Automatic transfer switches — return sequence and cooldown
Generator Will Not Transfer Back to Mains
- Applies to
- Contactor-based and motorised automatic transfer switches on utility and generator supplies, single- and three-phase
- Difficulty
- intermediate
- Competence required
- qualified electrician
- Diagnosis complexity
- Low. Most of these are a return timer doing its job, and the rest divide cleanly between mains sensing and the mechanism.
- Electrical system
- Utility and generator supplies 240 V / 415 V 50 Hz nominal; control supply per panel design
- Safety classification
- live electrical
- Author
- EmersonEIMS Engineering
- Technical review
- Mr. Kararaho
- Last reviewed
- 2026-07-27
Scope — read this before relying on the guide
This guide sets out diagnostic method, ranked causes and safety requirements. It deliberately does not state equipment-specific numbers — test voltages, resistance and component values, torque figures, acceptance windows and pin assignments are referred to the manufacturer's documentation for your exact model rather than given here, because a plausible-looking figure for the wrong machine is more dangerous than no figure at all.
Direct technical answer
Check the return timer before assuming anything is broken, because on most installations it is deliberately long and a set still running twenty minutes after the lights come back is usually correct behaviour rather than a fault. The return delay exists so the system does not transfer back onto a utility supply that has returned but is still unstable, and a cooldown period then lets the set run unloaded before stopping. If the timers have genuinely expired and the load is still on the generator, the question becomes whether the controller accepts the returned mains as healthy. It judges that on its own sensing terminals, so a blown sensing fuse, a lost phase or reversed rotation after upstream work will leave it correctly refusing to return to a supply it cannot verify. Only when the controller is commanding a return that does not happen should the mechanism be suspected, and at that point the distinction is whether the command reaches the switch or the switch cannot act on it. Running on generator is expensive but not dangerous, so there is no reason to force a transfer before understanding why it has not occurred.
01Symptom description
Controller / display
- Mains restored indication with the load still on generator
- Return or retransfer timer counting with no transfer following
- Controller showing mains not accepted despite utility being present
- Cooldown running after transfer, which is normal
- Position discrepancy alarm between commanded and actual state
Indicators
- Generator contactor still energised with utility available
- Mains available indicator absent on the controller
- Mode selector found in manual or test
- Control fuse or MCB open
Sounds
- Set continuing to run long after the utility returned
- Contactor chattering, which suggests marginal sensing or a weak control supply
- A clunk with no change of position, indicating a mechanical obstruction
- Motor operator running without completing travel
Smells
- Burnt smell from contactor coils or contacts
- Hot insulation smell at power terminations
Behaviour
- Returns eventually but only after a long delay, which is the timer rather than a fault
- Never returns until someone intervenes manually
- Returned normally until upstream electrical work was carried out
- Transfers to generator promptly but will not come back, which separates the two directions cleanly
- Returns on some occasions and not others, pointing at marginal sensing rather than a failed component
- Set runs on indefinitely, raising fuel cost and running hours
Visible
- Return delay and cooldown timer settings
- Mode selector position
- Mains-sensing fuses and where sensing is connected
- Contactor or switch contact condition, including signs of welding
- Mechanical interlock condition and free movement
- Evidence of recent upstream electrical work
02What the fault means
In plain language
The utility has come back but the load is still being fed by the generator. Usually the panel is simply waiting out a deliberate delay so it does not switch back onto a supply that has returned but is still unsettled. If the delay has passed, either the panel does not accept the returned supply as good enough, or it is trying to switch and the mechanism will not move.
Technical explanation
Return to the preferred source is governed by the same three-part logic as the outbound transfer, in reverse. The controller must judge the utility acceptable against its configured criteria, a return delay must expire, and the switching mechanism must execute the command. The return delay is normally set longer than the failure delay by design, because a supply that has just been restored is more likely to fail again, and transferring the load back prematurely risks a second interruption. A cooldown period usually follows the transfer, allowing the set to run unloaded and dissipate heat before stopping, so continued running after the load has moved is expected rather than faulty. Acceptance is judged at the controller sensing terminals, which is why the controller can refuse a utility supply the building is otherwise using: a blown sensing fuse, an open sensing lead, a lost phase or reversed rotation after upstream work all leave it unable to verify the source. On the mechanism side, motorised operators and contactor arrangements can fail to release from the generator position, and a mechanical interlock will physically prevent the return if the mechanism has not fully cleared. Contacts welded closed by an earlier fault current are a specific and dangerous case, because the controller may indicate a completed transfer while the load remains connected to the previous source.
