An office area in a building is feed with a 45 KVA 480/208 delta/wye transformer. The load on the transformer is expected to be below 40%. There will be 40 computers in the area and about 50 circuits feeding the area. Office outlets are the only load on the panel. Should the neutrals be dedicated or up sized for multiwire circuits or is diversity enough? Should the transformer be K rated?
Office Harmonics
- Thread starter DWH
- Start date
- Status
Mayimbe
Senior Member
- Location
- Horsham, UK
There is a paper that I used when I got this the same question a couple of months ago. Its available on the internet. the name is:
Mathematical Modeling of Current Harmonics
Caused by Personal Computers by
Rana Abdul Jabbar Khan, and Muhammad Akmal
It really helped me.
I got 35 computers in my case, and my transformer was 100kVA 480/208. My load was 50% and I recomended that it got to be a K-9 transformer. I guess you will have this same answer due to the numbers of computers you are delling with.
- Location
- Wisconsin
- Occupation
- PE (Retired) - Power Systems
I am one of the believers, that K-Rated transformers are not much more than a marketing gimic.
Yes, there are some applications that require an oversized (200%) neutral.
But, maybe only 1-2% of transformers feeding computer loads require K-ratings. And the only way to know what K-Rating is actually appropriate is to install all of the loads and then measure the actual harmonics.
The diversity of switched mode power supplies helps mitigate the affects of desktop computers.
Yes, there are some applications that require an oversized (200%) neutral.
But, maybe only 1-2% of transformers feeding computer loads require K-ratings. And the only way to know what K-Rating is actually appropriate is to install all of the loads and then measure the actual harmonics.
The diversity of switched mode power supplies helps mitigate the affects of desktop computers.
brian john
Senior Member
- Location
- Leesburg, VA
I have completed countless harmonic surveys and based STRICTLY on what I have seen in the field a typical office building seldom loads transformers above 50%. At this level, I saw no issues of concern.
I agree with Jim and the marketing ploy, for the money you could go with a 75 kva and have future capacity if you are really concerned.
As for the neutrals the copper industry really loves this (selling copper). I recently did a survey of a large data center, in no case did we see neutral currents higher than phase currents with close to balanced panels and typically the current was about 2/3's of a balanced panels load.
I agree with Jim and the marketing ploy, for the money you could go with a 75 kva and have future capacity if you are really concerned.
As for the neutrals the copper industry really loves this (selling copper). I recently did a survey of a large data center, in no case did we see neutral currents higher than phase currents with close to balanced panels and typically the current was about 2/3's of a balanced panels load.
Varczar
Member
- Location
- Birmingham, AL
skeshesh
Senior Member
- Location
- Los Angeles, Ca
I can agree with Jim for this particular case (and for most small scale applications). At the industrial level many large clients do look to meet IEEE std. 519. Overall I do think k-factor solutions are a bit crude given how far filtering technology has come. Sizing neutrals is still an ongoing debate in my mind, but I have to say I've heard from many quality engineers that have worked on impressive projects that there is meric to neutral oversizing, but I guess that's been pretty much entirely related to very large-scale projects with strong engineering crews, expensive software, and a LOT of data collection.
gar
Senior Member
- Location
- Ann Arbor, Michigan
- Occupation
- EE
091124-1158 EST
I have no advice on what you should do other than to determine your actual load.
Here is my load. Two computers, one 20 yrs old, one 10 yrs old; one 20" LCD (about 30 W), and one 20" CRT. Total for all of these is 388 VA, 271 W, and 3.38 A. Calculated PF = 0.7 . My average per computer is 388/2 = 194 VA.
High performance computers will require more power than mine, but the use of LCDs will be less power than my CRT. The CRT with a mostly white screen is 95 W and 140 VA. The LCD is 27 W and 44 VA.
Is the lighting load on the same transformer? How many coffee pots and electric foot heaters will there be?
.
I have no advice on what you should do other than to determine your actual load.
Here is my load. Two computers, one 20 yrs old, one 10 yrs old; one 20" LCD (about 30 W), and one 20" CRT. Total for all of these is 388 VA, 271 W, and 3.38 A. Calculated PF = 0.7 . My average per computer is 388/2 = 194 VA.
High performance computers will require more power than mine, but the use of LCDs will be less power than my CRT. The CRT with a mostly white screen is 95 W and 140 VA. The LCD is 27 W and 44 VA.
Is the lighting load on the same transformer? How many coffee pots and electric foot heaters will there be?
.
dbuckley
Senior Member
- Location
- Canterbury, New Zealand
If your load is less than 50% of the transformer rating, then you are unliklely to need a K rated tranny.
A K rated transformer is designed to handle the additional heating effects from a large neutral current caused by ugly harmonics. Many folks just specify a oversized tranny rather than a correctly K rated transformer.
A K rated transformer is designed to handle the additional heating effects from a large neutral current caused by ugly harmonics. Many folks just specify a oversized tranny rather than a correctly K rated transformer.
Dnkldorf
Senior Member
Been here, done that with K rated transformers.
They are boat anchors.
Use a zig-zag. Much better.
Harmonics will cause a normal transformer, with a 30% load, to operate above it's temp rating. This causes extreme heat coming out of the transformer. You can fry eggs on them.
Your A/C unit may not keep up, if not sized for this amount of heat.
Zig-zags are the way to go.
IMO
They are boat anchors.
Use a zig-zag. Much better.
Harmonics will cause a normal transformer, with a 30% load, to operate above it's temp rating. This causes extreme heat coming out of the transformer. You can fry eggs on them.
Your A/C unit may not keep up, if not sized for this amount of heat.
Zig-zags are the way to go.
