Transformers/Conductor Sizing/OCPD/the usual

[sinforoso]

Hello everyone,

My apologies if this question appears indicative of incompetence on my part, but for some reason I cannot wrap my head around the relevant code sections involving transformer & conductor OCPD--or the appropriate sequence of steps one must take, if any--despite having scoured the NEC and the Internet to try to dig up as much information on this as possible. So, here's my problem:

I have a 300KVA, 480V delta primary/208V wye secondary transformer, and I want to make sure that everything's installed and adequately protected per Code (as does everyone ). This installation is for a data center renovation, so I'm assuming load continuity for worst-case scenario/nonlinear loads. Below is a rough quasi-flowchart of the current design (using copper conductors):

500A brkr -> (2) parallel runs of (3)250MCM & (1)#1G -> 300KVA xfmr -> 1,200A fused disconnect -> (4) parallel runs of (4)400MCM & (1)#3/0G -> 1,200A MCB 208Y/120V panelboard

(bolded text represents "equipment," normal text represents conductors)

Reasoning for the above design:

• Xfmr Primary:
  • 300KVA @ 480V = 361A, therefore OCPD should be at least 125% of that, therefore 361A becomes 451A, therefore next size up is 500A.
  • Only 3 CCCs, therefore no derating, therefore (2) runs of 250MCM per the 75^o table gives 255A * 2 = 510A = sufficient conductor ampacity to withstand maximum possible current going through breaker.
• Xfmr Secondary:
  • 300KVA @ 208V = 833A, but because of harmonics and assumed continuous load, I assumed 125% of that, therefore 833A becomes 1041A, therefore--though the "next size up" rule does not apply to xfmr secondaries or to loads > 800A--the only "logical" choice for conductor OCPD from the way I'm interpreting the Code is a 1200A fused disconnect.
  • 4 CCCs, therefore conductor ampacity must be derated to 80% (from the 90^o table), therefore (4) runs of 400MCM * 0.8 = 380A * 4 * 0.8 = 1,216A = sufficient conductor ampacity to withstand maximum possible current going through disconnect = sufficient conductor ampacity to feed 1200A MCB panelboard downstream.

Point being, the above setup to me makes sense at first glance, but I have reason to suspect I messed up somewhere, as this design in theory allows the transformer to "see" currents continually above its rating without triggering any OCPDs. Realistically, the equipment upstream from the transformer will prevent this from happening, but something about this seems weird and I can't quite put my finger on it. As far as I can tell, this design meets the requirements of 450.3(B) for primary-only protection, so I guess it's possible I've just confused myself unnecessarily?

Anyway, I figured I'd check with you Code wizards out there and get some feedback on this . My apologies for the wall of text.

Thanks.

 

[augie47]

At first review, I see no problem.

 

[GoldDigger]

The upstream breakers will protect the transformer and both primary and secondary conductors (because the primary is delta?) against short circuit and ground fault.
On the assumption that a wiring fault will not, for long anyway, be current limited, the downstream protection will be good enough for long term low level overload.

Tapatalk!

 

[sinforoso]

Well, I'm not thinking of this so much in terms of fault currents, but rather plain old "regular" currents that could potentially reach the transformer.

As in, all other things equal, theoretically I could disconnect everything upstream from the 500A breaker and hook it up to a generator that delivers, say, 499A continuously (assuming the breaker is 100% rated). The breaker wouldn't trip, but there would still be 499A flowing to the primary windings even though they're only "rated" for 361A (300KVA @ 480V). Everything downstream from the transformer is also more than capable of handling the resulting secondary current, but the secondary windings themselves would be subjected to currents that exceed their ratings too. So, this scheme could conceivably go on forever so long as the transformer remains operational.

I guess what's confusing me is, what would be keeping the transformer operational in this situation? Wouldn't the transformer degrade rapidly if continuously subjected to currents above its rating?

Yes, this is quite possibly a newbie question, but I'm curious nonetheless. Thanks.