03Common causes, ranked
These are ordered by likelihood. Presenting every possible cause as equally probable is a failure of diagnosis, not thoroughness.
Most likely
start here- Return delay or cooldown timer still running — normal behaviour mistaken for a fault
- Mains-sensing fuse blown or sensing lead open, so the controller cannot verify the returned supply
- Mode selector left in manual or test
- Utility supply present but outside the acceptance criteria — voltage, frequency or a lost phase
Possible
check next- Phase rotation reversed after upstream electrical work
- Return timer configured far longer than the site expects
- Control supply weak or its fuse open
- Contactor coil or its control circuit failed on the mains side
- Auxiliary contact failed, so the controller has an incorrect view of position
Less common
after the above- Mechanical interlock jammed, preventing release from the generator position
- Motor operator failed part-way through travel
- Contacts welded closed on the generator side
- Controller configuration corrupted or changed
- Building management system inhibiting the return
Model specific
verify per unit- Return delay and cooldown durations are configurable and differ widely between installations — read them from the controller
- Acceptance thresholds for the returned supply are model-specific
- Some controllers require the utility to be stable for the full delay and restart the timer on any dip
- Interlock arrangements differ; mechanical on some designs, electrical on others
- Manual return procedures differ and must not be improvised
Environmental
site conditions- Unstable utility supply repeatedly restarting the return timer
- Moisture and dust causing tracking on control circuits
- Vibration loosening control terminations
- Insect ingress into the panel
Installation related
built in- Sensing connected to a point that does not represent the utility being monitored
- Return timer left at a commissioning default unsuited to the site
- Control wiring routed alongside power cables
- No documented test of the return sequence, so it is never proven
Maintenance related
deferred work- Return sequence never tested, only the outbound transfer
- Timer settings never recorded, so changes are undetectable
- Contacts never inspected for wear or welding
- Mode selector not returned to auto after intervention
Component level
electronics- Mains-sensing fuse open
- Mains contactor coil failed
- Auxiliary contact failed
- Motor operator failed
- Contacts welded on the generator side
04Safety requirements
Isolation
- The panel is fed from TWO independent live sources; isolating the utility leaves the generator side live and the reverse
- Isolate both sources and prevent the generator starting before working inside the panel
- Lock the generator control in stop and isolate its starting battery
- Prove dead on both incoming sides and the load side at the point of work
Lockout and tagout
- Lock and tag the utility supply, the generator supply and the generator control
- Confirm the load may lose supply for the duration, or arrange an alternative
- Keep the only key with the person doing the work
PPE
- Arc-rated protection appropriate to the prospective fault energy at the panel
- Insulated tools rated for the system voltage
- Eye protection
- Hearing protection while the set runs
Stored energy
- Motorised operators may hold stored spring energy; release it by the documented method
- Control circuits may remain live from a separate supply after the main sources are isolated
- The generator starting battery remains live
- The set may be running throughout the investigation
Specific hazards
- NEVER defeat the mechanical interlock to force a return. It exists to prevent utility and generator being connected together, which back-feeds the network and can kill someone working on supposedly dead lines.
- Welded contacts can hold the load on the previous source while the controller indicates a completed transfer — verify position physically rather than from the indicator.
- The generator can start or stop automatically during the work unless its control is locked off.
- Never open-circuit a current transformer secondary while load current flows.
- Running on generator is expensive, not dangerous. There is no reason to rush a forced transfer before the cause is understood.
Stop and call a qualified professional if
- The mechanical interlock is damaged, jammed or has been defeated
- Contacts are suspected welded
- There is a burnt smell or evidence of arcing in the panel
- The load cannot lose supply and no alternative exists
- Manual operation of this switch type is not documented or not understood
05Tools required
| Tool | Why it is needed |
|---|---|
| True-RMS multimeter rated for the system voltage | Utility voltage measured at the ATS sensing terminals, and control circuit checks |
| Phase rotation tester | Reversed rotation after upstream work disqualifies the returned supply instantly |
| Clamp meter | Confirming which source is actually carrying the load, independent of the indicators |
| Controller documentation and settings access | Return delay, cooldown and acceptance thresholds must be read rather than assumed |
| Proving unit and voltage indicator | Proving dead in a panel fed from two sources |
| Thermal camera | Power terminations and contact condition under load |
| Stopwatch or timing record | Establishing whether the delay observed matches the configured timer |
06Diagnostic decision tree
1. Is there a burnt smell or evidence of arcing in the panel?
Yes → Stop. Isolate both sources and escalate.
