- October 18, 2026
- Posted by: editor
- Category: RECD & Emission Control Devices
Here’s a design flaw that turns up on site after site. The Retrofit Emission Control Device controller is plugged into a convenient socket in the DG room. The socket is fed from the normal grid supply. Everything works perfectly during commissioning, weekly test runs and the acceptance test, because the grid is on during all of them.
Then the grid fails. The diesel generator starts, picks up the building load, and runs for six hours. For those six hours, the RECD controller is dark. No differential pressure reading, no temperature alarms, no regeneration control if the device needs it, and no data in the Building Management System.
A generator mostly runs because of a power failure. RECD monitoring therefore has to be designed for exactly that situation. This guide follows a power failure from the first second to the hours after the grid returns, sets out what should stay available at each stage, and explains how to power it, size backup, handle restarts and test the arrangement.
The Basic Principle
Anything needed to operate, monitor or protect the RECD while the DG runs must be powered from a supply that is live while the DG runs.
That usually means one or both of:
- An essential (DG-backed) supply, energised by the generator through the automatic transfer switch (ATS)
- An uninterruptible power supply (UPS) or battery-backed DC supply, which bridges gaps and covers the case where the DG itself fails
Normal grid-only circuits are fine for convenience loads, but not for RECD monitoring.
Following a Power Failure, Second by Second
T = 0: The grid fails
Everything on grid-only circuits goes dark. On a well-designed system, monitoring equipment on a UPS or battery-backed supply stays alive. The BMS records the grid failure.
What should stay available: BMS controllers, network equipment and RECD monitoring on UPS or DC backup, so the event is logged and the DG start can be seen.
T ≈ seconds: The DG starts and the ATS transfers
The DG cranks, reaches speed and voltage, and the ATS transfers essential loads to it. This usually takes seconds to tens of seconds, depending on the set and settings.
What should stay available: the same as above. The UPS carries monitoring through the gap until the essential supply returns.
DG running (minutes to hours)
This is the period the RECD exists for. The engine is producing exhaust, the device is filtering, and soot is accumulating or being burnt.
What must be available:
- RECD sensors and controller: ΔP, back-pressure, temperatures, regeneration status
- Regeneration control and any heater supply, for active-regeneration devices
- Alarm outputs, both hardwired contacts and the data link
- BMS controllers and network switches carrying RECD data to the control room
- Any remote-alert gateway, if used
- Local display and lighting in the DG room for operator checks
See how to design an RECD alarm response matrix for operators.
The grid returns
The ATS transfers the load back to the grid, sometimes with a brief break. The DG usually runs on for a cool-down period with no load.
What should stay available: monitoring through the re-transfer. Transfer transients shouldn’t reset the controller, and if they do, it must restart cleanly (see below).
Cool-down and shutdown
The engine idles, then stops. The RECD remains hot for some time, and a regeneration may even be in progress.
What should stay available: temperature monitoring and logging for a period after shutdown, especially for active-regeneration devices, so any abnormal temperature rise is still detected.
The scenario nobody plans for: the DG fails too
The grid fails and the DG doesn’t start, or it starts and then trips, perhaps on low fuel, overheating or a fault. Now the site is in a full blackout.
What should stay available: enough backup power for the BMS, network and alerting to tell people that the DG has failed and what the last readings were. RECD data is part of that picture. If the DG tripped on high back-pressure, for example, the RECD’s last ΔP and back-pressure readings are vital clues.
