October 27, 2026
Posted by: editor
Category: RECD & Emission Control Devices

A sensor that screws neatly into the existing port and plugs into the existing connector can still be the wrong sensor. When maintenance teams buy replacement RECD sensors, the commonest failure is not a part that refuses to fit. It is a part that fits perfectly, powers up, shows a number on the display, and quietly reports the wrong value for months.

Wrong range, wrong output signal or wrong scaling cause silent under-reading (high back-pressure goes unnoticed), over-reading (nuisance alarms that operators learn to ignore), or a dead BMS point hidden behind a plausible local display.

This guide treats the job like filling in a specification sheet: identify what is fitted, check each compatibility parameter, review the controller, write a purchase specification, then verify. An illustrative worked example shows how a range mismatch can hide a dangerous back-pressure rise.

First, Which Sensors Might Your RECD Actually Have?

RECD designs in India vary widely. Some are passive units with no instrumentation at all. Others carry one or more sensors connected to a local controller, a display, a data logger or a building management system (BMS). If you are unsure how your unit works, start with this overview of how an RECD works and the main design families.

Where sensors are fitted, the usual types are:

Differential pressure (ΔP) transmitters, measuring the pressure drop across a filter element through two impulse lines or hoses. A rising ΔP typically indicates soot or ash loading.
Exhaust temperature sensors, often thermocouples or resistance temperature detectors (RTDs), placed upstream and/or downstream of the treatment element.
Other instruments depending on design, such as a static back-pressure sensor at the engine outlet, a heater or regeneration status input, or a level switch on wet scrubber types.

Each feeds a decision (an alarm, a regeneration prompt, a maintenance trigger), so a replacement must reproduce the meaning of the signal, not just the fit. For background on why ΔP matters so much, see our explainer on how exhaust back-pressure affects a DG set.

Step 1: Build a Sensor Identity Card Before You Order Anything

Do not order from memory or from a photo of a dirty label. Create an identity card for every sensor on the RECD, using:

The sensor’s nameplate or marking, photographed clearly.
The RECD supplier’s part number or tag number.
Photos of the controller’s input configuration pages.
The wiring diagram and terminal schedule from the O&M manual.
The commissioning record, which may list ranges and alarm settings.
The RECD spare parts list supplied at handover, if your team requested one.

Sensor identity card template

Field	What to record	Example entry
Tag / location	Tag on drawing and physical position	ΔP-01, across filter inlet/outlet
Function	What decision the signal drives	Filter loading alarm, regeneration prompt
Make and model	Exactly as printed on the sensor	As per nameplate
RECD supplier part number	Supplier’s own code	As per spare parts list
Measured quantity and range	Units and span	0–10 kPa differential
Output signal	Type and span	4–20 mA, 2-wire
Sensor element type	For temperature sensors	Thermocouple type K / Pt100 3-wire
Supply voltage	At sensor terminals	24 V DC loop
Process connection	Thread, size, insertion length	1/4 inch fitting, hose barb, or compression gland
Electrical connector	Type and pin assignment	M12 4-pin / cable gland with flying leads
Cable	Length, screened or not, temperature rating	5 m screened, high-temperature sheath
Controller input	Terminal numbers, input type, scaling	AI-2, 4–20 mA, 0–10 kPa
Alarm settings	Value and units, delay	High at 8 kPa, 10 s delay
BMS point	Point name and units	RECD1_DP_kPa
Last good reading	Value, load, date	3.1 kPa at about 60% load

The last row gives you something to compare against after the swap. If you keep an operating log, the RECD shift-handover log is a natural place to find those historic values.

Step 2: The Compatibility Parameter Checklist

Compare every parameter below between the original and the proposed replacement. The third column says whether it must match or may differ with deliberate action.

