- November 14, 2026
- Posted by: editor
- Category: RECD & Emission Control Devices
When an RECD regeneration cycle ends early for no obvious reason, engineers look at the filter, the heater and the engine. Fewer look at the humble DC supply feeding the controller. Yet a controller that loses power for half a second can reset, forget it was in the middle of a cycle and quietly abandon the burn. The heater contactor drops out, the log shows nothing dramatic and the soot stays in the filter.
This article follows the power. We walk from the starter battery to the controller, sensors, relays and heater circuit, identify five ways RECD regeneration with battery-supported controls can be interrupted, explain what a well-designed controller should do when power returns and give you a test script to prove that your installation survives the disturbances it will actually meet.
The details differ between controllers. Treat this as a framework to discuss with your supplier and electrician, not as a wiring instruction. Where a point depends on your device, the article says so.
Why control power matters more than it seems
Regeneration is a controlled process that depends on a chain of small signals: pressure and temperature sensors, a decision logic, an output that closes a contactor, and sometimes a communication link to the DG controller. If any link loses power, the chain breaks.
Three features of a generator room make power disturbances more likely than in an office.
- Cranking dips. When the starter motor engages, battery voltage can fall sharply for a moment. Anything on the same DC bus may see that dip.
- Transfer events. When load moves between grid and generator, AC-fed supplies, chargers and contactors all see a change. Some pause, some restart.
- Electrical noise. Large motors, contactors and alternator regulation produce spikes and interference.
Add the fact that regeneration often runs at night or during outages, when conditions are least forgiving, and control power becomes a first-order design topic.
Follow the power: five paths to check
Walk through your installation in this order. Write down the answer for each path.
Path 1: The controller supply
What feeds the RECD controller? Options include the DG’s starting battery, a dedicated battery with charger, a 24 V DC supply fed from the essential board, an AC-DC supply with or without a DC UPS, or mains-fed supply with a small internal battery.
Questions.
- Is the supply derived from the DG starting battery? If so, what happens during cranking?
- Is it on a dedicated, labelled supply that cannot be switched off by mistake?
- Does it have undervoltage and overvoltage limits that match the controller’s rating?
- Is there a surge or reverse-polarity protection device?
Path 2: The sensor supply and signals
Pressure transmitters and temperature sensors are often powered from the controller. A sag can make readings freeze, drift or jump. The controller may then abort the cycle for “sensor fault” or, worse, keep going on false data.
Questions.
- Do sensors share the same supply as the controller or have their own regulated output?
- What does the controller do when a sensor signal goes out of range for a short time?
- Are cable routes away from heat and noise? See RECD cable routing near hot exhaust lines.
Path 3: The output relay and contactor coil
The controller commands a relay, which commands the heater contactor. The contactor coil has its own supply, perhaps AC from the essential board or DC from the same battery.
Questions.
- What happens to the contactor if the coil supply dips? Many contactors drop out at a fraction of the rated voltage and then need an explicit re-close command.
- Will the controller know that the contactor opened, or does it only know that it commanded closed?
- Is there auxiliary-contact feedback to the controller?
Path 4: The heater power circuit
This is the power circuit that actually warms the exhaust: a breaker, a contactor, perhaps a thyristor controller and the element. Our article on RECD controller and heater power supply requirements explains the electrical needs.
Questions.
- Is the heater circuit fed from the DG-backed essential board?
- Does heater power drop during a transfer, and for how long?
- Is the DG sized to carry the heater load while the building load is on?
Path 5: The communication and logging path
The controller records the cycle and may send status to the DG controller or the BMS. A communication gateway on a separate supply can lose power, so the data trail has a gap exactly when the cycle was interrupted.
Questions.
- Is the data logger on the same battery-supported supply?
- Does the log record supply voltage or a power-up event?
- Does the controller log a reason when it resumes?
Five interruption scenarios
Each scenario below describes what happens, why it harms the cycle, and what to look for.
Scenario 1: Cranking dip resets the controller
What happens. The DG starts, the starter motor pulls hard on the battery, and the controller’s supply drops below its minimum for a fraction of a second. The controller reboots.
Why it matters. If a regeneration was pending or running, the reboot may clear the state. This is most common when the controller shares the starting battery without a DC-DC converter or a hold-up capacitor.
What to look for. A power-up event or start-up message in the controller log at every DG start, or cycles that never begin after a start.
Scenario 2: Transfer interrupts AC-fed supplies
What happens. During a changeover, AC-fed parts of the control supply drop for the transfer time. A DC supply without a battery or capacitor passes the interruption straight to the controller.
Why it matters. A regeneration that was running on the grid side loses heat at the same moment.
What to look for. Controller power-up events or aborted cycles coinciding with transfer times in the building’s records.
Scenario 3: Charger or battery weakness
What happens. An ageing battery and a failing charger cannot hold voltage under load. Everything looks fine until a heavy demand arrives.
Why it matters. Low voltage can produce partial operation: sensors misread, relays chatter, a contactor half-closes and overheats.
