electronic ballasts, THD, and non-linear loads

Status
Not open for further replies.
electrofelon

electrofelon

Senior Member
Location
Cherry Valley NY, Seattle, WA
I have been doing lots of reading on the subject and I am just not clear on several things:

1. Isnt it true that we cant generalize that an electronic ballast is a non-linear load? With modern power electronics and designs, is it fair to say that a modern electronic ballast could be a linear load?

2. Does the term "total harmonic distortion" essentially quantify how non linear a load is? If so, about what would be the value that would differentiate between a linear and non linear load in the terms of the NEC?

3. Can the neutral current in a 3 phase MWBC be calculated from the THD value? Is it as simple as THD X current X 3 (assuming 3 identical loads, one on each phase)?

I appreciate any clarifications on the topic.
 
electrofelon said:
I have been doing lots of reading on the subject and I am just not clear on several things:
Click to expand...

1. Isnt it true that we cant generalize that an electronic ballast is a non-linear load? With modern power electronics and designs, is it fair to say that a modern electronic ballast could be a linear load?
I haven't come across any electronic ballast that was linear. They may exist. Sinewave input VFDs exist. But at significant cost premium.



2. Does the term "total harmonic distortion" essentially quantify how non linear a load is? If so, about what would be the value that would differentiate between a linear and non linear load in the terms of the NEC?
I've seen THD used that way. NEC, question I can't answer.

3. Can the neutral current in a 3 phase MWBC be calculated from the THD value? Is it as simple as THD X current X 3 (assuming 3 identical loads, one on each phase)?
For 3-phase+neutral third order harmonics add arithmetically in the neutral - sometimes called triple-n harmonics. Foe single phase non-linear loads, third is often the dominant harmonic.
 
Besoeker said:
2. Does the term "total harmonic distortion" essentially quantify how non linear a load is? If so, about what would be the value that would differentiate between a linear and non linear load in the terms of the NEC?
I've seen THD used that way. NEC, question I can't answer.
Click to expand...

What confuses me Is I was reading an article that had data on all sorts of ballasts and this statement:

"It is a common misconception that electronic ballasts increase THD. Currently
available electronic ballasts actually decrease the THD on an electrical system
compared to a system applying magnetic ballasts."

I dont understand. How can a magnetic ballast be a linear load and have a higher THD than an electronic ballast that is supposedly a non linear load?
 
electrofelon said:
I dont understand. How can a magnetic ballast be a linear load and have a higher THD than an electronic ballast that is supposedly a non linear load?
Click to expand...
Where was it stated that a iron core magnetic ballast a linear load?
Perhaps, nowhere because it is a non linear load, a rich source of 3rd harmonics.
 
140509-0929 EDT

electrofelon:

I have a Cree 10 W LED that has a measured power factor of 0.98 when connected directly to a sine wave source. That I would call a linear load.

Connect that bulb or a standard incandescent to the output of a phase shift dimmer, and that combination seen at the input to the dimmer is a non-linear load.

Most CFLs have rather poor power factors, and are non-linear. LEDs are better because of government regulations. CFLs were designed before the new rules.

This last weekend was U of M graduation. Our voltage waveform now is less flat topped compared with weeks before.

Power factor and non-linear do not have a one-to-one correspondence. An air core inductor has a poor power factor, but is a linear load (no harmonic content). An iron core inductor driven into some degree of saturation has a poor power factor, and is a non-linear load (has non-sinusoidal current waveform and harmonic content).

I do not have a criteria for when a waveform should be called non-linear. Generally a waveform is good or bad. In audio work and some other applications a very small amount of distortion can be significant. A small fractional percent harmonic distortion in audio to a critical listener is important, but not a problem in power system applications.

.
 
Sahib said:
Where was it stated that a iron core magnetic ballast a linear load?
Perhaps, nowhere because it is a non linear load, a rich source of 3rd harmonics.
Click to expand...

Maybe I am off base but I thought A magnetic ballast was generally a linear load.

I have a Cree 10 W LED that has a measured power factor of 0.98 when connected directly to a sine wave source. That I would call a linear load.
Click to expand...

Isnt power factor and non-linear loads two different things? My understanding was that power factor is a lead or lag of the current or voltage where a non linear load is a distortion of the sine wave. Not saying you couldnt have a load that does both things, but just a load with very low power factor doesnt mean it is non linear right? Ok you seem to elaborate further in your post, but the first sentence threw me off.
 
electrofelon said:
What confuses me Is I was reading an article that had data on all sorts of ballasts and this statement:

"It is a common misconception that electronic ballasts increase THD. Currently
available electronic ballasts actually decrease the THD on an electrical system
compared to a system applying magnetic ballasts."

I dont understand. How can a magnetic ballast be a linear load and have a higher THD than an electronic ballast that is supposedly a non linear load?
Click to expand...
Most likely that article was written for people outside of North America, or whomever researched it failed to differentiate data points.

CE and IEC rules now dictate that lighting ballasts be inclusive of power factor correction techniques and as part of that, harmonic mitigation. They accomplish this in various ways, but the end result makes that statement somewhat true. One of the ways is with what is called an FFE (Fundamental Front End), which is a semi-active rectifier. The ballast front end is partially controllable so that it can actually present a slightly leading power factor, which can be interpreted as having some corrective effect on existing harmonics. I think that's a bit of a stretch to apply that concept universally, because not all electronic ballasts have an FFE (they are more expensive).