Parameter	Why it matters	Must match / may differ with care
Measured quantity and range	Range sets the conversion from signal to engineering units. A different span changes every reading unless the controller is rescaled.	Prefer exact match. May differ only if the controller is rescaled and alarm values are re-entered.
Output signal type	A 4–20 mA input will not read a 0–10 V signal correctly, and vice versa. Some inputs can be damaged by the wrong type.	Must match the controller input, or the input must be reconfigured if it supports both.
Sensor type and curve	Thermocouple type K and type J generate different voltages at the same temperature. Pt100 and Pt1000 have different resistances.	Must match the controller’s configured curve.
Cold-junction handling	Thermocouple readings depend on compensation where the thermocouple wire meets copper. A head-mounted transmitter moves this point.	May differ with care; change the controller input type to suit.
Accuracy	A less accurate sensor widens the band of uncertainty around alarm thresholds.	May differ with care; equal or better than original.
Response time	A slow sensor may miss short temperature peaks; a very fast one may cause nuisance alarms without filtering.	May differ with care; check alarm delays.
Supply voltage and wiring	2-wire loop-powered, 3-wire and 4-wire devices are wired differently. RTDs may be 2, 3 or 4 wire.	Must match the available supply and terminals, or wiring must be changed and documented.
Process connection	Thread type, size, insertion length and sealing method affect fit, leak-tightness and where the tip sits in the gas flow.	Must match, or an approved adaptor must be used without changing tip position.
Temperature rating of body and cable	Parts near hot exhaust see radiant and conducted heat well above room temperature.	Equal or higher rating. Never lower.
Ingress protection	Rain, washdown and condensation can enter connectors and housings.	Equal or higher IP rating.
Vibration resistance	DG sets vibrate; sensors mounted on or near the exhaust see more.	Equal or better, especially for rigidly mounted transmitters.
Connector type and pinout	Same connector shell with a different pin assignment is a classic error.	Must match, or a documented adaptor cable is needed.
Cable length and screening	Long unscreened runs near alternator cables pick up noise.	May differ with care; keep screening and earth it at one end as per the original design.
Impulse line material and routing (ΔP)	Condensate and soot can block or bias impulse lines. Material must tolerate exhaust gas temperature at the tapping.	Must follow the original or supplier-approved routing, slope and material.
Calibration certificate	Gives traceable evidence of the sensor’s accuracy at purchase.	Ask for it with every replacement.

Measured quantity and range

Range is the parameter most often treated as “close enough”. A 0–25 kPa transmitter is not interchangeable with a 0–10 kPa one just because both output 4–20 mA; the worked example below shows how far wrong this goes.

Check units too. One kilopascal equals ten millibar, so a 0–100 mbar sensor has the same span as a 0–10 kPa sensor.

Output signal: 4–20 mA vs 0–10 V and others

The common output types you may meet are:

4–20 mA current loop: robust over long cables; a reading below about 4 mA usually indicates a broken wire.
0–10 V or 0–5 V: simpler but more affected by cable resistance and interference.
Millivolt from a bare thermocouple, or resistance from a bare RTD.
Digital such as Modbus RTU, where register map, address, baud rate and data format must all match.

If your controller input can be set to several types, changing it is a configuration change that must be recorded and approved.

Exhaust temperature sensor: thermocouple vs RTD

As general examples, a type K thermocouple and a Pt100 RTD are not interchangeable at the controller: one produces a small voltage, the other changes resistance. Type K and type J thermocouples also produce different outputs at the same temperature.

For RTDs, note whether the original is 2, 3 or 4-wire; dropping the compensating wire introduces a lead-resistance offset. Insertion length matters too, since a shorter probe sits nearer the pipe wall and may read low.

Heat, water and vibration

Check the temperature rating of both sensor body and cable sheath against the real location. Our guide to cable routing near hot exhaust lines covers separation and support practices that apply to replacement cables too.

Compare IP codes using the IEC explanation of IP ratings. An outdoor unit exposed to monsoon rain needs at least the original rating, with the connector sealed as well as the housing. Our post on protecting an RECD from rainwater and condensation is relevant here.