What to look for. Battery voltage trending low in logs, controller undervoltage warnings, heater contactor chatter or burnt contacts.
Scenario 4: Heater supply lost while control power remains
What happens. The controller is healthy but the heater breaker has tripped or the heater feed was load-shed. The controller believes the heater is on.
Why it matters. The cycle “runs” with no heat, ending on a timer. This produces the “complete by clock only” cases discussed in our guide to interpreting an RECD regeneration completion message.
What to look for. Heater current reading near zero while the controller shows heater on; no temperature rise.
Scenario 5: Backup runs out mid-cycle
What happens. A supply with only a few minutes of battery backup carries the controller through the DG start, but if the DG fails to take load, the backup runs flat during a long cycle.
Why it matters. The controller shuts down in the middle of a burn, possibly leaving a hot filter without monitoring.
What to look for. Controller shutdown messages with falling voltage, loss of data during outages.
What the controller should do when power returns
Hardware is half the story. The other half is the controller’s behaviour at restart. Ask your supplier for the answers below in writing.
| Question | Safe behaviour | Concerning behaviour |
|---|---|---|
| Does it remember it was regenerating? | Stores cycle state in non-volatile memory | Starts again as if nothing happened |
| Does it resume, restart or abandon? | Defined rule for each case, depending on how long power was out and filter temperature | Behaviour depends on luck |
| Does it check conditions before resuming? | Verifies temperature, ΔP, load and sensor health | Resumes blindly |
| Does it keep settings and clock? | Retains configuration; restores time or flags it as unsynchronised | Reverts to defaults or wrong date |
| Does it log the event? | Records power loss, restore, and action taken | Silent |
| Does it prevent a dangerous restart? | Blocks heater energisation until sensors are valid | Switches heater on at once |
A controller that resumes sensibly can ride through a short dip. One that restarts every cycle from zero will waste energy and may stress the filter. One that abandons the cycle silently may leave soot behind with no warning. Our article on repeated failed regeneration explains why repeated abandoned cycles matter.
Design options that reduce interruptions
None of these is mandatory. Discuss which suit your site with the supplier and a qualified electrician.
A dedicated supply for the RECD controller
Separate the controller from the engine starting battery, or isolate it with a DC-DC converter that tolerates cranking dips. A dedicated battery and charger, or a DC UPS with a monitored battery, gives a clean bus.
Hold-up capacity matched to the weakest event
A hold-up device should cover the longest expected disturbance with margin. For a transfer taking a few seconds, a modest hold-up is enough. For an outage with a failed DG start, a longer battery backup is needed to keep records. Our article on RECD monitoring during power failure covers the monitoring side, including a worked UPS sizing example, so this article does not repeat that arithmetic.
Heater power on the essential board, with feedback
Heater feeds should be on the DG-backed essential supply, protected, labelled and fitted with current or contactor feedback to the controller.
Contactor holding strategy
A contactor that drops out on a short dip should not be assumed to be closed on restoration. Use feedback contacts or choose a design that remains closed through short dips, as the supplier recommends.
Clear alarm for supply faults
Alarm on low battery voltage, charger failure, supply loss and heater current mismatch. Wire at least one summary alarm to a hardwired contact so that a communication fault does not hide it. See the RECD alarm response matrix for operators.
Battery maintenance as part of the routine
Test and replace batteries on a schedule. Heat shortens battery life, so keep batteries away from hot exhaust surfaces.
Protect the cycle at the process level
Instruct operators not to initiate transfers or load shedding during a running regeneration unless safety requires it. See RECD regeneration while supplying building loads.
A worked example: reading voltage at the controller
The following is an illustrative scenario, not a real customer. A logistics warehouse operates a 160 kVA DG set. Its RECD controller is fed from a 24 V DC supply that comes from the DG’s starting battery bank through a simple fuse. Over three months, the maintenance technician notices that out of eleven regeneration requests after DG starts, only four complete. The log shows a “controller power-up” entry in seven of the failures.
The test. The electrician connects a voltage logger to the controller’s supply terminals and starts the DG three times.
| Start | Minimum controller supply during crank | Duration below 18 V | Controller power-up logged |
|---|---|---|---|
| 1 | 16.8 V | 0.4 s | Yes |
| 2 | 17.2 V | 0.3 s | Yes |
| 3 | 18.9 V | 0 s | No |
The supplier’s datasheet shows the controller operates down to 18 V. The battery was also two years old and its voltage dropped more under load than the third start suggests, with the third start taking place after a long charge.
The findings. The supply dipped below the controller’s minimum during two of three cranks. A weak battery made it worse.
The remedy (illustrative). The team installs a DC-DC converter with a wide input range between the battery and the controller, replaces the battery, and adds a battery voltage alarm. They also ask the supplier to confirm the controller’s resume logic. After the change, the next ten DG starts show no power-up events and nine of ten regeneration requests complete. The tenth fails because of a load drop, which is a different matter.