IEEE 519, the standard we have been using for years, is changing to become more stringent with regard to correcting THD at each individual load, as opposed to TDD, the effect of harmonics on an entire system. But there is intense lobbying by mfrs to delay enforcement of the changes because it will mean cost increases. In most countries outside of the US, regulatory agencies have decided not to wait. That's why you see things like CFLs and LED lights already having better performance in that regard. Their power supplies are designed to be universal so they can sell them anywhere, which means we end up getting the better harmonic mitigation despite the lack of local regulatory influence. Eventually this will work its way into the stand-alone ballast world as well, when companies that resist find themselves unable to compete on a shrinking market share in North America alone.
 
A linear load with a phase shift will have a low displacement power factor but could have a high distortion power factor.
A linear load can be characterized by a complex impedance with constant coefficients.
With a non-linear load, doubling the voltage will not necessarily exactly double the current. And the shape of the current waveform will not match the shape of the applied voltage.
You can describe this shape difference in terms of phase delayed turn on or off, voltage sensitive turn on or off (neon bulb or rectifier circuit for example) or in terms if the amplitude of harmonics needed to reproduce that shape.

Tapatalk!
 
140509-1126 EDT

Besoeker:

Power factor as I measured it on the Cree is going to be close to the definition
PF = W/(Vrms*Irms). This was measured with a Kill-A-Watt.

I could use a definition of displacement PF where it was based on a Fourier analysis of the waveform and look at the phase shift of the fundamental component, but I don't know how or whether it makes sense to separate out something that would be called the distortion power factor when there is both a fundamental phase shift and then the other components from the higher harmonics.


electrofelon:

Power factor is as I defined above. Either waveform distortion or linear phase shift will cause a reduction in power factor from 1.00 .

Suppose we have a sine wave with the top flattened. That will have zero phase shift for the fundamental, but it has harmonic content and a reduced power factor.

Use a phase shift dimmer and turn on at 90 deg. That will have a large amount of harmonic content, and phase shift of the fundamental.

If we have a switch that turns on at 45 and off at 135, and on at 225 and off at 315, then the fundamental goes back to 0 phase shift, and there are many harmonics.

.
 
electrofelon said:
Maybe I am off base but I thought A magnetic ballast was generally a linear load.



Isnt power factor and non-linear loads two different things? My understanding was that power factor is a lead or lag of the current or voltage where a non linear load is a distortion of the sine wave. Not saying you couldnt have a load that does both things, but just a load with very low power factor doesnt mean it is non linear right? Ok you seem to elaborate further in your post, but the first sentence threw me off.
Click to expand...


I am followign this with interest. I was under the same conception (misconception) as electrofelon. My understanding of THD is that it is created by the high speed switching function of electronics, that cop the sine wave as opposed to suppressing the sine wave to control intensity or produce other voltages, DC etc. This function then sends harmonic pulses back along the line side. Those harmonics are all or many available harmonics, including second third fourth fifth sixth etc. The only ones that are a concern though being multiples of three because they add together on a shared neutral. Magnetic ballasts, being a coil did not produce harmonics back along the line from my understanding. They create the lag referenced.

Sorry for the rambling, but I wanted to spell out my conceptions so that they can be corrected, because I amnot sure what parts may be incorrect.
 
gar said:
140509-1126 EDT

Besoeker:

Power factor as I measured it on the Cree is going to be close to the definition
PF = W/(Vrms*Irms). This was measured with a Kill-A-Watt.

I could use a definition of displacement PF where it was based on a Fourier analysis of the waveform and look at the phase shift of the fundamental component, but I don't know how or whether it makes sense to separate out something that would be called the distortion power factor when there is both a fundamental phase shift and then the other components from the higher harmonics.
Click to expand...

If you take a plain rectifier, a B6U, on a three phase supply feeding a level load (big choke required) the best total power factor you can get is 0.958.
 
A magnetic ballast for a fluorescent tube will still have to deal with the fact that the arc will not strike until some minimum voltage is reached.
It can smooth out this current step some but there must still be some waveform distortion.
An electronic ballast could, at least in theory, use Besoeker's ideal rectifier to produce DC and hide all of the tube's non linearity from the supply circuit.

Tapatalk!
 
140509-1509 EDT

Strathead:

As soon as you distort the shape of a sine wave in any way you create harmonic content. Only an absolutely pure sine wave shape has no harmonics.

.
 
gar said:
140509-1509 EDT

Strathead:

As soon as you distort the shape of a sine wave in any way you create harmonic content. Only an absolutely pure sine wave shape has no harmonics.

.
Click to expand...

I understand, but I always thought a magnetic ballast was a pure inductive load, no sine wave distortion there.
 
GoldDigger said:
A magnetic ballast for a fluorescent tube will still have to deal with the fact that the arc will not strike until some minimum voltage is reached.
It can smooth out this current step some but there must still be some waveform distortion.
An electronic ballast could, at least in theory, use Besoeker's ideal rectifier to produce DC and hide all of the tube's non linearity from the supply circuit.

Tapatalk!
Click to expand...

The B6U has 20% 5th. Or worse.
 
Ok lots of technical discussion which is interesting and valuable, but bringing it back to the installation realm, lets do a survey: Assume both are 120Y/208 services with MWBC's.

1) You are asked to install a bunch of typical magnetic HPS ballasts that are the only load on the service. In terms of neutral sizing and the neutral as a CCC, what do you do?
2) You are asked to install a bunch of modern T8 light fixtures and they are the only load on the service. In terms of neutral sizing and the neutral as a CCC, what do you do?
 
140509-0908 EDT

Strathead:

As soon as a magnetic material (ferromagnetic) is added into the field of a coil of wire the current becomes non-sinusoidal. The more the material is saturated the greater the current waveform distortion. Under many conditions you may not see noticeable distortion because it is masked by other loads.

My photo at P8 at http://beta-a2.com/EE-photos.html shows the steady state magnitizing current to a transformer with no secondary load. This is far from a sine wave. But this is what can be described as an inductor.

.
 
Status
Not open for further replies.
Top