Typical Harmonics for modern computers and LED fixtures

[steve66]

I've tried searching online with zero luck. We know that most of the fluorescent ballasts had a 10% THD spec. Can anyone find any THD ratings for LED drivers or fixtures?

Since nobody seems to mention them, I'm guessing the THD is so low as to not be an issue. I'm also wondering if Energy Star of any other regulations put a limit on the harmonics that a fixture can draw.

Along with that, I'm wondering the same thing for a typical computer. What's the total THD that it creates? Again, it seems like there may be limits imposed by some listing or reg. agency.

 

[JoeStillman]

I can't imagine that the LED drivers have less THD than ballasts. But because they draw so much less current for an equivalent amount of light, the percentage of total harmonics is going to be way down.

As long as we're still using the NEC for residential service sizing, the LED current distortion is going to be insignificant, so I think the Energy Stars are in alignment. 3 VA/ft² for general lights and receptacles still applies, even though our lighting efficacy is in another galaxy.

 

[steve66]

I can't imagine that the LED drivers have less THD than ballasts. But because they draw so much less current for an equivalent amount of light, the percentage of total harmonics is going to be way down.

As long as we're still using the NEC for residential service sizing, the LED current distortion is going to be insignificant, so I think the Energy Stars are in alignment. 3 VA/ft² for general lights and receptacles still applies, even though our lighting efficacy is in another galaxy.

Well, with the 2014 NEC, if we follow the monitoring and alarming requirement, we can actually start using the Energy Code limits for lighting. So that brings it way down to around 1 VA/sf - which is much more reasonable.

But this has nothing to do with the watts per square foot. It has to do with harmonics on the neutral conductor, and if the neutral wire needs to be counted as a current carrying conductor.

 

[__dan]

Modern devices of the last 20 years or so are so much better than they used to be. Harmonics was a nasty problem in the 1980's. Now I generally regard modern PC's, ballasts, drivers, as pretty clean loads. Product manufacturers have mostly recognized the problem and mitigate it. It's very rare now to see a low quality PSU that runs hot and makes an interference.

For the purposes reduced neutral sizing of a feeder or counting the neut as a ccc, the 'generally true' statement above cannot be applied universally to all cases. Being conservative or not having test data of the actual load, the existential case would apply. There exists at least one case where the switching PSUs will be noisy.

If you cannot guarantee the (mostly rectifier front end) load will be a clean linear (sinewave) load, that is most cases so the neutral would have to be counted.

The feeder neutral will see a lot of cancellation but probably not perfectly. If it comes to a money question because it is increasing the derating factor, not counting it will not result in an overloading of anything, but it may be a decision made by others, for the AHJ, or for the design engineer so all the bidders are bidding the same thing.

It would only be an issue if someone wants to claim that it is required and they paid for it. Otherwise the owner could decide if they want the upsize that the derating would indicate, or they want to save the money and live with the decision.

 

[Besoeker3]

I've tried searching online with zero luck. We know that most of the fluorescent ballasts had a 10% THD spec. Can anyone find any THD ratings for LED drivers or fixtures?

Since nobody seems to mention them, I'm guessing the THD is so low as to not be an issue. I'm also wondering if Energy Star of any other regulations put a limit on the harmonics that a fixture can draw.

Along with that, I'm wondering the same thing for a typical computer. What's the total THD that it creates? Again, it seems like there may be limits imposed by some listing or reg. agency.

I don't think any computer on its own produced very high levels of THD. It's in aggregate they can be problematic.

A tale from about two decades ago. We had a project to supply and install about seven VFDs in a pumping station in the middle of a residential area. (In was the Roman town of Veralulamium as it happens.).The two largest drives were 200kW and, based on calculations we made them 12-pulse. As was usually the case, we had to prove that our installation complied with the Electricity Council THD regulations in force at the time.

This was done at the point of common coupling with other users which in this case was the 11kV. 400V 1500 kVA transformer in the yard of the pumping station. The THD exceeded the limits. Even with none of our drives running. In fact, there was no discernible difference with the drives running.

Needless to say, that somewhat surprised me. I was used to dealing with harmonics for industrial power electronics. What I was seeing was the combined effect of all that non-linear residential loads. That was a bit of an eyeopener....

 

[Jraef]

I've tried searching online with zero luck. We know that most of the fluorescent ballasts had a 10% THD spec. Can anyone find any THD ratings for LED drivers or fixtures?