The lesson. The filter, heater and engine were healthy. The faults were in the supply, and only a voltage log taken during a real start revealed them.
A test script: prove it before you rely on it
The best time to find a weakness is during a planned test, not in the middle of a night-time outage. The following script suits a commissioning check or an annual test. Adapt it with your supplier and electrician, and follow site electrical safety rules. Do not carry out live-electrical work unless you are qualified and authorised.
Preparation
- Agree the test with the facility manager. Warn affected staff.
- Have the controller manual, the supply drawing and the supplier’s contact at hand.
- Install a voltage logger on the controller supply and, if possible, a current clamp or logger on the heater feed.
- Note the starting state: battery voltage, charger status, ΔP, controller settings and clock.
Test 1: Start-up dip
Start the DG with the controller powered and no regeneration active. Record minimum supply voltage during cranking. Check for any controller reset, alarm or log entry.
Test 2: Transfer disturbance
During a planned transfer to or from DG, record the supply voltage and any controller events. If safe and agreed, repeat while a regeneration is pending.
Test 3: Brief control supply interruption
With the supplier’s permission, interrupt the control supply for a short, defined time, such as one or two seconds, and observe what the controller does when it returns. This tests the resume logic. Do not do this during a hot regeneration unless the supplier explicitly approves.
Test 4: Heater feed loss
Simulate loss of heater supply, for example by opening the heater breaker during a controlled test with the supplier. Check that the controller detects the loss and raises an alarm rather than continuing on a timer.
Test 5: Backup duration
Measure how long the controller and logger stay up on backup with the DG stopped. Compare with the required time.
Records
Document each test: date, who, conditions, measured values, observed behaviour and any action. Keep the record with the commissioning file and update it after changes. See also RECD commissioning with load bank testing and preserving backup power during an RECD retrofit.
Pass and fail criteria to agree in advance
| Test | Pass looks like | Fail looks like |
|---|---|---|
| Start-up dip | Supply stays above controller minimum; no reset | Reset, alarm or log gap |
| Transfer | No loss of state; cycle continues or resumes by rule | Cycle abandoned silently |
| Short interruption | Controller restarts, logs, checks and resumes or aborts safely | Heater energised immediately on restart |
| Heater feed loss | Alarm within the agreed time | No alarm, cycle ends on timer |
| Backup duration | Meets agreed time with margin | Controller shuts down early |
Signs in the data that control power is the culprit
You do not always need a test to suspect power. Look for these patterns in your logs.
- Aborted cycles clustered at DG start or at transfer times.
- Controller power-up or reset events without an operator action.
- Clock jumps or dates reverting to default.
- Data gaps that coincide with outages.
- Heater current readings of zero while the controller shows heater on.
- Cycles that complete by clock but with a low temperature rise.
- Battery or charger alarms in the DG controller.
For a broader view of separating electrical faults from filter problems, read separating a sensor fault from an RECD performance problem and RECD communication failure with the DG controller.
Who owns this?
Control power straddles trades. The RECD supplier owns the controller. The electrical contractor owns the supply and heater feed. The DG maintainer owns the starting battery. Facilities owns the transfer logic. When nobody owns the interface, problems drift. Write one owner against each path in the table at the start of this article, and record them in the project documents. See customer and installer responsibilities in an RECD project.
FAQ
Can a battery-backed controller completely prevent interrupted regeneration?
No. Backup protects the controller and logs from power dips, but regeneration can also be interrupted by load drops, generator stops, sensor faults and heater supply loss. Battery support removes one family of causes.
Should the RECD controller run from the DG starting battery?
It can, if the design is made for it, but cranking dips are a known risk. Many installations prefer a dedicated supply or an isolating DC-DC converter. Follow the controller manufacturer’s guidance and discuss with a qualified electrician.
How long should backup last?
Long enough to ride through transfers and DG start-up with margin, and, for sites that need records, through a failed start. The right figure depends on the site’s outage profile. Test it rather than estimate it.
What if my controller has no resume logic?
Then protect the cycle at the process level (stable supply, no transfers during cycles) and ask the supplier whether a firmware update or a different controller is available. Document the limitation in the operating procedure.
Is it safe to test interruptions during a hot regeneration?
Not without the supplier’s approval. Interrupting control during a hot cycle can leave the heater or burner in an unknown state. Test on a cold system or follow a supplier-approved procedure.
Where does this fit in the warranty?
Supply quality is often a customer responsibility under warranty terms. Keep test records and voltage logs, which help if a dispute arises. Our guide to RECD warranty terms explains typical exclusions.
Bringing it together
Interrupted regeneration is not always a filter problem. Follow the power through five paths: controller supply, sensor supply, relay and contactor, heater circuit, and logging. Know what the controller does when power returns, choose design options that reduce disturbances, and test your installation with a voltage logger and a written script. When you find a gap, fix it before the next outage finds it for you.
For help matching controller and electrical requirements to your DG set, see our retrofit emission control devices, read the FAQ or contact the Aceget team.