MOT Emissions Failure: Find the Cause Before the Retest

An MOT emissions failure gives you a set of numbers and a red line on a piece of paper. It does not tell you what is wrong. The single most expensive mistake at this point is to guess — to buy a lambda sensor because someone on a forum said so, fit it, and fail the retest. The failure sheet plus the car’s own fault codes will usually get you a long way towards the real cause for nothing.

Start by reading the sheet properly

The refusal notice lists the measured values next to the limits the car was tested against. Which numbers appear depends on whether it is petrol or diesel, and the two tests are completely different in nature.

Petrol

The basic test measures three things at fast idle, somewhere between 2,500 and 3,000 rpm: carbon monoxide, hydrocarbons, and lambda. It then measures carbon monoxide again at natural idle. The basic-test limits are CO up to 0.2% at fast idle and up to 0.3% at natural idle, HC up to 200 parts per million at fast idle, and lambda between 0.97 and 1.03. A car that fails the basic test goes on to a full catalyst test using either the manufacturer’s figures or default limits set by the registration date, so a fail on the basic test is not automatically the end of it.

Diesel

There are no CO or HC numbers. A diesel is measured for smoke opacity while the engine is accelerated to maximum revs, and the limit depends on when the car was first used: 2.5 m^-1 for a pre-July-2008 non-turbo, 3.0 m^-1 for a pre-July-2008 turbo, 1.5 m^-1 for July 2008 to December 2013, and 0.7 m^-1 from January 2014. On a car fitted with a diesel particulate filter, visible smoke of any colour during the metered test is a Major defect on its own, regardless of the opacity figure.

What each petrol reading is telling you

High CO

Carbon monoxide is unburnt fuel that only partly oxidised. High CO generally means the mixture is too rich — more fuel than the engine can burn cleanly. Look at leaking or dripping injectors, excessive fuel pressure from a failed regulator, a coolant temperature sensor telling the engine control unit the engine is stone cold when it is not, a heavily restricted air filter, or an upstream lambda sensor that has gone lazy and is no longer trimming the mixture properly.

High HC

Hydrocarbons are raw, unburnt fuel. High HC points at combustion that is not happening properly at all rather than happening too richly: worn or fouled spark plugs, a weak ignition coil, a misfire, a vacuum leak, low compression, or burning oil. It can also be a catalytic converter that has stopped doing its job, because the converter’s role is precisely to finish off what the combustion chamber left behind.

Lambda outside 0.97–1.03

Lambda is a ratio: 1.00 means the mixture was exactly stoichiometric. Below 0.97 the car was running rich, above 1.03 it was running lean. This is often the most diagnostically useful of the three because it tells you the direction of the error, and it is directly comparable with what you can see in the car’s own live data.

Now cross-reference with the car’s fault codes

Here is the part most owners miss. The engine control unit has usually been watching the same problem for weeks and has an opinion about it, stored as a fault code — and there may be pending codes that have not yet lit the warning lamp.

A high-HC petrol fail alongside a stored misfire code is not two problems, it is one. A lean lambda reading alongside a lean-condition code narrows the search to unmetered air. A diesel smoke fail with a stored DPF differential-pressure or EGR flow code tells you exactly where to look. The failure sheet gives you the effect; the codes give you the cause the car itself has already identified.

The free checks worth doing before you spend anything

Some of these sound too simple to matter. They all cause real emissions failures.

• Was the engine properly hot? A car tested from cold, or after ten minutes idling in a queue, reads worse than the same car after a decent run. If you failed marginally, this alone can be the difference.
• Air filter. A badly clogged filter richens the mixture. It is a five-minute check and a cheap part.
• Oil condition. Old, fuel-diluted oil raises hydrocarbon readings because crankcase vapour is fed back into the intake. On a car that only does short journeys this is a genuine cause, not a technicality.
• Exhaust leaks. A leak before the sampling point draws in fresh air and skews everything, and a leak ahead of the upstream lambda sensor will make the engine control unit richen the mixture for no reason.
• Coolant temperature. If the engine never reaches proper operating temperature, it stays in warm-up enrichment for far longer than it should. That is a thermostat, and it is one of the cheaper parts on the car.

Watching it happen in live data

Static numbers on a sheet only tell you where the car ended up. Live data tells you what it was doing on the way there, and it is the closest thing to seeing the emissions test from the inside.

The parameters worth watching on a warm petrol engine are short-term and long-term fuel trim, the upstream lambda sensor signal, coolant temperature and calculated engine load. A healthy warm engine in closed loop shows the upstream sensor switching briskly between roughly 0.1 V and 0.9 V several times a second while the trims hover near zero. Long-term trim sitting well positive means the engine control unit has been adding fuel to correct a lean condition; well negative means it has been taking fuel away from a rich one. That single number frequently identifies the direction of an emissions failure in under a minute.

Freeze frame is worth just as much. When a code was stored, the engine control unit saved a snapshot of the conditions — road speed, engine speed, load, coolant temperature, fuel trims. A fault captured at 70 mph under load is a different animal from the same code captured at idle from cold, and the snapshot tells you which drive to replicate when you are trying to reproduce the problem. Live data and freeze frame need AutoScan Enhanced or above; the Standard edition reads and clears codes but does not show live values.

Getting the retest right

Two things trip people up here.

First, do not clear the codes and immediately drive to the retest. Clearing wipes the readiness monitors, and more importantly it wipes your evidence. If the fault is still present it will simply come back, and you will have destroyed the freeze frame that would have helped you find it. Fix, clear, then drive the car properly for a while and confirm the code has not returned.

Second, get the engine hot before the retest. A proper run of fifteen or twenty minutes, ideally with some motorway or A-road driving rather than crawling, gets the converter up to temperature and, on a diesel, gives the particulate filter a fair chance of a passive regeneration. A marginal car that failed cold has a much better shot when it is properly warmed through.

One thing not to do

Removing or defeating emissions equipment is not a route to a pass. Original emissions control components that are missing, obviously modified or obviously defective are an explicit Major defect in the inspection manual — catalytic converters and lambda sensors and EGR valves on petrol, particulate filters and SCR systems on diesel. It is also illegal for road use. The tester looks under the car.

The route to a pass is the boring one: find out what the car is actually doing wrong, put that right, and let the numbers come back into range on their own.