OCP Question for an Autotransformer

[Smart $]

I don't think I can agree with the distinction you are making between connected and configured. Using you description of a delta connection, if I had a 240 Delta service and had a 208V, 3ph machine to install, and decided to use a 240 Delta-208/120V Wye transformer to stepdown to 208V for my machine, but only made 3 connections from the transformer to machine (120? out-of-phase relationship to each other,) I'd have a delta - delta connected transformer. I don't buy that.



I agree with your description of primary and secondary connections (in the case of an isolating transformer) but the problem, it seems to me, is that with the autotransformer one of your connections would be both primary and secondary...it is the same connection. I think this is where "primary" and "secondary" descriptions break down with autotransformers, and IMO, is why section 450.4 uses the descriptions of "input" and "output" for autotransformer, rather than primary and secondary.

I pretty much figured we would not see eye to eye on the matter

My main concern is conductor protection, as 450 covers xfmr protection. What NEC sections do you see as covering conductor protection of autotransformer-connected circuits?

For the sake of discussion, let use an example of two-wire in, two-wire out... and an abnormal condition to stress my point. Let's say we have 24kVA-rated autotransformer transitioning 120V in to 240V out. Max-rated in current is 200A. Max-rated out current is 100A. What ampacity conductors do you use, in and out, and what rating is your circuit ocpd?

 

[david luchini]

My main concern is conductor protection, as 450 covers xfmr protection. What NEC sections do you see as covering conductor protection of autotransformer-connected circuits?

I see Art. 240 as covering the conductor protection of an autotransformer-connected circuit. 240.21 says that overcurrent protection shall be provided in each ungrounded conductor at the point they receive their supply. Being that the autotranformer circuit is electrically continuous from input to output (vs. an isolating transformer where there is no electrical connection from primary to secondary,) I see the autotransformer input c/b as being the point where the circuit receives its supply.

For the sake of discussion, let use an example of two-wire in, two-wire out... and an abnormal condition to stress my point. Let's say we have 24kVA-rated autotransformer transitioning 120V in to 240V out. Max-rated in current is 200A. Max-rated out current is 100A. What ampacity conductors do you use, in and out, and what rating is your circuit ocpd?

I know you're stating an abnormal condition, but its taking me a while to get my head around this. If you are trying to boost 100% from 120V to 240V, you would need a transformer with 1:1 turns ratio. In a practical sense, you could use a 25kVA 480x240 primary 120/240 secondary xfmr, connected as 240V in each winding to get your 1:1 ratio. Your rated input current is 208, so 250A would be the max c/b. Use 4/0 on the input and output circuit to be protected by the 250A c/b.

If you wanted to put a 125A c/b in the output circuit somewhere, and then run #2 from that c/b to the load, I don't see why you couldn't do that. The #2 would then be protected at the point they receive their supply (the 125A c/b.) Of course, that connection is pretty much the same as a conventional connection of a 25kVA, 120-240V connection, so I don't see why you would just connect it conventionally.

 

[Smart $]

I see Art. 240 as covering the conductor protection of an autotransformer-connected circuit. 240.21 says that overcurrent protection shall be provided in each ungrounded conductor at the point they receive their supply. Being that the autotranformer circuit is electrically continuous from input to output (vs. an isolating transformer where there is no electrical connection from primary to secondary,) I see the autotransformer input c/b as being the point where the circuit receives its supply.

Perhaps this is where we differ.

Yes the input is where the circuit receives its power... but not where the solo output conductor receives its power. 240.21 says conductor, not circuit... right?. Only one input conductor is continuous with respect to the output (i.e. essentially spliced)... leaving two conductors?one input, one output?not continous because there is at least one transfomer winding or load between it and any other conductor. So by 240.21, the "solo" output conductor must have OCPD regardless of "primary" ocpd, right?

Short on time right now so I'll get to the latter part of your post later.

 

[david luchini]

Yes the input is where the circuit receives its power... but not where the solo output conductor receives its power. 240.21 says conductor, not circuit... right?.

Actually, 240.21 says "ungrounded circuit conductor."

Only one input conductor is continuous with respect to the output (i.e. essentially spliced)... leaving two conductors?one input, one output?not continous because there is at least one transfomer winding or load between it and any other conductor. So by 240.21, the "solo" output conductor must have OCPD regardless of "primary" ocpd, right?

No, the circuit conductors "receive their supply" at the input c/b. Consider, for example, the circuit in the OP. Assume the load current to be 65A @ 230V. Two, 2kVA 240-24V transformers are being used to boost the 208V input to 230V. Of the loads conductors, one is connected directly to the input, and two are connected in series with a 2kVA, 24V transformer winding. If the total load is 25.8kVA, how are the 2kVA, 24V transformer windings "supplying" that power to the load? I don't see that they are "supplying" the conductors.

Using your reasoning, if you had several loads that needed to be connected in series (another unusual condition, but just for the sake of argument) then you would need to install OCPDs between each of the loads. From the branch circuit breaker to the first load would be one circuit conductors, but then the "output" side of the first load would be a new conductors receiving its supply at the point where it leaves the first load so an OCPD would be required? Doesn't make sense to me.

Also, consider if the OP was a problem of bucking the voltage from 230 to 208V for a 71.9A, 208V 3ph load. We could buck the voltage using the same two 2kVA, 240-24V transformers. But in this case, two of the legs would connect to the series connected winding before connecting to the shunt windings. Would you, in this case, provide a new OCPD at the output of the series winding and then run your conductor to the load and to the shunt winding? If the output side of the series winding is the "point the conductors receive their supply" then you would be required to, but it would be in violation of 450.4.

 

[Smart $]

Actually, 240.21 says "ungrounded circuit conductor."



No, the circuit conductors "receive their supply" at the input c/b. Consider, for example, the circuit in the OP. Assume the load current to be 65A @ 230V. Two, 2kVA 240-24V transformers are being used to boost the 208V input to 230V. Of the loads conductors, one is connected directly to the input, and two are connected in series with a 2kVA, 24V transformer winding. If the total load is 25.8kVA, how are the 2kVA, 24V transformer windings "supplying" that power to the load? I don't see that they are "supplying" the conductors.

Using your reasoning, if you had several loads that needed to be connected in series (another unusual condition, but just for the sake of argument) then you would need to install OCPDs between each of the loads. From the branch circuit breaker to the first load would be one circuit conductors, but then the "output" side of the first load would be a new conductors receiving its supply at the point where it leaves the first load so an OCPD would be required? Doesn't make sense to me.

Also, consider if the OP was a problem of bucking the voltage from 230 to 208V for a 71.9A, 208V 3ph load. We could buck the voltage using the same two 2kVA, 240-24V transformers. But in this case, two of the legs would connect to the series connected winding before connecting to the shunt windings. Would you, in this case, provide a new OCPD at the output of the series winding and then run your conductor to the load and to the shunt winding? If the output side of the series winding is the "point the conductors receive their supply" then you would be required to, but it would be in violation of 450.4.

That's a great premise... but in refutation I point you to 210.9 and 215.11. Their titles say "Circuits Derived from Autotransformers". Note usage of the word "from"... and not the word "through", which would be the case if your "shunt winding in series" theory were correct.

PS: How does having an OCPD betweent the series winding and the load violate 450.4???