Transformer sizing for step up and step down transformers

[PVfarmer]

Can't you get 25kVA of 120V (or 208A) from a single 1ph 25kVA xfmr LV side, but to get the same 208A/25kVA/120V from a 240/120V high leg setup 3ph xfmr, you'd need a 75kVA xfmr, as it's only using 1 of the 3 windings?

I agree with the other posters here in that your lack of understanding of the fundamentals and your misuse of the terminology are real problems. I appreciate your curiosity and willingness to learn, but wwhitney and others have given excellent examples and explanations, which you seem to ignore. Focusing on the kva rating of the poco transformer is a waste of time for conventional power flow. (I grant that in PV applications it can matter.)

Thanks for the comment. Could you please explain how I misused terms in the comment above?
It seems like I'm using different terms to describe the same thing as other people, but they don't quite like it.

I understand your example with the house. The point I'm trying to get across is that yep, that's the way standard *residential* services work, but when you get into service upgrades, the POCO asks you to fill out the load sheet so they can size the xfmr accurately (whatever size that may be)
That's all.

About my quote above:
Let's forget about POCOs and PV for a minute. There's a 480V delta service and 480V delta inverters, amps of both are "more than enough".
So a new building goes up and the owner of the property and PV wants to power it.
It's a storage barn and the total equip. load in the building is eight 120V fans, 25A each.
That's 24kW of load.
You could supply that load from:
A. A single 480V to 120V 1ph transformer connected to 480V L1 and L2 to the xfmr H1 and H2, load connected to xfmr X1 (L1) and X2 (neutral/bonded to load panel).

or

B. A single 3 phase 480V delta to 240V delta/120V center tapped xfmr.

Xfmr A would be 25kVA. Xfmr B would have to be 75kVA.

Is that ok so far?

 

[PVfarmer]

System I have been working on the last few days is actually a near textbook example: 1000 amp 120/208 service (ampacity of service entrance conductors), 150 KVA transformer (~420 amps). Actual load hovers right around 115 KVA (~320 amps) during peak times, Peak demand from last two years is 156 KVA

I'm honestly not trying to argue here at all- just trying to return the favor of you answering my questions.

You mean 3 phase 208/120V wye service.
Looks to me like you're speaking of the 15th one down in the chart on pg. 71 of this ESB 750. -->https://www.nationalgridus.com/non_html/shared_construction_greenbk.pdf


1* Account: Commercial // Voltage:208/120 // Phases:3 // Amps:400 // Wires: 4 // Meter Terminals:7 // Sequence:Hot // Manual Bypass required:Yes // Diagram:7.3-6
*=1 A 400A service with a class 320 meter and socket is limited to 320 continuous amperes load capacity -
See Table 7.2-4 for rating requirements of service.

I'm thinking if the measured load is ~320A, and 320A is a "limit", and PV is "continuous"... 320A is also your limit for PV?

Does that seem like a factor to you?

edit: It seems like that 150kVA xfmr is sized just about right- it's running at 50-75% most of the time, hardly ever over 100%.

 

[wwhitney]

But the amps of the main switch doesn't always equal the amps of the service, so next I'd ask "what size is the xfmr"

This is one of your misuses of terminology. The "amps of the main switch" is the service size, by definition. The size of the POCO transformer is a separate matter.

2 I'm trying to say a 50kVA xfmr wouldn't ever be supplying a 200A service, because it would be oversized 2x, and the POCO doesn't do that.

I agree with you. But to be clear, 200A at 240V = 48kVA, so you've just accepted that the POCO will never provide a transformer that matches the nominal service size. Great!

If it was 600A of heaters running 50% of the time, 75kVA would be the one. That's why I'm saying 75kVA would Be "standard" for someone using 600A on a regular basis.

That may be. But somebody using 600A on a regular basis wouldn't have a 600A service, they'd have at least an 800A service.

And why is there a 1000A service when the load never gets close to 500A?

Because service size means the size of the main disconnect and service entrance conductors, which are typically sized via the NEC, which is very conservative.