What Should Stay Available, and From Which Supply
| Function | Needed while DG runs? | Needed in total blackout? | Recommended supply |
|---|---|---|---|
| RECD sensors and controller (monitoring) | Yes | Useful (last values, logging) | Essential supply plus UPS or DC battery |
| RECD heater (active regeneration) | Yes, if the device uses one | No | Essential (DG-backed) supply; usually too large for a UPS |
| Hardwired alarm contacts and their receiving panel | Yes | Yes | Essential supply plus UPS |
| BMS controller reading RECD data | Yes | Yes | Essential supply plus UPS |
| Network switches and gateways | Yes | Yes | Essential supply plus UPS |
| Cellular gateway for remote alerts | Yes | Yes | Essential supply plus UPS, or own battery |
| DG room lighting | Yes | Emergency lighting | Essential supply and emergency lighting |
| Cloud dashboard | Optional | Optional | Not on site; depends on mobile network |
The key distinction is between small monitoring loads, which are tens of watts and are practical to back up with a UPS or battery, and heater loads, which can be kilowatts and must come from the DG-backed essential supply.
Powering the Heater for Active Regeneration
Some RECDs use an electric heater to help regenerate the filter, especially on sets that run at light load. See RECD controller and heater power supply requirements.
If the heater is fed from a grid-only circuit, it can’t operate while the DG is running during an outage, which is precisely when it’s needed. Soot then builds up, ΔP rises and regeneration fails.
Points to check:
- Is the heater fed from the essential board?
- Is the DG sized to carry the heater load in addition to the building load? Heaters may draw several kilowatts. See does an RECD change a generator’s usable power output.
- Is the heater circuit protected and labelled so it isn’t load-shed by mistake?
- Does the controller know whether heater supply is available, and does it alarm if not?
Sizing a UPS or Battery for Monitoring Equipment
Monitoring loads are small, so backup is inexpensive. A worked example, with illustrative figures:
| Equipment | Typical power |
|---|---|
| RECD controller and sensors | 20 W |
| Modbus gateway | 6 W |
| Cellular gateway | 6 W |
| Total monitoring load | 32 W |
Suppose the site wants monitoring to survive 4 hours if the DG fails completely.
- Energy needed: 32 W × 4 h = 128 Wh
- Allow for inverter losses (assume 85% efficiency): 128 ÷ 0.85 ≈ 151 Wh
- Allow for usable battery depth (assume 80%): 151 ÷ 0.80 ≈ 188 Wh
- Add an ageing and temperature margin (assume 25%): 188 × 1.25 ≈ 235 Wh
- At 12 V DC, that’s about 235 ÷ 12 ≈ 20 Ah of battery capacity
A modest UPS or a battery-backed 24 V DC power supply can easily cover this. Many controllers and gateways run on 24 V DC, so a DC UPS avoids inverter losses altogether.
Practical tips:
- Use a dedicated UPS or a dedicated, labelled output on the BMS UPS. Avoid sharing an IT UPS that might be load-shed or switched off for maintenance.
- Monitor the UPS itself. A UPS with a dead battery gives false confidence. Bring its alarm into the BMS.
- Mind the DG room environment. Heat shortens battery life, so locate batteries away from hot exhaust surfaces. See assessing an RECD for very hot ambient conditions.
- Replace batteries on a schedule, not after they fail.
Restart Behaviour: What Happens When Power Returns?
Even with good backup, some equipment may lose power briefly, for example during transfers or UPS maintenance. Check how each device behaves on restart:
- Does the RECD controller restart automatically and resume monitoring and regeneration control, or does it wait for a manual reset?
- Does it keep its settings, such as alarm limits and regeneration parameters, or revert to factory defaults?
- Does it keep its clock? A controller that loses time after a power cut stamps data with the wrong date, which makes trends and alarm logs unreliable. A real-time clock with battery backup, or time synchronisation from the network, solves this.
- Does it keep its logged data? Data stored in volatile memory may be lost.
- Does it handle a restart mid-regeneration safely? Ask the supplier what happens if power is lost while a heater-assisted regeneration is underway.
- Do the gateway and BMS reconnect automatically, without manual intervention?
Ask the RECD supplier to confirm each point, and test them during commissioning.
Transfer Transients
ATS transfers can cause short voltage dips, interruptions or spikes. Sensitive electronics may reset or behave erratically. Protection measures include:
- Feeding monitoring equipment through the UPS, which smooths transfers
- Using surge protection on supplies to controllers and gateways
- Checking that the controller’s power supply is rated for the site’s voltage variations
If controllers reset during every transfer, that’s a design problem to fix, not a quirk to live with.