No → Continue
2. Have the return delay and cooldown timers actually expired?
Yes → Continue
No → Not a fault. The panel is waiting by design — confirm the settings suit the site.
3. Is the mode selector in AUTO?
Yes → Continue
No → That explains it, and it is a common finding after earlier work
4. Is utility present at the ATS SENSING terminals, all phases, correct rotation?
Yes → Continue
No → The controller cannot verify the supply — sensing fuse, lost phase or rotation
5. Does the controller report the utility as available?
Yes → The decision is correct — investigate the mechanism
No → Sensing or acceptance thresholds, not the switch
6. Is a return command being issued at the controller output?
Yes → Command present but no movement indicts the coil, operator or mechanism
No → The fault is in sensing, configuration or the controller
7. Does the mechanism move freely by the documented manual method?
Yes → Suspect the coil, operator or its control circuit
No → Interlock or obstruction — do not force it
8. Is the load physically confirmed to be on the source the indicator claims?
Yes → Proceed
No → Suspect welded contacts — this is a safety matter, not an indication fault
07Step-by-step diagnosis
Step 1Read the timers before treating anything as a fault
- Inspect
- Return delay and cooldown settings, and how long the utility has actually been back
- Where
- Controller configuration and display
- Instrument
- Controller interface, timing record
- Expected result
- Behaviour consistent with the configured delays
- If the result is abnormal
- A return delay is deliberately long, and a set still running well after the lights come back is usually correct. Establishing this first prevents a great deal of unnecessary investigation.
- Next
- Step 2
Step 2Check the mode selector and read the controller
- Inspect
- Selector position, active alarms and what the controller reports about the utility
- Where
- At the controller
- Instrument
- Display or service interface
- Expected result
- Auto selected; a clear statement about the utility source
- If the result is abnormal
- The controller usually says plainly whether it regards the utility as available, which is faster than measuring and points immediately at sensing or mechanism.
- Next
- Step 3
Step 3Measure the utility at the ATS sensing terminals
- Inspect
- Voltage on every phase, and phase rotation, at the ATS rather than a nearby board
- Where
- ATS utility input and sensing terminals
- Instrument
- True-RMS multimeter and phase rotation tester
- Expected result
- All phases present, balanced, correct rotation
- If the result is abnormal
- A lost phase or reversed rotation leaves the controller correctly refusing a supply the rest of the building is using. Reversed rotation after upstream work is a classic finding here.
- Next
- Step 4
Step 4Compare what the controller senses against your measurement
- Inspect
- Controller-reported utility values against your own readings at the same terminals
- Where
- Controller display versus ATS terminals
- Instrument
- Multimeter and controller display
- Expected result
- Agreement
- If the result is abnormal
- Disagreement isolates the fault to sensing — a blown sensing fuse or open lead. The controller is then making a correct decision on wrong information.
- Next
- Step 5
Step 5Establish whether a return command is being issued
- Inspect
- Controller output and the corresponding coil or operator circuit
- Where
- Controller output terminals through to the mechanism
- Instrument
- Multimeter
- Expected result
- Command present and arriving at the mechanism
- If the result is abnormal
- This separates a decision problem from a mechanism problem, and it is the step that stops people replacing switches that were never commanded to move.
- Next
- Step 6
Step 6Assess the mechanism and interlock
- Inspect
- Free movement, interlock condition, contact condition and auxiliary contacts
- Where
- At the switching mechanism, both sources isolated and proven dead
- Instrument
- Manual operation by the documented method
- Expected result
- Mechanism moving freely and completing travel in both directions
- If the result is abnormal
- A mechanism that hums or partly moves is usually obstructed or not fully released from the generator position. Never force it and never defeat the interlock.
- Next
- Step 7
Safety: Isolate both sources and lock the generator control before touching the mechanism.
Step 7Verify the load position physically, not from the indicator
- Inspect
- Which source is actually carrying the load
- Where
- On the load conductors
- Instrument
- Clamp meter
- Expected result
- Load on the source the controller indicates
- If the result is abnormal
- Welded contacts can hold the load on the previous source while the controller indicates a completed transfer. That is a safety matter and must be resolved before the panel is returned to service.