Cheers, Wayne

 

[electrofelon]

1 Sure, but the service is provided by the xfmr, the same xfmr the PV goes thru, so the size of it has to be taken into account at some point- why not at the first point?
If a customer says "hey, I want some PV", you don't run out and buy inverters without asking "what voltage service do you have?"

I dont think anyone is saying to ignore what the service voltage is

2 I'm trying to say a 50kVA xfmr wouldn't ever be supplying a 200A service, because it would be oversized 2x, and the POCO doesn't do that.

It could, like iron ben said, it could and often does serve multiple houses/buildings. In some cases it may be super small or it could be larger depending on how many buildings it serves and what the loads are. Ive seen the utility "steal" off two legs of a 120/208 three phase bank that was serving some commercial occupancies, to serve some dwellings.

4 Doesn't the NEC end at the service point? So it's not factoring in the service xfmr at all?

No

You mean 3 phase 208/120V wye service.
Looks to me like you're speaking of the 15th one down in the chart on pg. 71 of this ESB 750. -->https://www.nationalgridus.com/non_h...on_greenbk.pdf

Actually the job I mentioned is in Seattle city light territory. There are other utilities out there ya now

I'm thinking if the measured load is ~320A, and 320A is a "limit", and PV is "continuous"... 320A is also your limit for PV?

Does that seem like a factor to you?

Sure. Generally you would go by what the "weakest link" is in the system. There are situations where it gets more complicated to figure out what the service rating is, such as when you have multiple service disconnects. You could have the ampacity of the service entrance conductors to be less than even one of the multiple service disconnects. OR you could have that class 320 meter socket with hefty conductors for some reason and multiple or one large service disconnect, but you would be stuck with that 320 amp continuous limit.

 

[kwired]

Could you elaborate? Is that single phase or three ph 120/208V? And why is there a 1000A service when the load never gets close to 500A?

I have only skimmed over the posts since my last post so I'm not certain what was all said but this caught my attention and I wanted to reply to it.

There are many services out there that are oversized for the actual load. Most are intentional to allow future addition. POCO still generally only provide the size of transformer needed (from their calculation methods) to supply the current load. I have several times had to go to POCO to get an increase in transformer size because we added load to a facility but never upgraded service equipment or service conductors. Actually it was more like contact them and tell them we are adding load and let them decide if and how much increase in transformer size they wish to install.

Now if you are adding PV, that changes things a bit. If you were only feeding a high KVA into the grid for a short period it really wouldn't be much different then having a motor that only draws a high demand for a short period - they may leave a somewhat undersized transformer in place, but PV is putting a fairly continuous VA into the system for several hours at a time and that transformer is seeing different heating effects with such a load for that duration compared to heavy loading for short duration with sufficient cooling off periods.

 

[PVfarmer]

1
But the amps of the main switch doesn't always equal the amps of the service, so next I'd ask "what size is the xfmr"
This is one of your misuses of terminology. The "amps of the main switch" is the service size, by definition. The size of the POCO transformer is a separate matter.
2
I'm trying to say a 50kVA xfmr wouldn't ever be supplying a 200A service, because it would be oversized 2x, and the POCO doesn't do that.
I agree with you. But to be clear, 200A at 240V = 48kVA, so you've just accepted that the POCO will never provide a transformer that matches the nominal service size. Great!
3
If it was 600A of heaters running 50% of the time, 75kVA would be the one. That's why I'm saying 75kVA would Be "standard" for someone using 600A on a regular basis.

That may be. But somebody using 600A on a regular basis wouldn't have a 600A service, they'd have at least an 800A service.
4
And why is there a 1000A service when the load never gets close to 500A?
Because service size means the size of the main disconnect and service entrance conductors, which are typically sized via the NEC, which is very conservative

1 ok- The rated amperage of the main switch doesn't equal the amps of the service xfmr, ever!
Service xfmr = 150kVA.
Available amps at 100% xfmr rating = 180.6A

| | |-----480/277V service lines (4 total)----->>>> 225A main fused service disconnect switch----->>>200A meter-----^^^^----->>>>225A- customer's load/distribution panel.

