Is a 480 V 3-wire single mains transformerless UPS a separately derived source?

[Matt_from_CA]

By single main I am referring to how the internal bypass (static switch) is connected via internal bus links to the main input, and the overall UPS system is fed from a single 3-ph 3-wire 480 V AC source.

According to IEEE 3003.1-2019 9.5, the answer is simply no, because "Transformerless UPS units do not provide electrical (galvanic) isolation between the UPS ac output and its ac input. As such, transformerless UPS units should not be treated as a separately derived system. Any ground-fault current on the output of the transformerless UPS returns to the source of the ac input and not the UPS inverter output." There is similar language in 9.2.

It appears this covers only normal operation. Consider either the upstream UPS input circuit breaker being opened unexpectedly, or if the UPS is applied in conjunction with a standby diesel generator, then during the brief period of time during the break-before-make transition between sources: the UPS is still providing power downstream via battery/inverter, but is presumably operating ungrounded (no neutral being brought to the UPS since it is a 3-wire system)? You can find some information online saying that due to the duration of exposure this is a non-issue, which seems reasonable, but it seems to be something worth considering. It also looks like some UPS manufacturers provide the required ground detection circuity per 250.67.

It's a bit funny to install a UPS to provide backup power in such scenarios, but then for it's classification as a separately derived source to change resultingly.

Posting to encourage some discussion, maybe I am interpreting things wrong.

 

[topgone]

By single main I am referring to how the internal bypass (static switch) is connected via internal bus links to the main input, and the overall UPS system is fed from a single 3-ph 3-wire 480 V AC source.

According to IEEE 3003.1-2019 9.5, the answer is simply no, because "Transformerless UPS units do not provide electrical (galvanic) isolation between the UPS ac output and its ac input. As such, transformerless UPS units should not be treated as a separately derived system. Any ground-fault current on the output of the transformerless UPS returns to the source of the ac input and not the UPS inverter output." There is similar language in 9.2.

It appears this covers only normal operation. Consider either the upstream UPS input circuit breaker being opened unexpectedly, or if the UPS is applied in conjunction with a standby diesel generator, then during the brief period of time during the break-before-make transition between sources: the UPS is still providing power downstream via battery/inverter, but is presumably operating ungrounded (no neutral being brought to the UPS since it is a 3-wire system)? You can find some information online saying that due to the duration of exposure this is a non-issue, which seems reasonable, but it seems to be something worth considering. It also looks like some UPS manufacturers provide the required ground detection circuity per 250.67.

It's a bit funny to install a UPS to provide backup power in such scenarios, but then for it's classification as a separately derived source to change resultingly.

Posting to encourage some discussion, maybe I am interpreting things wrong.

First, you have to understand the definitions of the words or phrases involved here. A separately-derived system is a separate system other than the service. The service system is the source used to provide power to your load; generation and the respective transmission/ service lines.
2nd, you have the UPS. Depending on the architecture, a UPS could be directly connected to the service (normal operation) and will only become a "separate" source (change-over) once the service fails (that's why it is called an Uninterruptible Power Supply). Or, your UPS could be always "online", meaning the UPS is always supplying the load whether the service is out or not (as in a double-conversion online UPS). This blurs the distinction between the two because the UPS can independently supply the load and there is no observed change-over (service to UPS or vice versa), IMO.

 

[Matt_from_CA]

Thanks - it is a double conversion UPS (not operated normally in Eco mode, which is just this particular manufacturer's way of saying not normally through the static bypass).

Blurry indeed.

Well, I suppose one could argue then that from the downstream critical loads perspective, the service is always the same - the inverter output.

But from the DC links perspective, the service is either the battery, or the rectifier output.

Edit: and this is not even considering the other permutations involving the static bypass when there is a fault on the downstream critical system.

 

[topgone]

Thanks - it is a double conversion UPS (not operated normally in Eco mode, which is just this particular manufacturer's way of saying not normally through the static bypass).

Blurry indeed.

Well, I suppose one could argue then that from the downstream critical loads perspective, the service is always the same - the inverter output.

But from the DC links perspective, the service is either the battery, or the rectifier output.

Edit: and this is not even considering the other permutations involving the static bypass when there is a fault on the downstream critical system.

Viewed from the load side, there are 2 sources, the inverter output and the utility (bypass switch).

 

[Matt_from_CA]

Understood. In non-fault scenarios, the static bypass is N/A to this discussion right?