- October 15, 2026
- Posted by: editor
- Category: RECD & Emission Control Devices
Ask most DG operators how long a generator needs to warm up and you’ll hear “ten to fifteen minutes, until the coolant gauge settles.” For running the plant, that’s a reasonable answer. For an emission test on a generator fitted with a Retrofit Emission Control Device, it is one of the most common reasons a good device gets a bad report.
The coolant tells you about the engine block. It tells you nothing about a large steel canister full of ceramic or metal substrate several metres downstream. That canister is the part being tested, and it is usually the last thing in the system to reach a steady temperature.
This article walks through what actually happens, thermally, between pressing the start button and the moment a sample is valid. It explains what goes wrong when sampling starts too early, and how to write a warm-up rule into the test plan that both buyer and supplier can accept.
Why Warm-Up Matters More With an RECD Than Without
On a bare exhaust pipe, the gas reaching a sampling port has cooled a little on the way, but there is nothing in the path that stores large amounts of heat or holds pollutants. Once the engine is warm, the exhaust at the port settles fairly quickly.
An RECD changes that in three ways:
- It adds a lot of thermal mass. The housing, insulation, substrate and end cones together can weigh tens or hundreds of kilograms on mid-size sets, and much more on large ones. All of it has to heat up before the gas leaving the device is at a steady temperature.
- It contains temperature-sensitive chemistry. An oxidation catalyst has a light-off temperature below which it does little. A catalysed filter burns soot passively only above a certain temperature.
- It stores material. Soot, organics and water can sit in the device after shutdown and be released, burnt or evaporated as it warms.
So the question isn’t “is the engine warm?” It’s “has the RECD reached the condition it will be in during normal operation at this load?”
The Five Thermal Phases of a DG Emission Test
It helps to divide a test day into phases. The times below are indicative only. Actual durations depend on DG size, load, ambient temperature, insulation and how far the device sits from the engine.
Phase 0: Cold soak (before start-up)
The set has been off, perhaps overnight. Everything is at or near ambient temperature. In a winter basement in north India, that might be 8–12 °C. On a summer rooftop it could be 40 °C or more.
During cooldown, moisture in the remaining exhaust gas condenses inside the pipework and the RECD. Heavier hydrocarbons and some soot settle on cool surfaces. Our article on protecting an RECD from rainwater and condensation explains why this matters for the device’s life as well.
What to record: ambient temperature, time since last shutdown, how long the last run lasted.
Phase 1: Engine warm-up (typically the first 10–20 minutes)
Coolant and lube oil climb towards operating temperature. Exhaust temperature at the manifold rises quickly, often within a few minutes, but the gas is losing heat to cold pipework and to the RECD all the way downstream.
During this phase you may see:
- White vapour from the stack as water condensed overnight is pushed out (see why a generator emits white smoke)
- Higher HC and CO, because the engine and catalyst are both cold
- Some visible smoke on older engines until combustion chambers warm up
What to record: coolant temperature, exhaust temperature before the RECD, RECD outlet temperature, every minute.
Phase 2: The device catches up (often another 20–60+ minutes)
The engine reads “warm,” but the RECD is still absorbing heat. The temperature difference between the RECD inlet and outlet is large at first and shrinks slowly. This is the phase most often mistaken for steady state.
Inside the device:
- The catalyst moves through light-off, and CO and HC conversion improve rapidly once it passes its threshold
- Condensed water evaporates and leaves the stack
- Stored organics evaporate or begin to oxidise
- Back-pressure rises as the gas gets hotter and expands
Sampling in Phase 2 measures a device that is still settling.
What to record: the RECD inlet-to-outlet temperature difference, back-pressure, and CO and HC trends if a gas analyser is running.
Phase 3: Equilibrium at the test load
Inlet and outlet temperatures have both levelled off, back-pressure is steady, and gas concentrations have stopped trending. This is the state the device will be in during normal operation at this load, and it is the only phase where a steady-state sample means anything.
What to record: the time equilibrium was declared, and the readings that justified it.
Phase 4: Mode changes, and starting the clock again
When load steps from one test point to the next, the system leaves equilibrium. Stepping up in load makes the device heat up again. Stepping down makes it cool, and that can be slow, because the stored heat has to dissipate.
Sampling right after a step down catches the RECD hotter than it would normally run at that load. That can make light-load performance look better than it really is, because the catalyst and filter are still benefiting from heat left over from the previous mode.
What to record: time of each load change, and when equilibrium was re-established.
How Big Is the Thermal Lag?
The lag between engine and device depends on a few predictable factors:
| Factor | Effect on warm-up time |
|---|---|
| Larger RECD (more mass) | Longer |
| Light test load (cooler exhaust) | Much longer, and may never reach catalyst activity |
| Long exhaust run between engine and device | Longer; see where to position an RECD along a long exhaust run |
| Uninsulated pipework | Longer; see RECD insulation and heat shields |
| Cold ambient (winter, basement, rooftop at night) | Longer |
| Device recently run (still warm) | Shorter, but now you’re testing a different starting state |
On a light-loaded set in winter, it isn’t unusual for the RECD outlet to need well over an hour to settle. Fixed rules like “run for 20 minutes, then sample” often fail in exactly these conditions.