Impulse lines for ΔP transmitters

A new ΔP transmitter on old impulse lines can inherit the old problem. Before fitting, inspect the hoses or tubes for soot blockage, cracking, hardening or kinks. Confirm that the lines slope as originally designed so condensate drains away rather than collecting in a low point, and that any condensate trap or drain is clear.

Also confirm which port is high pressure and which is low. Swapping the hoses gives a negative or near-zero reading that may be mistaken for a clean filter.

Step 3: Check the Controller Side Before You Swap

A sensor is only half the measurement chain. Before the new one arrives, record (with photos) how the controller, logger or BMS interprets the signal:

Input type configured for each channel (mA, V, thermocouple type, RTD type).
Scaling: what engineering value corresponds to the low and high ends of the signal.
Any offset or trim already applied.
Alarm thresholds, their units, and whether they are stored as engineering values or as a percentage of span.
Alarm delays, hysteresis and any latching behaviour.
Firmware version, if displayed.
BMS or remote monitoring point names, units and scaling (the BMS may apply its own scaling separately).

Who is allowed to change settings

Agree in advance who may change controller settings. On many sites, settings are password protected and the RECD supplier or a nominated service provider holds access. If a different contractor will do the work, read our guidance on whether a different service company can maintain an installed RECD and agree responsibilities in writing first.

In regulated environments such as pharmaceutical plants, a sensor change with rescaling usually falls under formal change control. The approach in our post on RECDs for pharmaceutical plants and change control applies well beyond pharma.

Whoever makes the change should record the before and after values for every setting touched. A simple two-column “as found / as left” sheet is enough.

Step 4: Write the Purchase Specification

Many replacement problems start with a purchase request that says “ΔP sensor for RECD, urgent”. Give purchasing or the supplier a proper specification instead. The same discipline described in our guide to writing a technical RFQ for an RECD project applies to a single spare.

Model RFQ text for replacement RECD sensors

Adapt the following to each sensor. Replace bracketed items with values from your identity card.

Item: Replacement [differential pressure transmitter / exhaust temperature sensor] for RECD tag [ΔP-01], installed on [kVA] DG set, RECD serial [number].
Reference: Original make/model [as nameplate]; RECD supplier part number [code]. Photos of original label and controller input page attached.
Measured quantity and range: [0–10 kPa differential] / [0–600 °C] or as confirmed by RECD supplier.
Output: [4–20 mA, 2-wire loop powered] / [type K thermocouple, ungrounded junction] / [Pt100, 3-wire].
Accuracy: equal to or better than [value or original specification], stated over full range.
Supply: [24 V DC] at sensor terminals.
Process connection: [thread type and size], insertion length [mm] for temperature probes.
Electrical connection: [connector type and pinout] or [cable gland with flying leads], cable length [m], screened.
Environmental ratings: ingress protection not lower than [IP code of original]; body and cable temperature rating suitable for [measured ambient at location] °C; suitable for engine-mounted or exhaust-adjacent vibration.
Documentation required with supply: datasheet, wiring diagram, calibration certificate with as-tested points, and declaration of any difference from the original specification.
Deviations: any deviation from the above must be declared in writing before dispatch, with the supplier’s statement of whether controller rescaling or rewiring is required.
Warranty: state warranty period and whether fitting by site staff or a third party affects it.

The “deviations” line is the most valuable one. It forces the vendor to tell you about a range or output change before it reaches site rather than after.

Step 5: Installation and Verification After Fitting

Fitting the sensor is not the end of the job. Verification proves the measurement chain still tells the truth.

Safety first

Isolate the DG set and RECD controller using your site’s lockout/tagout procedure; the OSHA guidance on controlling hazardous energy is a useful reference. Let the exhaust cool before touching probes or impulse lines, which stay hot long after shutdown.

