NCT emissions failure

quenching

I brought my '94 E220 W124 estate for its NCT last night and all was well with the exception of the emissions test, failed on High Idle, Lambda was 1.04 (limits 0.97 to 1.03). It shows as a pass on the CO at 0.07 (below 0.30 is a pass for 1994), however the tester explained that during the 5 minute emissions test the CO value was too high and the 0.07 figure was the reading in the last second of the test.

Its been suggested that the cause could be the lambda sensor, air getting into the exhaust system, or the catalytic convertor needs replacing. The car has covered 220k miles and was serviced 8k miles ago, so due in another 1k.

Is a service and some fuel treatment such as Dipetane likely to sort out the emissions values, or should I try the Dipetane on its own first?

When its retested does it have to pass ALL the exhaust emissions tests or just the one it failed on? The NCT report just shows the Lambda value as a FAIL/REFUSAL.

Plug

Its only a tiny bit off. Maybe there is a crack in the exhaust somewhere.

dr.fuzzenstein

High idle? As in idles at too high a rev? That wouldn't help.

quenching

The idle speed wasn't too high, I mean the "High Idle" section of the exhaust emissions test (3020rpm indicated on the NCT report)

RTT

Maybe have a look at this http://www.ncts.ie/pdf/nctmanual.pdf

Don't mean to highjack here but on a related topic I failed mine last night as one wiper blade was frayed slightly, (possibly due to the snow and ice). This didn't affect the cleaning of the windscreen as stated on p9 on the handbook. Anyway just have to bring it back for a visual inspection which is free.

I used to have one of those. As said above, check for exhaust leaks (block the back of the pipe). If the downpipe was removed at any stage the union from exhaust manifold is very tricky to assemble without leaks occurring. High lambda means too much oxygen in the exhaust gases.

If there are no leaks, your best bet is to check the error codes. It's possible to read them from the system if you invest on three banana plugs, a 5 volt LED and a switch. PM me and I'll send you details.

taintabird

Have it serviced properly and that should do it also give it a decent drive before you bring it back for test and that should do it

quenching

So, went off to a recommended exhaust repair centre and they resealed both ends of the centre section of the exhaust - it had been replaced with a new one last summer. They said there was a small leak at either end. Went back for the NCT retest and it failed again, same reason of high Lambda value, this time a bit higher, results of both tests are below if thats helpful. As far as I can tell from reading around a high Lambda value is almost always due to air getting into the exhaust system or a faulty Lambda sensor.

Is it possible that there is a leak in the exhaust between the manifold and centre section thats not easily spotted, or am I looking at a new Lambda sensor? Or anything else, exhaust guy said the catalytic convertor was probably fine.



1994 Mercedes E220 petrol, 220k miles.

First test:
Low Idle (0rpm)
CO 0.00% - Limit 0.5% - PASS
HC 9ppm - Limit 0ppm - Not Applicable

High Idle (3020rpm)
Lambda 1.04 - Limit 0.97 to 1.03 - FAIL
CO 0.07% - Limit 0.30% - PASS
HC 43ppm - Limit 200ppm -PASS



Second Test
Low Idle (770rpm)
CO 0.00% - Limit 0.5% - PASS
HC 9ppm - Limit 0ppm - Not Applicable

High Idle (2670rpm)
Lambda 1.05 - Limit 0.97 to 1.03 - FAIL
CO 0.21% - Limit 0.30% - PASS
HC 51ppm - Limit 200ppm -PASS

taintabird

could even be a cracked manafold

Scottie99

As someone said before, give it a good service. On the way to the the NCT centre add some dipitane, and drive the car hard so it up to temperature. I had a similar issue some years ago, the additive and a hard drive did the job.

biko

Take it to a garage that have a emissions tester.
Otherwise you'll get people guessing and fixing things and you have to pay for the NCT retest every time.

This is mainly for VW but a good read http://type2.com/s/?tag=exhaust

Is the exhaust totally leak free now (i.e. the engine stops if you block the tail pipe)?