The | | | are the three grid wires providing 3ph service. The ^^^^ is where the PV is connected, from its' own fused disco.
Before the PV, the loads were 225A max, so a 225A panel and switch.
You can add 200A of PV (breaker) line side (load side of meter), but with the 225A load panel, that 225A main fused switch has to become a 400A fused switch.


That service, with that load, without PV would most commonly be a 150kVA pad mount. Because it fits- it's also a "200 amp service", not a 180.6A service.
---
2 No- the kVA ratings of any possible xfmr or combination thereof divided by any normal voltage will *never* come out to either 100, 200 or 400A, which are standard services offered...why would they offer a 208.6A service?
How about this- when the POCO says "200A service", they mean "in the ballpark of 200A". The continuous will nearly always be less than max load (maybe a data center they'd be close to =?), so the continuous kVA load is always less than the service xfmr kVA rating, while the max might be over the kVA rating. Better for the max to be above xfmr kVA rating than below.

3 So your #3 aboce and my #2 above are basically in agreement?

4 He didn't say what the main switch was. 1000 amp 120/208 service (ampacity of service entrance conductors)

I'm not gonna guess on the main switch there because someone will get upset. But aren't conductors/switch the same as service xfmr/main switch, and never the same? You size the conductors based on the loads before the main switch is what I mean.

 

[PVfarmer]

1
It could, like iron ben said, it could and often does serve multiple houses/buildings. In some cases it may be super small or it could be larger depending on how many buildings it serves and what the loads are. Ive seen the utility "steal" off two legs of a 120/208 three phase bank that was serving some commercial occupancies, to serve some dwellings.

2
Actually the job I mentioned is in Seattle city light territory. There are other utilities out there ya now

3
Sure. Generally you would go by what the "weakest link" is in the system. There are situations where it gets more complicated to figure out what the service rating is, such as when you have multiple service disconnects. You could have the ampacity of the service entrance conductors to be less than even one of the multiple service disconnects. OR you could have that class 320 meter socket with hefty conductors for some reason and multiple or one large service disconnect, but you would be stuck with that 320 amp continuous limit.

1 My take is that the POCO is thinking that undersized xfmrs is the way to go, as long as they've computed that it won't cause issues (with the xfmr itself as well as the grid overall). Oversized xfmrs will of course always work, but no reason to oversize if smaller will work "nearly equally" as well.
And the higher the amperage of the service loads, the more undersizing matters.

2 Ok, well I'm still unclear as to if you meant 1ph or 3oh- if it is 1ph, nothing I've said makes any sense at all!:huh:

3 So we've got:
A service rating (meaning xfmr kVA?)
B service conductor rating
C main service disconnect switch/breaker (fuse) rating
D the "service amperage" you are "ordering" from the POCO, meaning 100, 200 and 400 "standard".

Right. No wonder it's confusing!!!

But out of those...C and D would most often be "the same".

 

[PVfarmer]

I have only skimmed over the posts since my last post so I'm not certain what was all said but this caught my attention and I wanted to reply to it.

me (PVF)-
Could you elaborate? Is that single phase or three ph 120/208V? And why is there a 1000A service when the load never gets close to 500A?

1
There are many services out there that are oversized for the actual load. Most are intentional to allow future addition.

2
Now if you are adding PV, that changes things a bit. If you were only feeding a high KVA into the grid for a short period it really wouldn't be much different then having a motor that only draws a high demand for a short period - they may leave a somewhat undersized transformer in place, but PV is putting a fairly continuous VA into the system for several hours at a time and that transformer is seeing different heating effects with such a load for that duration compared to heavy loading for short duration with sufficient cooling off periods.

1 He said "1000A service conductors", not sure if he meant "the service from the POCO catalog" also, but you're talking "service xfmrs".

*IF* it is a "400A of 208/120V wye service" (according to POCO), I can see why the conductors would be 1000A- 400A * 208V * 1.73 = 144kVA
144kVA also = 1200A at 120V.

if someone lost their mind and put 120,000w of 120V bulbs on a single panel, those 1000A rated conductors would be quite handy...
---
2 I agree- if the PV output is going to be "maxed" for over 3 hours a day (it probably/almost always is?), and you knew the xfmr you were getting for loads would run a lot also, and that the xfmr was most efficient when loads = 50% of xfmr rating...you'd multiply the max load by 2, and get that size xfmr. If you owned the xfmr that is.

