MOT Emissions Issues - very strange

 My 1998 Jimny has done very little work in the nearly 2 years owned but runs perfectly. Not that good on fuel but around low thirties to the gallon. Smokes a little on start when not used for weeks and cold and £147k miles.
The first two attached pics are last years Emissions. A terrible first test and then around 10 minutes later a fantastic test.
The third pic shows this years(todays) test, another really bad result. Lambda very rich and a lot of incompletely burnt fuel. The tester tried for quite a while to get her through but unlike last year there was no magic improvement.
I've never seen a more contrasting emissions tests than those of last year, from terrible to pretty much perfect but they would be genuine. I get on well with these guys but no way they did anything clever here and it failed disappointing on front numberplate; tester did not like the stick on which was unexpected.
Think I have found the obd1 plug behind the glovebox. Appears to have six ports and four wires - black, red and two white. Guessing the jumper is black(earth?) to red? Perhaps that will give a clue.
 bit gutted really and hoping someone can shed some light on this.           

Attachments have not worked

Test 1 6/7/23 - 11.39
fast idle
co 1.6
HC 86ppm
lambda 0.962

Test 2 6/7/23 - 11.49
fast idle
co 0.02
HC 6ppm
lambda 1.003
natural idle 
co 0.5

Test 3 3/7/24 - 13.00
co 1.61
HC 177ppm
lambda 0.967   

Low lambda means rich mixture. 
It could be as simple as the thermostat is missing/stuck open and the engine is actually cold and running "on choke", but warmed up enough last year for the second test t o drop back to normal.
Or the temperature sensor that feeds the ecu is dodgy and it thinks the engine is colder than it is.

You want live data on the temperature sensors, Coolant & IAT (inlet air temperature).

It could be the pre-cat oxygen sensor gone sluggish or even the MAP sensor on top of the manifold, but as it magically got better last year with a bit of extra running time (and the weather was a lot warmer) I'd lean towards a temperature sensor reading cold, (either a faulty sensor or an over cooled engine)

The high CO is the fault of the low lamba, it could be low catalyst efficiency, but I'd fix the lambda first- and I'd expect the post cat sensor to trigger the EML on emissions if the catalyst was bad

Thank you for the reply, much appreciated. I am an MX5 specialist but know very little about my Jimny and no reference material. Too busy dealing with MX5 issues to put this right but now have to pull my finger out to get this 1998 Jimny 1.3 JLX through its MOT.

Agree with what you say, so a sensor problem, probably temperature sensor but maybe lambda sensor. I intend to replace both but really struggling to get the lambda loom connector plug undone around the back of the engine. Is there a trick to undoing this please?

This is a 1998, so no post cat lambda and no OBD2. Can anyone advise on how to jumper the plug behind the glovebox to get an OBD1 flash sequence with ignition on, I was going to try black wire connector to red as a guess but would appreciate assurance that this is correct please?

Also this car has a three pin temperature sensor on the thermostat housing but does not appear to have another sensor at the back of the engine as expected; is that correct?

The ambient temperature sensor on the airbox appears to be the same as the one on the MK2 MX5. I usually have at least one knocking around but not at the moment. Is it worth replacing that too?

Yes, the IAT is the one on the airbox, the ECU guesses the  initial fuelling from the stored map based on the rpm, manifold pressure and air temperature, then it tweaks it to get the correct remaining oxygen level from the oxygen sensor reading.

I'd expect it to have high values of fuel trim if the sensors were out of calibration- if they fail and it can't achieve what it thinks is "correct" to match the sensors then the EML is triggered.

When the engine is cold it runs an enrichment fuel map, which is why I'd start with checking that the engine isn't physically overcooling, and the coolant sensor.

AFAIK, the engine coolant sensor (ECT) is the one at the front of the engine

I can never part any of the connectors without breaking them either!  I'm always searching for affordable Magic Tools that can open them.

I've never messed with the G series engine, according to the manual there is a separate connector hidden somewhere behind the dash to get the MIL to flash out the codes, I wouldn't expect any (except 12 = everything is awesome..) if a sensor is reading incorrectly, they have to be obviously broken for the ECU to realise. 

Oh- you have checked the air cleaner? A blocked one will drop the manifold pressure  and make it think the engine is working harder.

 

 
 

Thank you again facade, appreciate your continued interest in this, especially as I posted in the wrong section. Sorry Mods, please move the thread to the right section.
The old skool diagnostic does a repeated 12, so I think that lets the ambient temp sensor, and MAP sensor off the hook. Possibly the coolant sensor too but I guess only the heater circuit tested on the lambda.
I'm going to replace the lambda sensor, coolant temp sensor, air filter and thermostat. Parts all ordered and hopefully when they get here will have recovered from Covid, into 4th day now.
One last question please regarding required drain down to fit thermostat and temp sensor. Is it best to drain down using the radiator cock or bite the bullet by undoing the thermostat housing. Either way hoping to dump coolant in a controlled way into a container and assuming it is clean add it back after parts replaced.

As anticipated, 12 means as far as the ECU is concerned everything is tip-top, and it is fuelling correctly according to its measurements.

It has no way to tell if a sensor is giving duff readings unless they are outside the realms of possibility (it is more a test for short circuit to ground or open circuit due to bad connections), so if the ECT reads low it assumes the sensor is right and the engine is cold.

You can check the IAT & ECT sensors with your multimeter, the manual gives the curves of resistance vs temperature, so you can tell if they are miles out. Just measure at room temperature, between 2K & 3K for the IAT, and a little higher for the ECT.

 

 


Otherwise, firing the parts cannon at it is going to be the only way without paying for a mechanic to diagnose & fix- which will probably be by firing the parts cannon, and charging you a mark-up + labour, unless you can find someone with the old diagnostic tool that can read live sensor data.


I had an M series, but I'd certainly drain as much coolant as possible into a bucket from the radiator cock, and hope it empties the thermostat housing. Take the radiator cap off first, so it leaves the expansion bottle full.

 

Well the good news is she breezed the retest
Fast Idle test...........
CO - 0.01
HC - 12 ppm
Lambda 1.007
Natural Idle Co - 0.27

Natural Idle CO aside, as marginal, the other figures pretty much perfect.
What a huge relief.

Well done!

The high idle CO is probably low catalyst efficiency when it cools down a bit at idle, I wouldn't worry about it until it fails an MOT again.

Can't think of anything else, other than cat that would cause the high CO here but surprised there is none on fast idle.
I watched him do the test and he did not mess around, thrashed the hell out of it on fast idle and then dropped to natural idle very quickly.
The cat is aftermarket and as it's a 1998, unlikely to be type approved.
I find the extremely high HC allowable, hugely out of proportion to the low CO for an MOT pass.
A company I deal with that sell aftermarket cats, BM Catalysts, are very conservative and generous with their warranty returns but will not accept a return where the HC level is 60+ppm. Kind of puts the absurdly high MOT pass level of 200ppm into lala land.