I'm not 100% clear if the disconnector is before or after the UPS - if before, questions about the N-PE reference of the output come to mind.

Otherwise, I'd agree with Mike about the order of poles breaking. It's common enough on single pole supplies to have RCDs trip when switching off DP switches - if the N opens first and there's a relatively high impedance fault between N and PE. In normal working conditions there's only a small voltage difference between N and PE so the leakage current is often small enough for the RCD to tolerate - but if the N contact opens first, the downstream N conductors are pulled towards line voltage by the load, so so the increased voltage between N and PE results in more current leaking - often enough to go beyond the RCD's threshold. 3-phase could be similar if the loads aren't well balanced (or otherwise have non-cancelling N currents - e.g. from single phase switch-mode PSUs).

  - Andy.

I'm not 100% clear if the disconnector is before or after the UPS - if before, questions about the N-PE reference of the output come to mind.

Otherwise, I'd agree with Mike about the order of poles breaking. It's common enough on single pole supplies to have RCDs trip when switching off DP switches - if the N opens first and there's a relatively high impedance fault between N and PE. In normal working conditions there's only a small voltage difference between N and PE so the leakage current is often small enough for the RCD to tolerate - but if the N contact opens first, the downstream N conductors are pulled towards line voltage by the load, so so the increased voltage between N and PE results in more current leaking - often enough to go beyond the RCD's threshold. 3-phase could be similar if the loads aren't well balanced (or otherwise have non-cancelling N currents - e.g. from single phase switch-mode PSUs).

  - Andy.

Thanks Andy,

The disconnector is before the UPS (at the input to UPS), on the N-PE reference of the output are you referring to when the UPS is in battery mode?

Or is there something else to be considered?

Regards

are you referring to when the UPS is in battery mode?

Indeed. Some UPS don't establish their own N-PE link but rely on the N-PE loop of the supply even when in battery mode (not a good policy in my opinion, but apparently not uncommon) - so breaking the supply N might have the effect of separating load N from PE ... which can again cause odd effects as the N is no longer held close to 0V but all the conductors get to be loosely referenced to Earth via filters and so on.

   - Andy.