?

I wonder what the grid looks like at a server farm, such as 350 E. Cermak. I know one of the builders. There are multiple [Redundancy is Job 1!] 132KV vaults in the basement.

A plurality of that load will be switching supplies on the servers. Not sure how much of the total is HVAC load; the cooling is actually outsourced a block away but there are surely many air handlers on site.

Side thought: How [un]well do distribution transformers cope with non-linear loads? I'd think they'd help clean up the waveform their feed sees, but at what cost?

 

[Besoeker]

Agreed, the topic is about harmonics.

I am simply stating that you can calculate the maximum possible value of those specific harmonics by using power factor. If the only information that is easily available is power and power factor, then you can calculate a maximum bound for the triplen harmonics, run your calculation to see if there is a problem, and if there isn't a problem then you can move on. This maximum bound comes by assuming that _all_ of the non-unity power factor is caused by triplen harmonics, which is a conservative assumption.

Well, both the compressor motors on a site we support are about 0.8pf on fixed speed with no harmonics.
My method for calculating harmonics was to plot the calculated current waveform in Excel and extract the harmonics from that.

 

[Besoeker]

I deliberately buried my lede:

Well, resistive heaters, incandescent lighting, across the line [no VFD] motors, and ........

My silent point was these are all disappearing from the grid.....leaving what?

I wonder what the grid looks like at a server farm, such as 350 E. Cermak. I know one of the builders. There are multiple [Redundancy is Job 1!] 132KV vaults in the basement.

A plurality of that load will be switching supplies on the servers. Not sure how much of the total is HVAC load; the cooling is actually outsourced a block away but there are surely many air handlers on site.

The probably have Active Harmonic Filter(s). That's what we did for one in London.

Side thought: How [un]well do distribution transformers cope with non-linear loads? I'd think they'd help clean up the waveform their feed sees, but at what cost?

You can mitigate harmonics using transformers with multiple secondaries. Larger drives often have Ddyn11 unit transformers.

 

[Smart $]

...
My method for calculating harmonics was to plot the calculated current waveform in Excel and extract the harmonics from that.

hmy: We must have actual data.

 

[Besoeker]

hmy: We must have actual data.

If you know the application, you can calculate the waveform(s).
That's what I did for the CBRD project.

 

[Smart $]

If you know the application, you can calculate the waveform(s).
That's what I did for the CBRD project.

I swear some things just make their way right past you. :slaphead:

 

[Besoeker]

I swear some things just make their way right past you. :slaphead:

What exactly do you imagine went right past me?

 

[mivey]

What exactly do you imagine went right past me?

The sarcasm.

See post #33. That was after a calculated ceiling was proposed for an application.

IOW, you can calculate a ceiling. Once there, apply the knowledge and experience you have with the application to decide if it is reasonable to say the 3rd harmonic content will stay below that ceiling. See winnie's brief.

 

[Besoeker]

The sarcasm.

A poor attempt but water under the bridge.

IOW, you can calculate a ceiling. Once there, apply the knowledge and experience you have with the application to decide if it is reasonable to say the 3rd harmonic content will stay below that ceiling. See winnie's brief.

For the projects I was involved with there were mandatory limits you had to comply with so you had to do the calculations prior to submitting a bid. Getting it wrong wasn't an option.