 

[jumper]

Yes, this is quite possibly a newbie question

Your questions are perfectly valid and quite good to post here.

Welcome to the forum.

 

[david luchini]

As in, all other things equal, theoretically I could disconnect everything upstream from the 500A breaker and hook it up to a generator that delivers, say, 499A continuously (assuming the breaker is 100% rated). The breaker wouldn't trip, but there would still be 499A flowing to the primary windings even though they're only "rated" for 361A (300KVA @ 480V). Everything downstream from the transformer is also more than capable of handling the resulting secondary current, but the secondary windings themselves would be subjected to currents that exceed their ratings too. So, this scheme could conceivably go on forever so long as the transformer remains operational.

I guess what's confusing me is, what would be keeping the transformer operational in this situation? Wouldn't the transformer degrade rapidly if continuously subjected to currents above its rating?

Yes, subjecting a transformer to currents above its rating will reduce the transformer's lifetime. But, just hooking up a generator that can deliver 499A to the primary will not cause 499A to flow in the primary. The current flow in the primary will be based on the load connected on the secondary. If you had 250kVA of load connected on the secondary, then the primary will see 300A even though it is connected to the 499A generator through the 500A breaker. To prevent reducing the transformer life, don't connect more load than it is rated for.

Regarding your feeders, you exceed code requirements, so you could probably save some money by reducing some of the sizes. For instance the grounds could be...
(2) parallel runs of (3)250MCM & (1)#2G and (4) parallel runs of (4)400MCM & (1)#1/0G
There is no requirement for the primary conductors to for the whole 500A, so you could probably reduce them to 4/0.

I would also consider reducing the secondary to 1000A instead of 1200A.

 

[petersonra]

I think you may be mixing up conductor sizing requirements and OCPD requirements.

I think you are Ok as is but I think how you got there might be a little screwy.

I am not a fan of 1200A fuses. probably more cost effective and simpler to use a CB.

 

[sinforoso]

Yes, subjecting a transformer to currents above its rating will reduce the transformer's lifetime. But, just hooking up a generator that can deliver 499A to the primary will not cause 499A to flow in the primary. The current flow in the primary will be based on the load connected on the secondary. If you had 250kVA of load connected on the secondary, then the primary will see 300A even though it is connected to the 499A generator through the 500A breaker. To prevent reducing the transformer life, don't connect more load than it is rated for.

Regarding your feeders, you exceed code requirements, so you could probably save some money by reducing some of the sizes. For instance the grounds could be...
(2) parallel runs of (3)250MCM & (1)#2G and (4) parallel runs of (4)400MCM & (1)#1/0G
There is no requirement for the primary conductors to for the whole 500A, so you could probably reduce them to 4/0.

I would also consider reducing the secondary to 1000A instead of 1200A.

Well, I was merely assuming a worst-case scenario with the 499A generator analogy. Even if we looked at it from the load on the secondary, though, I believe I'd run into the same problem. Say I loaded the 1,200A panel to 80% of rating, aka 960A. The transformer secondary windings are presumably only "rated" for 833A (300KVA @ 208V), so the windings would still be taking a beating, and yet my design is apparently code-compliant based on the comments I'm reading here.

If an electrician is adding load to the 1,200A panelboard all willy-nilly, I somehow doubt he's gonna bother to notice if the upstream transformer's being unduly stressed. So, I guess my question now is, why would this design pass muster in the first place?

As for ground sizes, I based those on Table 250.122 (though it appears you are correct on the primary conductors, aka I messed that one up). As for the secondary conductors, what's the rationale for #1/0? Isn't the "Automatic Overcurrent Device in Circuit Ahead of Equipment, Conduit, etc., Not Exceeding (Amperes)" the 1,200A panelboard, thus requiring #3/0 at a minimum? (even if I reduced it to 1000A, the Table still requires #2/0)

Lastly, as for "There is no requirement for the primary conductors to for the whole 500A, so you could probably reduce them to 4/0," that's another one of the things I'm never too clear on either. Isn't it (or shouldn't it be) the case that the 500A breaker cannot protect #4/0 conductors because the 500A breaker can "accept" currents greater than what the conductors can handle?

p.s. I am that rare breed of engineer that cannot be concise even if his life depended on it, so once again, my apologies for the verbosity.