- Next
- Correct the identified cause and prove the full sequence
08Repair procedure
Timers and mode
configuration- Set return delay and cooldown to values that suit the site rather than commissioning defaults
- Return the mode selector to auto and make that a documented final step
- Verify acceptance thresholds against the supply the site actually has
- Record all settings after any change
A large share of these callouts end here, having found the panel behaving exactly as configured.
Sensing and control
cleaning and connections- Replace blown mains-sensing fuses after establishing why they operated
- Repair open or corroded sensing leads and control wiring
- Re-torque power and control terminations and survey thermally under load
- Clear moisture and insect ingress and seal panel entries
Switching components
component replacement- Replace failed contactor coils and auxiliary contacts
- Replace contactors with pitted or welded contacts rather than dressing them
- Replace a failed motor operator
Mechanism and interlock
mechanical- Free or correct a jammed mechanism only by the documented method
- Repair or replace a damaged mechanical interlock — never bypass it
- Confirm travel completes in both directions after any work
Interlock integrity is a life-safety matter because of the back-feed risk.
Sensing connection point
wiring- Correct phase rotation at its source after upstream work
- Move sensing to a point that genuinely represents the utility being monitored
- Label sensing fuses so the next person finds them quickly
09Post-repair validation
- Simulate a mains failure and prove the FULL sequence including the return, not just the outbound transfer
- Time the return delay and cooldown against the configured values
- Confirm the load is carried by the expected source, verified by measurement rather than indicator
- Verify phase rotation on both sources
- Confirm the mode selector is left in auto and record that it was
- Thermally survey power terminations under load after the work
- Confirm the mechanical interlock operates correctly and has not been disturbed
- Record all settings, measurements and the test result in the panel documentation
10When not to repair
- Switch assemblies with damaged mechanical interlocks, where safety cannot be assured by repair
- Contactors with welded or heavily eroded contacts, which are replacement items
- Panels undersized for the present load
- Obsolete controllers and switch assemblies that are unobtainable
- Any situation where restoring function would require defeating the interlock
11Prevention
- Test the RETURN sequence at every service, not only the transfer to generator — the return is the half that goes untested for years
- Record return delay and cooldown settings so unintended changes are detectable
- Return the selector to auto as a documented step at the end of every intervention
- Verify phase rotation after any upstream electrical work
- Inspect contacts for wear and welding at service intervals
- Seal panels against moisture and insect ingress
- Confirm load position by measurement periodically rather than trusting indicators
12Questions engineers actually ask
The power came back ten minutes ago and the generator is still running. Is it broken?
Almost certainly not. The return delay exists so the system does not transfer back onto a supply that has just returned and may fail again, and a cooldown then lets the set run unloaded before stopping. On many installations that adds up to a considerable wait. Check the configured timers before treating it as a fault, and confirm they match what the site actually wants.
It transfers to the generator fine but never comes back. Why one direction and not the other?
Because the two directions use different information. Transferring out depends on detecting that the utility has FAILED; returning depends on verifying that it is healthy again. If the mains-sensing fuse has blown or a phase is lost, the controller can still see a failure but can no longer confirm a good supply, so it correctly refuses to return. Measure the utility at the ATS sensing terminals, not at a nearby board.
Can we just force it across manually to save fuel?
Only by the documented manual method for that switch, and only after understanding why it has not returned. Running on generator is expensive but not dangerous, so there is no need to rush. What you must never do is defeat the mechanical interlock — it prevents utility and generator being connected together, which back-feeds the network and can kill someone working on supposedly dead lines.
The controller says it transferred back, but the generator is still loaded. What is happening?
Verify the load position with a clamp meter rather than trusting the indicator. Contacts welded closed by an earlier fault current can hold the load on the generator while the controller believes the transfer completed. That is a safety issue as well as an operational one, and the panel should not be returned to service until it is resolved.
Standards and references
- IEC 60947-6-1 — low-voltage switchgear and controlgear: transfer switching equipment
- IEC 60364 — low-voltage electrical installations, including provisions for safety services and supplies
- ISO 8528 — generating sets, including control and switchgear considerations
- The ATS and controller manufacturer's documentation for the specific equipment, which is the only valid source for return delay and cooldown functions, acceptance thresholds, manual operation procedure and interlock arrangements referred to throughout
This guidance is written from engineering principle and is not a substitute for the manufacturer's model-specific documentation. Where a figure is model-specific, confirm it against the service data for your unit before acting on it.
Need this diagnosed properly?
Send us the make, model and any fault codes shown, and photographs of the controller display if you have them. Our mobile workshop covers all 47 counties.