Verification sequence

Zero check at rest. Engine stopped: a ΔP transmitter should read near zero (about 4 mA); a cooled temperature sensor near ambient.
Signal check at the terminals. Measure loop current or voltage with a calibrated multimeter and compare with the display.
Plausibility against a reference instrument. Compare with a calibrated portable manometer or thermometer at one or two loads.
Comparison with past readings. Run near the load of the “last good reading” and compare; differences should be explainable.
Alarm function test. Simulate a signal above the threshold and confirm the alarm appears locally and remotely, with the correct delay.
BMS point check. Confirm the BMS shows the same value and units as the local display, and that alarms reach the right people.
Update records. Record the new sensor’s serial number, calibration certificate reference, as-left settings and verification readings.

Instruments used for verification should themselves be calibrated and traceable. Our post on calibration records for RECD test instruments explains what to keep.

Note the change in the maintenance file and handover log. Our general RECD maintenance guide shows where this fits in routine upkeep.

Illustrative Worked Example: A Range Mismatch That Hides High Back-Pressure

This example is illustrative. The figures are chosen to show the arithmetic, not taken from any particular RECD or controller.

The situation

Original ΔP transmitter: range 0–10 kPa, output 4–20 mA.
Controller input: 4–20 mA, scaled 4 mA = 0 kPa and 20 mA = 10 kPa.
High ΔP alarm: 8 kPa.
The transmitter fails. Stores buy a 0–25 kPa, 4–20 mA unit from the same family because it was in stock. Nobody rescales the controller.

What the new transmitter outputs

A 4–20 mA transmitter has a 16 mA span. Output current is:

Output = 4 + 16 × (true pressure ÷ full-scale range)

At a true ΔP of 6 kPa on the new 0–25 kPa transmitter:

6 ÷ 25 = 0.24
16 × 0.24 = 3.84
4 + 3.84 = 7.84 mA

What the unchanged controller displays

The controller still assumes 20 mA means 10 kPa:

Displayed = (current − 4) ÷ 16 × 10

7.84 − 4 = 3.84
3.84 ÷ 16 = 0.24
0.24 × 10 = 2.4 kPa

The true value is 6 kPa, the display says 2.4 kPa. The error is (6 − 2.4) ÷ 6 = 3.6 ÷ 6 = 0.6, so the system under-reads by 60%. Put another way, the display always shows 10 ÷ 25 = 0.4 times the true value.

What happens to the alarm

The alarm triggers when the display reaches 8 kPa. Since displayed = 0.4 × true:

True ΔP at alarm = 8 ÷ 0.4 = 20 kPa

The alarm meant for 8 kPa now waits for two and a half times that pressure drop, while operators see a comfortable 2–3 kPa and believe the filter is clean. The DieselNet overview of diesel particulate filters gives general background on why loading must be controlled in filter-type designs.

The fix

The best fix is to buy the correct 0–10 kPa transmitter. If the 0–25 kPa unit must be used temporarily, and the RECD supplier agrees it is acceptable, rescale the controller:

4 mA = 0 kPa, 20 mA = 25 kPa.
Re-enter the high alarm as 8 kPa in engineering units.

Watch for one trap. If the controller stores the alarm as a percentage of span (here 80%), rescaling to 25 kPa would silently move the alarm to 0.8 × 25 = 20 kPa. Always check the alarm value in kPa after rescaling.

Also accept the loss of resolution. If both transmitters have an illustrative accuracy of ±0.5% of span, the original’s uncertainty is 0.005 × 10 = ±0.05 kPa, while the replacement’s is 0.005 × 25 = ±0.125 kPa. For a filter loading alarm that may be tolerable; for fine trend analysis it may not be.

Verification readings after rescaling (illustrative)

Check	Reference / expected	Loop current	Controller display
Engine stopped, zero	0 kPa	4.00 mA	0.0 kPa
Running, mid load	6.0 kPa on reference manometer	7.84 mA	6.0 kPa
Simulated alarm signal	8 kPa equals 4 + 16 × (8 ÷ 25) = 9.12 mA	9.12 mA injected	8.0 kPa, alarm raised

Check the middle row: (7.84 − 4) ÷ 16 × 25 = 0.24 × 25 = 6.0 kPa. Check the alarm row: 8 ÷ 25 = 0.32; 16 × 0.32 = 5.12; 4 + 5.12 = 9.12 mA. The display now agrees with the reference, and the alarm fires at the intended pressure.