 

[david luchini]

Well, I was merely assuming a worst-case scenario with the 499A generator analogy. Even if we looked at it from the load on the secondary, though, I believe I'd run into the same problem. Say I loaded the 1,200A panel to 80% of rating, aka 960A. The transformer secondary windings are presumably only "rated" for 833A (300KVA @ 208V), so the windings would still be taking a beating, and yet my design is apparently code-compliant based on the comments I'm reading here.

If an electrician is adding load to the 1,200A panelboard all willy-nilly, I somehow doubt he's gonna bother to notice if the upstream transformer's being unduly stressed. So, I guess my question now is, why would this design pass muster in the first place?

Yes, that may be code compliant, but that doesn't make it good design. If you have a 300kVA transformer and are concerned about reducing the transformer life, don't load it beyond 300kVA. No one should be adding load willy-nilly. Just as you have the responsibility to make the original installation within code requirements, someone else would have the responsibility to make any revisions within code requirements. But if they did add loads willy-nilly, they wouldn't necessarily be creating an unsafe condition, they would just reduce the transformer life, as you noted. Of course, that would be another reason for going to 1000A instead of 1200A.

As for ground sizes, I based those on Table 250.122 (though it appears you are correct on the primary conductors, aka I messed that one up). As for the secondary conductors, what's the rationale for #1/0? Isn't the "Automatic Overcurrent Device in Circuit Ahead of Equipment, Conduit, etc., Not Exceeding (Amperes)" the 1,200A panelboard, thus requiring #3/0 at a minimum? (even if I reduced it to 1000A, the Table still requires #2/0)

The "ground" on the secondary size isn't an equipment grounding conductor, it is an equipment bonding jumper for the separately derived system. It should be sized per 250.30(A)(2), which directs you to 250.102(C), which directs you to 250.66. Per table 250.66, you need a 1/0 for 400mcm phase conductors.

Lastly, as for "There is no requirement for the primary conductors to for the whole 500A, so you could probably reduce them to 4/0," that's another one of the things I'm never too clear on either. Isn't it (or shouldn't it be) the case that the 500A breaker cannot protect #4/0 conductors because the 500A breaker can "accept" currents greater than what the conductors can handle?

p.s. I am that rare breed of engineer that cannot be concise even if his life depended on it, so once again, my apologies for the verbosity.

With regards to the 2 sets of #4/0 on the 500A c/b, 240.4(B) permits the next standard size up OCPD to protect conductors as long as the next standard size up does not exceed 800A (you referenced the "next size up" rule in your original post.) Two sets of 4/0 has an ampacity of 460. The next standard size up from that is 500A.

 

[kingpb]

Couple things:

First; the NEC does NOT mean a good design. This cannot be stressed enough.

Second; a breaker set at 500A, like you propose is shown. The breaker shown does have LSIG in order to meet the inrush requirements. A standard breaker without an adjustable Instantaneous setting may not work. The breaker works fairly well, but slight overloads for a longer period of time could still have an accumulating effect on transformer health. Again, this is a device with LSIG so depending on which breaker you choose, the characteristic will be different.

You have to understand that on a solidly grounded delta-wye transformer the through fault curves have to be adjusted by 58% for proper selection of HV side protection to account for unbalanced secondary faults. SqD has a good paper that pertains to this subject.

You are correct in that the transformer could be continuously overloaded. This is often overlooked because people base the protection on the maximum NEC allowable. But, maximum doesn?t mean good design. Your quandary is actually an engineering issue that is overlooked time and time again.

IMO - For a data center, I would probably select a 400KVA transformer and de-rate it a little to keep it cooler and set protection to keep it under the damage curve, at all times.

 

[sinforoso]

I appreciate all the responses so far; I don't mind taking a proverbial beating every once in a while if I can learn something valuable in the process, heh.