IMO
ohmhead
Senior Member
- Location
- ORLANDO FLA
Well its all about phase shift with a zig zag and yes it kinda eats the harmmonics with a 30 deg phase shift but who will pay the money for this ?
Its still there you can buy a k transformer and there nothing more than a electro static shield around each phase and bigger windings meaning conductor phase windings thats for the heat but does not limit the harmonics it just carrys the load better .
Back feed in a branch circuit do to computers on line spikes and voltage changes are always present its called todays electrical .
There are filters just like in the old days install a bigg bad MOV on each branch to limit the spikes to that circuit but its still there on that line .
TVSS to me are just MOV S they control the max voltage spike but the spikes are still on line and its money wasted because your computer has its own inside so you can spend lots of money on nothing .
I think you can limit harmonics to your service backfeed but its still on the branch circuit so it needs to be a the source of the computer at that point .
Bill needs to solve this for us
Its still there you can buy a k transformer and there nothing more than a electro static shield around each phase and bigger windings meaning conductor phase windings thats for the heat but does not limit the harmonics it just carrys the load better .
Back feed in a branch circuit do to computers on line spikes and voltage changes are always present its called todays electrical .
There are filters just like in the old days install a bigg bad MOV on each branch to limit the spikes to that circuit but its still there on that line .
TVSS to me are just MOV S they control the max voltage spike but the spikes are still on line and its money wasted because your computer has its own inside so you can spend lots of money on nothing .
I think you can limit harmonics to your service backfeed but its still on the branch circuit so it needs to be a the source of the computer at that point .
Bill needs to solve this for us
brian john
Senior Member
- Location
- Leesburg, VA
So if I have any level harmonics I should spend the extra money on a zig zag transformer, kind of a broad statement. Do you sell transformers?
Dnkldorf
Senior Member
Any level harmonics?......I wouldn't say that. (1) Fluro lay-in will cause harmonics, but I wouldn't go spend $9K on a zig zag. I wouldn't spend $4K on a K one either.
But if transformers start getting hot, and you confirm you a harmonic problem, then yup, I would go with a zig zag.
And yup, I sell transformers. I install them, rip them out and move them too.
brian john
Senior Member
- Location
- Leesburg, VA
I just think if you are going to make such a broad statement you should offer some back up for your thoughts. I also sell and install equipment, but before I recommended an up-sell I would have some backup for my recommendations.
We don't often have that opportunity when we install new equipment. We usually have to provide a harmonic analysis with the bid and demonstrate compliance post installation.
That said, my main filed of activity is the variable speed drives business and harmonics has long been known to be a characteristic of such systems - I have a copy of a technical paper from 1948 dealing with this issue.
For commercial and residential installations, the harmonics issue has crept in by degrees and sometimes, probably more often than not, it isn't considered until it sneaks up on someone and bites them in the backside.
I have seen this happen on a number of occasions, sometimes within our own organisation. We have a lighting division. It's architectural lighting and the application is usually commercial or up-market residences and some of the installations are fairly extensive. And that's when it can be a problem. The products themselves comply with required standards for individual units. But put enough of them in the same place and there can be an unacceptable level of harmonics. Then it becomes a very messy situation both technically and commercially.
brian john
Senior Member
- Location
- Leesburg, VA
I have often thought about when we all get rectifiers for our electric cars, and VFD for furnaces and AC units.
Maybe the manufactures will think about this before we all change over?
- Location
- Wisconsin
- Occupation
- PE (Retired) - Power Systems
I did not say that harmonics were not always an issue.
K-rated transformers are designed to handle only the overheating that a 100% loaded unit might experience based on a general amount of harmonics - this is a gimmick. The casual way they are typically specified indicates the specifier has no idea what a K-rating means.
It is not possible/practicable to predict the harmonic profile of an office or an industrial plant. This is partly due to the mitigation that will occur in the 'as installed' conductor impedance, and partly from the mutual cancellation of switched mode power supplies. Dealing with a single harmonic producing device, like a VFD or a UPS front end, is a different matter. Properly written standards acknowledge the issue of 'harmonics' at a "point of common coupling".
Maybe, but I don't think it will be voluntary.
You have a point and I have no disagreement with it. In my experience, harmonics and related problems, potential or otherwise, is seen as something of a black art by many.
Well, given that I regularly have to do this and commit to figures, I respectfully disagree with you on that point.
Getting it wrong would be bad and expensive.
Yes, source impedance from the PCC to point of use does have a bearing on harmonic current magnitudes and consequent harmonic voltage distortion.
Source impedance is usually known or can be calculated with a sufficient degree of accuracy.
You have mentioned this a couple of times.
The input for most switched mode PSUs is a plain rectifier bridge.
So no input cancellation of harmonics.
Again I agree.
I generally deal with systems with a multiplicity of harmonic producing devices.
- Location
- Wisconsin
- Occupation
- PE (Retired) - Power Systems
Do you really model an entire facility before bid time? Do you really know the behavior of all the other harmonic sources prior to them actually being purchased much less installed?
When I bid the engineering portion of a study I know I will not have any control over the conductors, raceways, and lengths that will be finally installed. One of the last projects I was involved in was switched from aluminum to copper conductors after the original approval process had been completed.
Yes the input is a simple rectifier just like most VFDs have. The harmonics result from the 'pulses' of current drawn based on the load requirements Because not all supplies will be drawing equal amounts of currents at all times there is a probability of cancelation of the harmonics. This is one reason IEEE discusses 'mitigating' harmonics at the point of common coupling rather than at each harmonic source. Cancelation like this would be more common with 1-phase sources (i.e. office computers) rather than 3-phase loads (i.e. VFDs).
- Status