A single-phase UPS supplies a panel through two fundamentally different fault-current paths: the inverter path and, when available, the bypass path. Follow the power from the source through the UPS output main breaker, panel bus, branch breaker, conductors, and fault. The protective device that opens first depends on the current delivered through that active path and the devices’ time-current behavior—not simply on their ampere ratings.
Where does the output fault current flow?
Under normal protected operation, the inverter supplies the panel. A branch short circuit therefore reaches the inverter before it reaches any alternate source. Industrial UPS inverters commonly limit output current electronically to protect their switching devices. That limit may be too low, or may persist for too little time, to operate the instantaneous magnetic element of a branch breaker.
If the UPS recognizes the overload and an acceptable bypass source is available, the intended path changes:
- The load produces an overcurrent at one branch.
- The branch breaker remains closed while the inverter reaches its current-limited region.
- The UPS commands a transfer to bypass. The described design objective is transfer within a fraction of a cycle, but the selected UPS documentation must provide the actual transfer behavior.
- The bypass source supplies the prospective fault current through the UPS output main and the branch breaker.
- The branch breaker, output main breaker, a bypass protective device, or an upstream device opens according to its operating curve and the current passing through it.
Layer one first. Before evaluating trip curves, confirm that the bypass source is physically present, connected, within the UPS acceptance window, and capable of supplying the required fault current. A bypass selection on a display does not prove that the complete power path can carry and clear a downstream fault.
Why might the branch breaker fail to trip on the inverter?
A breaker’s handle rating is not its instantaneous pickup value. A 20 A thermal-magnetic breaker can carry currents above 20 A for a time determined by its thermal curve, while its magnetic element requires a substantially higher current determined by the specific breaker design. The exact pickup band and clearing time must come from the manufacturer’s time-current curve.
Consider the stated 10 kVA, 120 V single-phase example:
I = 10,000 VA / 120 V = 83.3 A
The eight branches each carry 10 A, giving a total indicated load current of 80 A. The remaining difference between nominal current and load current is only 3.3 A. However, 83.3 A is the nominal apparent-power current, not automatically the inverter’s maximum instantaneous fault current. Read the UPS overload curve, current-limit value, permitted duration, crest-factor restrictions, and shutdown behavior to determine what it can actually deliver.
The same kVA rating at 230 V produces a lower nominal output current:
I = 10,000 VA / 230 V = 43.5 A
This calculation explains why limited inverter fault current can become more noticeable at the higher voltage for an otherwise equal kVA rating. It does not predict breaker operation by itself. The breaker curve and the UPS fault-current envelope still decide whether the branch device clears before the inverter transfers or shuts down.
| Observed symptom | Likely mechanism | Measurement or document that decides it |
|---|---|---|
| Branch breaker does not open while the UPS remains on inverter | Inverter current limit stays below the breaker’s fast-trip region | UPS current-versus-time envelope and branch breaker time-current curve |
| Entire UPS output drops after a branch fault | Bypass was unavailable, transfer was rejected, or an upstream device opened first | UPS event log, bypass status, and protective-device indication |
| UPS transfers, but the output main opens | Main and branch curves overlap at the available bypass fault current | Prospective fault current and overlaid curves including tolerances |
| Protection behaves differently at 230 V | Equal kVA provides lower nominal current, while breaker pickup remains device-specific | UPS output rating, overload data, and branch breaker curve |
How do breakers, fuses, and independent sources compare?
| Approach | What it addresses | Primary limitation | Decision data |
|---|---|---|---|
| Coordinate branch and main breakers on bypass | Attempts to isolate only the faulted branch | Curves may overlap at high fault current, creating a race between devices | Bypass fault current and breaker curves |
| Use branch fuses | A selected fuse may clear faster than a selected breaker in part of the fault-current range | Speed is not universal; space, replacement, and high circuit count affect practicality | Fuse clearing curve, breaker curve, enclosure layout, and fault current |
| Add current sensing with a shunt trip | Permits a separately selected trip threshold and logic | Adds relays, wiring, trip power, testing, and significant cost | Sensor range, relay logic, shunt-trip requirements, and failure modes |
| Use dual-corded loads from independent sources | Keeps a load operating when one supply path is lost | Requires equipment with two inputs and each source sized for the full required load | Load input architecture and capacity of each source |
Fuses are not automatically the answer merely because they are described as faster. Compare the total clearing curve of the proposed fuse with the branch and upstream breaker curves at the calculated minimum and maximum fault currents. A large installation also has to accommodate fuse holders, spare-fuse control, replacement access, and the number of branch circuits.