 

[wwhitney]

Service xfmr = 150kVA.
Available amps at 100% xfmr rating = 180.6A

| | |-----480/277V service lines (4 total)----->>>> 225A main fused service disconnect switch----->>>200A meter-----^^^^----->>>>225A- customer's load/distribution panel.

The | | | are the three grid wires providing 3ph service. The ^^^^ is where the PV is connected, from its' own fused disco.
Before the PV, the loads were 225A max, so a 225A panel and switch.

You can add 200A of PV (breaker) line side (load side of meter), but with the 225A load panel, that 225A main fused switch has to become a 400A fused switch.

There's no need to change the main fused switch, the currents don't add that way.

The current through the main switch will be (Customer Demand - PV Production), where if that comes out negative the current is goes out to the grid from the PV. If the "Customer Demand" is between 0 and 225 amps, and "PV Prodution" is between 0 and 200 amps, then the current through the main switch will be between -200 and 225. The disconnect switch doesn't need changing.

That service, with that load, without PV would most commonly be a 150kVA pad mount. Because it fits- it's also a "200 amp service", not a 180.6A service.

No, 150kVA is too big a POCO transformer for that service, unless the customer is actually using the full capabilities of the service. That would be unusual. The POCO will determine the transformer size based on the actual usage pattern.

How about this- when the POCO says "200A service", they mean "in the ballpark of 200A".

No, they mean a maximum of 200A at any time. The average could be anywhere from 0A to 160A. Because of NEC rules on continuous loads, the average should never exceed 80% of the nominal rating. So let me revise something I said earlier, if you take the POCO transformer size in kVA, and divide by the service size in kVA, you'll get a number from 10% to 80%. Very, very, rarely will you get the full 80%.

BTW, this a source of possible confusion: POCO amp ratings are continuous ratings, while NEC amp ratings are non-continuous ratings. So to compare you have to multiply the NEC rating by 80%. That is why a 320A meter (a POCO rating, continuous) is used for a 400A service (an NEC rating, non continuous).

Cheers, Wayne

 

[wwhitney]

if the PV output is going to be "maxed" for over 3 hours a day (it probably/almost always is?), and you knew the xfmr you were getting for loads would run a lot also, and that the xfmr was most efficient when loads = 50% of xfmr rating...you'd multiply the max load by 2, and get that size xfmr. If you owned the xfmr that is.

I'd do what was most economical. The extra cost of the larger transformer could well exceed the lifetime savings you'd get from the increased efficiency. So you have to calculate a net present value for each scenario and pick the one that is most advantageous.

Cheers, Wayne

 

[wwhitney]

3 So we've got:
A service rating (meaning xfmr kVA?)

Nobody uses that terminology. It would just be service transformer size or rating.

B service conductor rating
C main service disconnect switch/breaker (fuse) rating
D the "service amperage" you are "ordering" from the POCO, meaning 100, 200 and 400 "standard".

Right. No wonder it's confusing!!!

For a service with a single main breaker, there is just one service size. It is the rating of the main breaker, your (C). That's the terminology. After the service point, the NEC will require that conductors protected only by that main breaker will be sized according to that main breaker rating.

The equipment under the POCO's control is their responsibility. Their transformer size and conductor size up before the service point will be smaller than (C). In fact for the conductors before the service point, the POCO sizing rules and the NEC sizing rules are different for a given ampacity.

Cheers, Wayne

 

[don_resqcapt19]

...
I'm saying with 400A/50kVA you can run 399A of 240V motors (and one 120w 120V lightbulb) and overload the xfmr to 200%, but you really shouldn't. Sure, you could run 800 120w 120V bulbs.
...

I don't see how you think that you can supply 96kW of load from a 50kVA transformer.

 

[PVfarmer]

I don't see how you think that you can supply 96kW of load from a 50kVA transformer.