Data Continuity During Outages
Monitoring data during a DG run is valuable, so make sure it’s kept:
- Local logging in the controller or BMS should continue throughout. Check that log storage has enough capacity for long outages.
- Store-and-forward: if data goes to a cloud platform, the gateway should buffer readings when the mobile network is congested or down, and upload them later. See does an RECD need internet access to operate.
- Time stamps must be correct, which depends on the clock behaviour described above.
This data is useful for diagnosing problems, planning maintenance and showing regulators the device was monitored during real operation. See why RECD test results may differ from day-to-day site readings.
Special Situations
Several DG sets running in parallel
Large sites often run two or more DG sets synchronised on a common bus, starting extra sets as load rises. Each RECD needs its own monitoring supply from a source that’s live whenever its set might run. A single shared supply for all RECD controllers is acceptable only if it’s on the essential bus and backed up. Check that monitoring of a set that starts late, or is held in standby, doesn’t depend on that set itself running.
Sites with solar and battery storage
Where a battery energy storage system can carry the site through short outages, the DG may start only after the batteries run low, sometimes hours into an outage. The DG then runs during a period when parts of the electrical system may already be in reduced or emergency modes. Make sure RECD monitoring and heater supplies are available in that operating mode too, not only in the simple “grid off, DG on” case. Review the control philosophy with whoever designed the hybrid system.
Load shedding schemes
Many sites shed non-essential loads automatically when on DG power, to keep within the set’s capacity. Check that RECD heater and monitoring circuits are not on a shed list. A heater that’s shed to protect the DG can lead to regeneration failure on long runs. If the heater must be sheddable for capacity reasons, the controller should alarm when its supply is removed, and operators should know what that means. See how to design an RECD alarm response matrix for operators.
Sets that run under restrictions
In Delhi NCR, GRAP measures can restrict DG operation to essential and emergency uses during severe pollution episodes. Sets that do run during these periods are likely to be closely watched, which makes reliable RECD monitoring during outages even more valuable as evidence. See how GRAP restrictions affect businesses using DG sets.
Common Mistakes, and How to Spot Them
| Mistake | How it shows up | Fix |
|---|---|---|
| Controller plugged into a grid-only socket | RECD data goes blank whenever the DG runs on outage | Move to essential supply plus UPS |
| Heater on a non-essential board | Regeneration fails during long outages | Move heater to essential board; check DG capacity |
| Gateway or network switch not backed up | BMS shows “comms fail” during outages | Back up network equipment |
| Shared IT UPS switched off for maintenance | Monitoring lost unexpectedly | Dedicated UPS or labelled, protected output |
| Controller loses time after power cut | Trends and alarms with wrong dates | Battery-backed clock or time sync |
| Controller reverts to defaults after power loss | Alarm limits change without anyone knowing | Confirm non-volatile settings; test |
| UPS battery dead | Everything fails at the first real outage | Monitor UPS alarms; replace batteries on schedule |
A Power-Source Schedule for the DG Room
A short table on the electrical drawings, or on the inside of the DG room’s distribution board door, prevents most of the mistakes above. An illustrative example:
| Load | Feeder / circuit | Supply type | Backed up by | Sheddable? | Notes |
|---|---|---|---|---|---|
| RECD 1 controller and sensors | EDB-1, circuit 7 | Essential (DG-backed) | 24 V DC UPS, 4 h | No | Do not move to general sockets |
| RECD 1 heater | EDB-1, circuit 12 | Essential (DG-backed) | None | No | 3-phase; check DG margin |
| Modbus gateway | EDB-1, circuit 8 | Essential | 24 V DC UPS, 4 h | No | |
| Network switch (OT) | EDB-1, circuit 8 | Essential | 24 V DC UPS, 4 h | No | Shared with gateway |
| Cellular gateway | EDB-1, circuit 9 | Essential | Built-in battery, 2 h | No | External antenna |
| DG room lighting | EDB-1, circuit 3 | Essential | Emergency lighting units | No | |
| DG room convenience sockets | DB-G, circuit 4 | Grid only | None | Yes | Not for monitoring equipment |
Label the relevant breakers clearly, for example “RECD MONITORING – DO NOT SWITCH OFF”, so they aren’t turned off during unrelated work.