OEM Sensor vs Equivalent: Weighing the Options

Factor	OEM / RECD supplier part	Equivalent sensor
Compatibility	Usually matched to the controller configuration and documentation.	Must be proven parameter by parameter.
Lead time	Can be longer if not stocked locally.	Often faster from local distributors.
Cost	May be higher.	May be lower, but add time for checking and rescaling.
Documentation	Usually referenced in the O&M manual and spare parts list.	Needs new datasheets and records.
Warranty	Less likely to raise questions.	May affect warranty on the sensor, controller or RECD; check the terms.
Future replacements	Same part number simplifies stocking.	Creates a second variant to track.

Warranty implications

If your RECD is under warranty, read the terms before fitting a non-supplier part or changing settings; some cover only work done by the supplier or an authorised party. Our post on RECD warranty terms for parts, labour and exclusions explains what to look for.

If you suspect the original sensor failed early and may be a warranty item, collect evidence before you throw it away. Our sibling guide on what to record when raising an RECD warranty claim lists the photos, readings and records that help.

Before deciding a high ΔP means a new filter element, confirm the measurement is genuine. Our guide to replacing an RECD filter core vs the complete assembly covers that larger decision.

Common Mistakes When Buying Replacement RECD Sensors

Ordering by appearance or connector type alone.
Accepting a “same family, different range” substitute without rescaling.
Rescaling but forgetting the alarm is stored as a percentage of span.
Fitting a type J thermocouple on a type K input, or a Pt1000 on a Pt100 input.
Reconnecting ΔP impulse hoses the wrong way round, or reusing blocked lines.
Accepting a lower temperature or IP rating to save cost.
Updating the local display but not the BMS scaling.
Skipping the alarm test because “the reading looks normal”.
Not recording as-found and as-left settings.

None of these mistakes ever justify running the DG set with the RECD bypassed or its monitoring disabled while waiting for parts. That is a compliance risk; if a sensor is unavailable, discuss interim arrangements with your RECD supplier and check your consent conditions with your State Pollution Control Board. The Central Pollution Control Board publishes the national RECD framework, and in Delhi NCR the Commission for Air Quality Management issues directions on DG set use.

FAQ

Can I use a non-OEM sensor as a replacement on my RECD?

Often yes, if every compatibility parameter matches or is deliberately managed, and the controller is checked. Confirm with your RECD supplier first, and read your warranty terms, because some warranties depend on approved parts.

What is the most important parameter to match on a ΔP transmitter?

Range and output signal together. A different range with the same 4–20 mA output looks compatible but changes every reading unless the controller is rescaled and alarms are rechecked in engineering units.

Is a 4–20 mA sensor interchangeable with a 0–10 V sensor?

Not without changing the controller input configuration, and only if the input supports both types. Connecting the wrong type can give false readings and may damage some inputs.

Can a type K thermocouple replace a Pt100 RTD?

Not directly; they produce completely different signals. Treat it as a design change and agree it with the supplier.

How do I know the new sensor is reading correctly?

Do a zero check at rest, compare against a calibrated reference instrument, compare with past readings at similar load, test the alarm with a simulated signal, and confirm the BMS shows the same value.

Should a sensor replacement be recorded?

Yes. Record the serial number, calibration certificate, as-found and as-left controller settings and verification readings, and note the change in the handover log.

Need Help Matching a Sensor to Your RECD?

If you are unsure which sensor your unit uses or whether a substitute is acceptable, Aceget can help you work through the details. Explore our retrofit emission control devices for DG sets, review our type approval and certifications, or contact the Aceget team with your sensor identity card, photos of the label and the controller input page. We will tell you what information is still missing before you place an order.

How to Specify Replacement RECD Sensors Without Losing Compatibility