Since nobody seems to mention them, I'm guessing the THD is so low as to not be an issue. I'm also wondering if Energy Star of any other regulations put a limit on the harmonics that a fixture can draw.

Along with that, I'm wondering the same thing for a typical computer. What's the total THD that it creates? Again, it seems like there may be limits imposed by some listing or reg. agency.

All SMPS type drivers / electronic ballasts / power supplies etc. use the same basic technology and to achieve Energy Star and IEC 61000-3-2 requirements for efficiency and True Power Factor*, they must all now use Boost Converter Switched Mode Power Supply technology. A Boost Converter SMPS adds a feedback loop in the system that is used to alter the PWM pattern in order to decrease the distortion power factor as much as possible and that has the effect of decreasing the I-THD. Given that the proliferation of LED lighting has taken place AFTER the advent of Boost Converter technology, you would be safe in assuming it is 10% I-THD or less (I've seen one report claiming 8.79% I-THD and bragging that it is the best in the industry).

To Besoeker's point though, the effect of LED lighting systems on the total I-THD of a facility is only going to end up relevant if the ratio of I_load to I_{short circuit} is very low and given that the very use of LED lighting drastically reduces the I_load in the first place, it has the effect of curing it's own problem somewhat.

*TPF is different in that it takes into account the effects of displacement AND distortion PF. Utility metering systems often don't see distortion PF, but the Energy Star and IEC regs require that it be taken into account.

 

[GoldDigger]

Utility metering systems often don't see distortion PF, but the Energy Star and IEC regs require that it be taken into account.[/I]

An interesting thought!
Since utility meters are designed on basic physics to correctly measure real power (small overestimates resulting from not measuring neutral voltage for reference notwithstanding), this must mean that the limitation is in the way that the Irms and Vrms (or more properly Irms x Vrms) are measured directly or calculated.
Use of true RMS metering for both of those quantities would automatically take both varieties of PF into account. Is the utility meter instead just looking at the average phase shift angle between voltage and current and calculating (Irms x Vrms) from real power and the assumption that there is no waveform distortion?

 

[gar]

190326-2216 EDT

I believe utility meters measure the instantaneous product of voltage and current and average this over some time like 1 second, and then accumulate those measurements. See the kind of chip used in a TED System.

.

 

[Besoeker3]

*TPF is different in that it takes into account the effects of displacement AND distortion PF. Utility metering systems often don't see distortion PF, but the Energy Star and IEC regs require that it be taken into account.

Yes. agreed. But the problem is how to police it in residential areas. The total on a distribution transformer supplying say, 100 residences may exceed the levels permitted without any single consumer remotely contravening those limits. How do you enforce the regs in this situation?

And that's kinda where we are. The aggregate as I mentioned earlier.

 

[steve66]

All SMPS type drivers / electronic ballasts / power supplies etc. use the same basic technology and to achieve Energy Star and IEC 61000-3-2 requirements for efficiency and True Power Factor*, they must all now use Boost Converter Switched Mode Power Supply technology.

I don't have a copy of IEC 61000-3-2, but from what I can find by googling it seem to imply there may be no requirements to follow this standard in the US.

And even if a manufacturer did follow that standard, it doesn't look like it applies to anything operating at less than 240 volts or anything operating at less than 75 watts. That would exempt most computers and LED interior lights.

Also, I do see Energy Star has limits on standby power consumption, but again, I can't find any limits on harmonics.

 

[Jraef]

I don't have a copy of IEC 61000-3-2, but from what I can find by googling it seem to imply there may be no requirements to follow this standard in the US.

And even if a manufacturer did follow that standard, it doesn't look like it applies to anything operating at less than 240 volts or anything operating at less than 75 watts. That would exempt most computers and LED interior lights.

Also, I do see Energy Star has limits on standby power consumption, but again, I can't find any limits on harmonics.

Sorry, I didn't mean to imply that Energy Star had anything to do directly with harmonics, it doesn't. But part of the Energy Star certification does involve true power factor (TPF) so to meet it, they use Boost Converter technology, which INDIRECTLY results in lower harmonics than older designs.

 

[steve66]

All good responses, but I still haven't been able to find any specific requirments or even typical harmonics generated by either computers or LED lights.

Back when we had fluorescent ballasts, we knew. It was either 10% or 20% depending on the ballast specified.

Now nobody even mentions harmonics, which makes me think they are much lower, at least for LED light fixtures.

I may have to see if I can find a way to measure some.