Signs That a Sample Was Taken Too Early
When you review a test report, or your own data, these are signs the device hadn’t stabilised:
- RECD outlet temperature was still climbing during the sample, by more than a few degrees
- CO or HC trended down during the sampling period, a sign the catalyst was still lighting off
- Back-pressure trended up during sampling
- Visible white vapour was noted at the stack during or just before sampling
- The “after” result is much worse at the first test point than at later ones, even though later points are at similar or lower loads
- A repeat sample at the same load later in the day gave a clearly better result
- There’s no temperature log at all, so stability can’t be demonstrated either way
If two or more of these appear, the first sample is likely to understate the device’s real performance.
Why Warm-Up Bias Usually Hurts the RECD
The baseline and the post-RECD test aren’t affected equally by early sampling.
Baseline test (no RECD): the exhaust path has little thermal mass. Once the engine is warm, the port settles quickly. A 15-minute warm-up is often close enough.
Post-RECD test: the device is the slowest part of the system. The same 15-minute warm-up leaves it in Phase 2.
So if both tests use the same fixed time instead of the same stability criterion, the baseline is sampled at equilibrium and the RECD test is sampled during warm-up. The comparison is biased against the device. This is common in sequential tests (bare pipe one week, RECD fitted the next), where the “after” test is often squeezed in on commissioning day with little run time.
Upstream/downstream sampling on the same day reduces this problem, because the device has to be at equilibrium for both readings. Even then, the upstream reading settles first. Don’t start either sample until the downstream side is stable.
Writing a Warm-Up Rule Into the Test Plan
Use a stability criterion, not a stopwatch
Define readiness by what you can measure. A practical rule that both parties can agree:
Stability criterion: Sampling at any load point starts only when (a) load has been within tolerance for at least 10 minutes, and (b) the RECD outlet gas temperature has changed by less than 5 °C over the previous 5 minutes, and (c) exhaust back-pressure has changed by less than an agreed amount over the same 5 minutes.
The 5 °C and 5-minute figures are sensible starting points, not regulatory values. Large devices or very light loads may justify a tighter rule or longer window. What matters is that the rule is written down, the same for both tests, and backed by a log.
Add a minimum time as a safety net
Some test plans add a minimum run time on top of the stability criterion, for example “and not less than 30 minutes after start-up.” This stops a sample being taken during a brief flat spot in the warm-up curve, which can happen when condensed water is evaporating.
Repeat after every load change
Apply the same criterion after each step between load points. Stepping down usually takes longer to settle than stepping up, so plan the test day with that in mind.
Agree on the break-in question
A brand-new RECD can behave differently in its first hours. Coatings settle, binders burn off, and the filter has no soot layer yet. Agree in writing whether the acceptance test:
- happens on commissioning day,
- happens after a short break-in run (for example, a few hours at moderate load), or
- happens after a set number of operating hours in normal service.
There is no universal right answer, but there should be an agreed one. RECD trial installation acceptance criteria covers this kind of pre-agreement.
What to Measure, and Where
A useful warm-up record needs only a few instruments:
| Measurement | Where | Why |
|---|---|---|
| Coolant temperature | DG controller | Confirms the engine is warm |
| Exhaust temperature | Before the RECD | Shows what the device is receiving |
| Exhaust temperature | After the RECD | Shows when the device itself has settled |
| Back-pressure | Engine exhaust outlet | Tracks gas expansion and device loading |
| CO and HC | Downstream port (if analyser available) | Shows catalyst light-off progress |
| Load (kW) | Controller or power analyser | Confirms the load was steady during warm-up |
| Ambient temperature | Near the DG | Explains seasonal differences |
Thermocouples should be fitted into the gas stream through proper ports, not taped to the outside of the housing. Skin temperature lags gas temperature and is affected by insulation and airflow.
A Sample Warm-Up Log
Here is what a convincing record looks like for a mid-size set warming up at about 50% load in winter. The figures are illustrative:
| Minutes from start | Coolant (°C) | Before RECD (°C) | After RECD (°C) | Back-pressure (mbar) | Status |
|---|---|---|---|---|---|
| 0 | 14 | 14 | 12 | 0 | Start |
| 10 | 68 | 288 | 74 | 31 | Engine warming |
| 20 | 82 | 318 | 162 | 39 | Engine warm; device lagging |
| 30 | 84 | 326 | 228 | 43 | Device catching up |
| 40 | 84 | 330 | 266 | 45 | |
| 50 | 84 | 331 | 284 | 46 | |
| 60 | 84 | 332 | 292 | 47 | |
| 65 | 84 | 332 | 295 | 47 | Δ < 5 °C over 5 min → sample start |
In this example, a “20 minutes and go” rule would have started sampling with the RECD outlet still at 162 °C, more than 130 °C below its eventual operating temperature. Any conclusion from that sample would describe the warm-up, not the device.