I'm not saying I'd DO it, just that it is allowed- 200% for 30 minutes is the rule.
https://books.google.com/books?id=B...=overloading transformers 200% ansi&f=false

 

[don_resqcapt19]

I'm not saying I'd DO it, just that it is allowed- 200% for 30 minutes is the rule.
https://books.google.com/books?id=B...=overloading transformers 200% ansi&f=false

It is not permitted under the rules of the NEC.

 

[PVfarmer]

1
Nobody uses that terminology. It would just be service transformer size or rating.
2
For a service with a single main breaker, there is just one service size. It is the rating of the main breaker, your (C). That's the terminology.

1
The words at the top of the amp column in the POCO booklet (for selecting a voltage and service) are "Service Size (Amps)". Isn't that really just the meter rating?
And the setups to supply the (Amps) of service are called "transformer installations".

2
I have to dispute that.
Say there are 240V motor loads of 125A, not often at all and additionally there are under 125A of loads most of the rest of the time. 125 * 240 = 30kW peak.
Also, 125A * 120V = 15kVA.
There is a 200A panel for the 240 motors, and another 200A panel for the other 125A of loads.

The main breaker will be 400A, while the service size will be 200A.

 

[PVfarmer]

It is not permitted under the rules of the NEC.

Sure, you can't do that with a xfmr that you own- but the POCO side isn't NEC, so I'm saying it can happen, and I'd suppose the POCO uses the 200% thing in their calculations, if they have to for some reason.

 

[wwhitney]

1
The main breaker will be 400A, while the service size will be 200A.

If the main breaker is 400A, the service size is 400A. Period. That's the definition. The conductors from the service point to the main breaker will have an NEC ampacity of 400A.

Now the POCO may choose to provide that 120/240V service with a 50 kVA transformer, and they may provide conductors from the transformer to the service point that the NEC would say are only good for 200A. But that's the POCO's business. If the fact they have done that ever causes you trouble, then you have to talk to the POCO. E.g. if you want to add more than 50kVA of PV in this example.

Cheers, Wayne

 

[electrofelon]

It is not permitted under the rules of the NEC.

But is it Don? As you know, the best violation you could find is 110.3(A) and (B) which are fairly general and it could be argued that they leave the door open for intermittently exceeding the name plate, doing an operating temperature analysis, accounting for low ambient temp, etc. I know that isnt really common practice for NEC installations, just sayin'

 

[electrofelon]

He didn't say what the main switch was. 1000 amp 120/208 service (ampacity of service entrance conductors)

There was not a single main switch. That service has 6 service disconnects, so you go by the ampacity of the service entrance conductors

 

[PVfarmer]

If the main breaker is 400A, the service size is 400A. Period. That's the definition. The conductors from the service point to the main breaker will have an NEC ampacity of 400A.

Now the POCO may choose to provide that 120/240V service with a 50 kVA transformer, and they may provide conductors from the transformer to the service point that the NEC would say are only good for 200A. But that's the POCO's business. If the fact they have done that ever causes you trouble, then you have to talk to the POCO. E.g. if you want to add more than 50kVA of PV in this example.

I'm thinking that the POCO did do that in one case I know of, and it didn't cause any trouble.
Skip the whole PV thing for a minute- a farm here, before PV, had a 25kVA xfmr, (200A service, 120V * 200A = 24kVA).
A list of the 240V motors on premises, before even getting to the bigger ones, hit 125A of 240V loads (or 30kVA). That's without the 3hp and 5hp motors!

The main switch was 400A, because well... I don't know. I'd like to say there were 2 200A panels on that switch, but it was setup originally in the 1960s and modified by 4 different electricians since then.
Are you telling me the whole thing was a violation that was let slide for 20+ years?

I thought it was just a 200A service with a 400A main fuse because it only drew over 200A intermittently.

Either way, I was very happy when the engineer saw it and said, to paraphrase efelon, "yeah, you're gonna want to throw that conglomeration of panels in the river".

This makes sense to me- a 200A service means you have a service xfmr which delivers ~200A at 120V when running at near 100%.
Sure, you can also draw 200A for 240V loads from that- hence the 200% overload allowance.
If you were drawing 200A for 240V loads all the time, you'd want a 400A service.