Dual independent sources solve a different problem. They improve load availability after one source path trips, but they do not remove the need to clear the fault selectively. The two inputs must be electrically independent, and either source must have enough capacity to carry the required load after the other source is lost.
Which protection approach should be selected?
For a conventional UPS-fed panelboard, retain branch breakers and evaluate selective operation with the UPS bypass path as the fault source. This approach fits multi-circuit panels and avoids adding a separate sensing and shunt-trip system to every branch. Treat the UPS as part of the path when checking transfer eligibility and current withstand, then analyze the bypass circuit like a coordinated distribution system after transfer.
Do not infer selectivity from a 20 A branch breaker below a 60 A main breaker. Molded-case device curves can overlap when fault current rises above 1,000 A or 2,000 A, so either breaker may win the race in that region. Those values describe the cited coordination difficulty, not universal pickup thresholds. Use the curves for the exact installed devices.
Select fuses only when their published curves demonstrate a material coordination advantage and the physical installation can support them. Select independent dual feeds when continuity of the individual loads justifies the additional source and equipment architecture. For a critical application where curve analysis cannot establish the result, specify a controlled fault-clearing test on a representative UPS and protection assembly.
How should breaker coordination be evaluated?
- Draw the complete one-line path from the bypass source to the fault. Include upstream protection, UPS bypass protection, UPS output main, panel main if separate, branch breaker, and conductor impedance.
- Record the exact protective-device catalog data and settings. Obtain time-current curves with tolerance bands rather than comparing only handle ratings.
- Obtain the UPS inverter overload and short-circuit response. Record current limit, duration, transfer conditions, transfer behavior, and self-protection action from the selected unit’s technical data.
- Determine prospective fault current at the panel and at the end of each relevant branch. Include source and conductor impedance; the minimum current checks whether the branch device trips promptly, while the maximum current checks device duties and coordination.
- Overlay the branch, main, bypass, and upstream curves. Evaluate both inverter-limited current before transfer and bypass-source current after transfer.
- Check the transition interval. A downstream device cannot clear from bypass current until the UPS has accepted and completed the transfer.
- Identify every overlap where the branch and upstream devices can open at the same current. Mark that region as nonselective rather than predicting a winner from nominal ratings.
- Confirm that the output main and bypass path can carry the fault for the downstream device’s total clearing time.
If exact curves or settings are unavailable, obtain them from the device manufacturer. Generic breaker families are not interchangeable for a coordination study.
What changes when bypass is unavailable?
Loss of utility power removes the common bypass source precisely when battery-backed operation matters most. A fault during this condition remains on the inverter path. If inverter-limited current cannot operate the branch protection, the UPS protects itself and shuts down, leaving all loads on that output without power.
No downstream coordination study can create fault current that the inverter cannot deliver. The design choices are to reduce the protected zone, use protection that operates within the documented inverter current envelope, divide loads among smaller independently protected UPS paths, or provide dual-corded equipment from independent sources. An adjustable overcurrent relay and shunt trip can lower a branch trip threshold, but the design must discriminate normal starting or inrush current from a fault and must define which branch to trip.
Check UPS event records after a test or incident. Distinguish inverter current limit, bypass transfer, transfer rejection, overload shutdown, branch trip, and upstream trip. A dark panel alone does not identify where the sequence stopped.
How is the design verified before service?
- Confirm voltage and loading measurements on each branch and at the UPS output. For the stated example, reconcile the measured total with the calculated 83.3 A nominal current at 10 kVA and 120 V.
- Verify bypass availability under the operating modes that matter, including normal utility operation and utility loss.
- Review curve overlays at the calculated minimum and maximum branch-fault currents. Record any region where the branch and main devices are not selective.
- Perform an approved controlled test on a like UPS and protective-device arrangement when loss of the full panel is unacceptable. Some UPS manufacturers may decline short-circuit testing for safety or liability reasons, so establish test acceptance before procurement.
- During the test, capture output current, UPS operating state, transfer indication, event log, and the device that opened. Pass only when the intended branch device clears without an unacceptable loss of the remaining loads in every required source condition.
What do engineers ask about UPS output faults?
What happens if a UPS branch faults while bypass is available?
The inverter may current-limit and command a bypass transfer. After transfer, the branch breaker, output main, bypass breaker, or an upstream device opens according to the available fault current and their time-current curves.
What happens if the bypass source is unavailable?
The inverter remains the only fault-current source. If its documented current envelope cannot trip the branch breaker, the UPS protects itself and shuts down, interrupting all loads on that output.
How do I verify that the 20 A branch breaker clears first?
Overlay the exact branch and upstream device curves at the calculated fault current, then run an approved representative test. Capture current, UPS state, transfer indication, event log, and the device that opens as the final verification step.