With regards to the 2 sets of #4/0 on the 500A c/b, 240.4(B) permits the next standard size up OCPD to protect conductors as long as the next standard size up does not exceed 800A (you referenced the "next size up" rule in your original post.) Two sets of 4/0 has an ampacity of 460. The next standard size up from that is 500A.

I might have misstated my previous question. Yes, I'm aware of the "next size up" rule, but I'm curious perhaps as to the rationale behind it altogether? As a few people have said so far, "code-compliant" is not necessarily synonymous with "well-designed." Similar to my transformer example, it is theoretically possible to create a situation that meets code and yet in which the conductors (with an ampacity of only 460A) are continually subjected to currents greater than 460A without tripping the breaker, at which point the breaker is effectively not doing its job.

Not trying to argue for the sake or arguing, btw. I'm genuinely curious (and eager to correct my errors in thinking).

 

[david luchini]

I appreciate all the responses so far; I don't mind taking a proverbial beating every once in a while if I can learn something valuable in the process, heh.

No beatings intended. We're all here to learn.:happyyes:

I might have misstated my previous question. Yes, I'm aware of the "next size up" rule, but I'm curious perhaps as to the rationale behind it altogether? As a few people have said so far, "code-compliant" is not necessarily synonymous with "well-designed." Similar to my transformer example, it is theoretically possible to create a situation that meets code and yet in which the conductors (with an ampacity of only 460A) are continually subjected to currents greater than 460A without tripping the breaker, at which point the breaker is effectively not doing its job.

If you have conductors with an ampacity of only 460A which are continually subjected to currents greater than 460A, then you have a Code Violation. That condition is not permissible under Code rules.

See, for instance, 215.2(A)(1).

Feeder conductors shall have an ampacity not less than required to supply the load as calculated in...Article 220.

In addition, the minimum conductor size shall have an ampacity which is not less than 100% of the non-continuous load plus 125% of the continuous load.

 

[sinforoso]

Touch?; thanks.

 

[Smart $]

The upstream breakers will protect the transformer and both primary and secondary conductors (because the primary is delta?) against short circuit and ground fault.
...

240.4(F) says conductors of 4-wire transformer secondaries shall not be considered to be protected by primary OCPD.

 

[Smart $]

...
500A brkr -> (2) parallel runs of (3)250MCM & (1)#1G -> 300KVA xfmr -> 1,200A fused disconnect -> (4) parallel runs of (4)400MCM & (1)#3/0G -> 1,200A MCB 208Y/120V panelboard
...

Didn't read all recommendations, but having two 1,200A OCPD's is not required. You can change panel to 1,200A MLO if so desired... providing it is not in another building or some other parameter which would require the MCB panelboard.

(OCPD rating as appropriate)

 

[kingpb]

What you may find is that since the breaker cannot be loaded to more than 80% of its rating, jumping to the next standard size will be less than an additional 25%, in which case if the conductors are sized per minimum NEC requirements, the ampacity, even at 75 deg C will still be within the rating.

Example: Current = 361A, so 125% is 451A. Cable has to handle 451A at 75 deg C, breaker can be upsized to 500A.
But, in accordance with code, neither will be loaded to more than 361A.

Even if breaker was loaded to 80% that is only 400A, and you should have 51A margin on cable at 75 deg C. Remember the cable can handle additional current and is rated up to 90 deg C, even though the terminations may be rated 75 deg C at the breaker. The breaker is going to trip on overload long before cable damage occurs.

 

[lielec11]

concise post but it looks good to me.

 

[infinity]

The "grounds" run with the secondary conductors would be supply side bonding jumpers (SSBJ) and as David stated sized according to 250.66 for the secondary conductors in each parallel raceway. The system bonding jumper would be sized according to 12.5% of the total kcmil size of the secondary conductors which in this case would be 4*400 kcmil=1600 kcmil * 12.5% = 200 kcmil or minimum of #4/0. {250.28(D)(1)}