Site and Seasonal Factors That Stretch Warm-Up
Winter and cold basements
A DG in an unheated basement in Delhi NCR in January starts from a much lower temperature than the same set in May. Warm-up can take noticeably longer, and condensation during Phase 0 is heavier. If the baseline was taken in summer and the post-RECD test in winter, warm-up differences add to every other seasonal difference. Planning guidance for these rooms is in RECD installation in a basement DG room.
Light-load sets
Some standby sets rarely run above 25–30% load. At such loads the RECD may stabilise at a temperature where the catalyst is only partly active. That isn’t a warm-up problem as such, since the device has settled where it will normally sit. The test should report it honestly instead of trying to warm the device artificially with a higher load first. See choosing an RECD for a generator running at low load.
Sets that only run briefly
If your set typically runs for 10–15 minutes during weekly tests and short outages, it may spend most of its life in Phases 1 and 2. Steady-state test results won’t describe that operation, and cold-start behaviour deserves a separate look. See can an RECD work reliably during cold starts and RECD selection for DG sets with frequent short start-stop cycles. For sets that run rarely, see is an RECD suitable for a DG set that runs only occasionally.
Should You Test Cold-Start Emissions Too?
A steady-state comparison is the standard basis for proving a particulate reduction. Cold-start behaviour still matters, especially for sets that run in short bursts. If you want to characterise it:
- Keep it as a separate test, clearly labelled, so it isn’t averaged into the steady-state result
- Use a consistent cold soak (for example, at least 8 hours off) before each run
- Sample over a defined window from start-up (for example, the first 20 minutes) and repeat the same window in both tests
- Expect wider run-to-run variation, and plan more repeats
This gives you useful information about real-world behaviour without contaminating the main acceptance result.
Warm-Up Checklist for Test Day
- Record ambient temperature and time since last shutdown
- Fit gas thermocouples before and after the RECD
- Bring the set to the first test load and hold it steady
- Log temperatures, back-pressure and load every minute
- Apply the agreed stability criterion and minimum time
- Note the exact time sampling starts, and the readings that justified it
- Re-apply the criterion after every load change
- Avoid shutdowns between load points
- Attach the full temperature-versus-time log to the report
How Warm-Up Interacts With the Other Test Variables
Warm-up is closely linked to the other variables in an RECD comparison:
- Load sets the temperature the device will settle at. See why generator load must be matched in RECD comparison tests.
- Regeneration can begin during warm-up as the filter gets hot enough to burn stored soot. See RECD testing during regeneration: what should be recorded.
- Sampling location determines which temperature the sample sees, and so how much condensable material it collects. See how exhaust sampling location affects RECD test results.
- Dilution problems such as air leaks often show up as unusual temperature readings. See how dilution can mislead RECD before-and-after measurements.
Frequently Asked Questions
How long should a DG set run before an RECD emission test?
There’s no fixed number that works everywhere. Run until the RECD outlet temperature is stable at the test load, for example changing less than about 5 °C over 5 minutes, and add a sensible minimum time. On many sets that takes 45–90 minutes from cold, and longer at light load or in cold weather.
Isn’t coolant temperature a good enough indicator?
No. Coolant tells you the engine block is warm. The RECD has much more thermal mass and sits downstream, so it settles much later. Measure the gas temperature after the device.
Can I pre-heat the RECD by running at high load, then drop to the test load?
You can, but you must then wait for the device to cool to its natural equilibrium at the test load before sampling. Otherwise the result reflects leftover heat and can flatter light-load performance. Do the same in both tests.
Does warm-up affect the baseline (no-RECD) test as well?
Yes, but much less, because a bare pipe has little thermal mass. The danger is using the same fixed warm-up time for both tests. The baseline settles and the RECD test doesn’t, which biases the comparison against the device.
Our RECD has an electric heater. Does that remove the warm-up problem?
A heater may shorten warm-up or support regeneration, depending on the design, but you still need to show stability before sampling, and record when the heater was on. See RECD controller and heater power supply requirements.
Build Warm-Up Into the Plan, Not the Excuses
A warm-up rule costs nothing to write and saves days of argument later. Aceget supplies CPCB type-approved Retrofit Emission Control Devices for DG sets from 25 kVA to 2500 kVA, and we help clients set temperature measurement points and stability criteria before acceptance testing. For the overall test approach, read our guide on planning an RECD emissions test before and after installation. For background on the oxidation chemistry inside many devices, DieselNet’s overview of diesel oxidation catalysts is a useful technical reference. When you’re ready, contact our team with your DG details.