The Blackout Test
The only way to be sure is to test under real conditions. A blackout test for RECD monitoring:
- Plan it with the facility manager, BMS contractor and DG maintainer, at a time when a planned outage is acceptable.
- Before the test, note current readings on the RECD controller and BMS.
- Open the grid supply to simulate a power failure.
- Observe:
- Did monitoring stay powered through the DG start and transfer?
- Did the BMS keep receiving RECD data throughout?
- Did any controller reset? Did it restart correctly, with settings and clock intact?
- Is the heater supply live (for active-regeneration devices)?
- Did remote alerts still arrive?
- Run the DG on load for long enough to see normal RECD readings, ideally at least 30–60 minutes.
- Restore the grid and observe the re-transfer and cool-down.
- Optional advanced step: with the DG stopped and the grid still open, check how long monitoring and alerts survive on backup alone.
- Record results on a signed test sheet and fix any gaps.
Repeat the test periodically, for example annually, and after any electrical changes in the DG room. It fits naturally alongside routine DG checks. See the monthly maintenance checklist for diesel generators.
Questions to Settle at the Design Stage
- Which supply will feed the RECD controller, sensors, heater, gateway and network equipment?
- Is the essential board capacity, and the DG’s spare capacity, enough for the heater?
- What backup duration is needed for monitoring if the DG fails?
- Who maintains the UPS, and is its alarm on the BMS?
- How does the controller behave on power loss and restart?
Answer these in the RECD order and electrical design, not after commissioning. See how to write a technical RFQ for an RECD project, RECD general arrangement drawing: what to include and customer and installer responsibilities in an RECD project.
Related Reading
- Integrating RECD monitoring with a building management system
- How to design an RECD alarm response matrix for operators
- RECD Modbus data: units, scaling and register mapping explained
- Does an RECD need internet access to operate?
- RECDs for hospitals: retrofit around critical backup loads
Frequently Asked Questions
Does the RECD stop filtering if its controller loses power?
The filter substrate keeps trapping particulate as long as exhaust flows through it. What’s lost is monitoring, alarms and, for active-regeneration devices, regeneration control. Over a long run without regeneration, soot can build up and back-pressure can rise, which is why the controller and any heater need a supply that’s live while the DG runs.
Can I power the RECD heater from a UPS?
Usually not practical. Heaters can draw several kilowatts, far more than a typical monitoring UPS. Feed the heater from the DG-backed essential board, and back up only the low-power monitoring equipment with a UPS.
How long should monitoring run on backup if the DG fails?
Long enough for the right people to be alerted and for last readings to be recorded and retrieved. For many sites, one to four hours is a reasonable target. Critical sites may choose longer, in line with their wider BMS backup policy.
Our controller resets every time the ATS transfers. Is that acceptable?
No. Frequent resets can lose data, clocks or settings and interrupt regeneration. Feed the controller through a UPS or DC backup, add surge protection if needed, and ask the supplier to confirm the controller’s tolerance to supply interruptions.
Do I need to test this if the electrical drawings look correct?
Yes. Drawings and reality often differ. A socket gets swapped, a breaker is relabelled, a UPS battery has aged. Only a real blackout test confirms that monitoring stays available when it matters.
Design for the Moment the DG Actually Runs
RECD monitoring that only works when the grid is on is monitoring that fails when it matters. Aceget supplies CPCB type-approved Retrofit Emission Control Devices for DG sets from 25 kVA to 2500 kVA. We can advise on controller and heater power requirements, and support blackout testing at commissioning. See our certifications, learn how an RECD works, or contact our team.