UPS bypass — static bypass, maintenance bypass and the backfeed risk (IEC 62040-1)
UPS bypass — static bypass, maintenance bypass and the backfeed risk (IEC 62040-1)
The guide on UPS continuity for climate computers and critical processes covers the VFI/VI/VFD classification per IEC 62040-3 and battery autonomy time. This article covers a different aspect of the same UPS, standardised in IEC 62040-1: the bypass paths that exist alongside the inverter path, and the safety pitfall that arises when the distinction between the static bypass and the maintenance bypass is not properly understood.
Static bypass: automatic, electronic, for inverter faults
The static bypass is an electronic switch (typically thyristor-based) that automatically switches the load from the inverter path to the direct mains path within a few milliseconds — for example on an internal inverter fault, an overload the inverter cannot carry, or an inrush current spike that exceeds the inverter's rated capacity. The static bypass is intended to preserve continuity of the load when the inverter path itself (temporarily) cannot supply it — not to enable servicing of the UPS. While the static bypass is active, the load sits directly on the mains rather than behind the inverter, and the load therefore temporarily loses the protection against mains disturbances that the UPS normally provides.
Maintenance bypass: manual, mechanical, for de-energising the UPS unit itself
The maintenance bypass is a separate, usually manually operated switch (sometimes implemented as a make-before-break mechanical transfer switch) intended to keep the load fed via the mains while the UPS unit itself is fully isolated and de-energised for inspection, servicing, or replacement — without the load experiencing an interruption. The essential difference from the static bypass: the maintenance bypass is intended to enable working on the UPS itself, not to automatically absorb an inverter fault.
The backfeed risk: why the switching sequence is critical
An incorrectly executed maintenance-bypass procedure can result in part of the UPS installation — for example the inverter output terminals, or part of the internal cabling path — unintentionally remaining energised while a technician assumes the entire unit is isolated. This backfeed risk typically arises from:
- Engaging the maintenance bypass before the UPS output has been correctly disconnected, so that during an overlap period two sources (mains via bypass, and the still-active inverter) can be present at the same node simultaneously.
- A battery connection that is not disconnected during the maintenance procedure — the battery can continue feeding the inverter's DC bus, so parts of the inverter electronics remain energised even after the AC supply to the UPS has been removed.
- A parallel-connected UPS module in an N+1 configuration that, via the shared output rail, still backfeeds voltage into the part being worked on.
The correct sequence: confirm bypass first, only then isolate
The manufacturer's procedure for a maintenance-bypass changeover typically follows a fixed sequence — first activate the maintenance bypass and explicitly confirm that the load is genuinely being fed via the bypass path (for example via a voltage indicator or status message on the panel), only then isolate the UPS output and the battery connection. Reversing this sequence — attempting to isolate the UPS first and only then activating the bypass — is the most common cause of either an unintended loss of supply to the load or, conversely, of a component unexpectedly remaining live.
Practical relevance
Before working on a UPS unit, it must not only be verified that the maintenance bypass has been activated, but also explicitly checked — with a voltage tester on the actual terminals to be worked on, not solely on the basis of a status indication on the panel — that every part of the UPS being worked on is genuinely de-energised, including the battery connection. This aligns with the general five-step procedure (see the LOTO five-step guide), but for a UPS specifically requires attention to the battery path as a separate voltage source that is not disconnected via the mains bypass.
Common mistakes
- Treating the static bypass and the maintenance bypass as the same thing — the static bypass is an automatic, fault-absorbing path; the maintenance bypass is a manual provision to enable working on the UPS itself.
- Executing the maintenance-bypass changeover in the wrong sequence — attempting to isolate before the bypass has genuinely been confirmed, resulting in a backfeed risk or unintended loss of supply.
- Not disconnecting the battery connection during maintenance work — the battery can keep the inverter's DC side energised even after the AC supply has been removed.
- Relying solely on the bypass panel's status indication without verifying with a voltage tester on the actual work terminals that they are de-energised.
Related
Further reading
- §8-1 / IEC 60364-8-1NEN 1010 Part 8-1 — Energy efficiency of electrical installations (IEC 60364-8-1)
- §445 (IEC 60364-4-45)§445 — Protection against unintended automatic restart after undervoltage
- §721Electrical installations in caravans and motor caravans (§721)
- NEN 1010 §729 (gangpaden)Operating and maintenance gangways for switchgear and distribution boards (NEN 1010 §729)
- IEC 62305-2Lightning protection — risk assessment and LPL class under IEC 62305-2
- §512.2 (AD/AE) / §411.3.3Outdoor socket-outlets — minimum IP44, mandatory 30mA RCD and the distinction from fixed outdoor lighting (§714)