If the exhaust is sound, your next step *really* is to read the fault codes (Merc hasn't got CEL). This can point out faults in oxygen sensor and air flow meter circuits etc and allows resetting the codes to exit from limp home mode (unlike in some other cars disconnecting battery does not reset fault codes in your Merc).

There's lots of variance between the two high idle results and the tester also said that CO was too high for most of test. Dipetane definitely won't fix the root cause of the failure.

I'm guessing that oxygen sensor will turn out to be the culprit if exhaust turns out to be sound. Broken air flow meter in my E220 caused really bad over fuel with black smoke and yours is the opposite.

taintabird

I would stills a cracked manifold or a leak in the exhaust system I would have that checked first before I would go replacing oxygen sensors or other expensive bits, I have seen the 124 series with cracked manifolds,

taintabird

taintabird wrote: »

I would stills a cracked manifold or a leak in the exhaust system I would have that checked first before I would go replacing oxygen sensors or other expensive bits, I have seen the 124 series with cracked manifolds,

+1 Cracked exhaust manifold can have a huge effect on emissions as that's where the oxygen sensor is situated and exhaust pulses easily draw in fresh air.

I think the cracked manifold syndrome is mainly seen on older (pre M111) petrols though. Could be wrong with that.

taintabird

samih wrote: »

+1 Cracked exhaust manifold can have a huge effect on emissions as that's where the oxygen sensor is situated and exhaust pulses easily draw in fresh air.

I think the cracked manifold syndrome is mainly seen on older (pre M111) petrols though. Could be wrong with that.

its actually surprising how many different makes and models can suffer from manifolds cracking

quenching

Thanks for all the advice so far, all much appreciated and all seems to point in the same general direction.

Yesterday I took it back to the exhaust place to have it checked again, they did an emissions test and it failed, but this time on the CO reading and passed the Lambda test. So they checked the exhaust and found another pinhole leak, fixed that and did the emissions test again (the engine temp was higher this time), passed all the tests this time.

So off to the NCT again next Wed, this time I'll sit in the car until the tester actually comes to take it in (last time it was sitting for 30 minutes as they were behind schedule), and as lots of people, and the exhaust guy, have said, get it really hot and rev it hard just before testing.

No cracks in the manifold I'm told so unless theres a small leak in a hose somewhere thats yet to be found then all leaks should be now dealt with! Fingers, toes and everything else crossed as it seems very hit and miss as to what temp and engine speed they test at as to what emissions values they get.

BTW, I checked the service record and it was done 5k miles ago and not due for another 4k.

kona

Isnt there a test for Lambda sensors, you use a Multimeter to check the resistance or something?.
Lean running could be a leak anywhere behing the airflow meter, so check your intakes here too.

Also the reason your two tests were different as they were taken at very different hight idle engine speeds.

dougie-lampkin

kona wrote: »

Isnt there a test for Lambda sensors, you use a Multimeter to check the resistance or something?.

Although they do work on resistance, i.e., too much oxygen causes low resistance, and a high voltage, and a richer mixture until it returns to range. I have read that putting a ohmeter across the lambda sensor will damage it. Then again I have also read success stories of people troubleshooting emissions problems by testing the lambda sensor like this so who knows, maybe it's an urban legend...

Best bet is to measure voltage across the sensor I guess, so you can see if you're burning too lean, which would cause too much oxygen in your emissions readings. Maybe this info is part of your OBD diag port, which could possibly also include the current fuel ratio and AFM reading, making fault finding a lot easier. I'm not sure, never owned a car with an OBD port, carbs ftw, especially the NCT emissions expemtion :pac:

shamwari

A healthy sensor usually oscillates between 0.1 & 0.8 volts each second at tickover. Three ways to test the Lamda sensor:

1. A voltmeter with a fast rise time measuring the above voltage variations.
2. An oscilliscope reading same
3. Read the sensors output on live data using a scan tool.

quenching

NCT'd again this afternoon and passed , Lambda value of 1.02, now have a valid NCT until April 2013. It would seem the problem was caused by poor fitting of the centre section of the exhaust that was replaced last summer, at least I now know where not to go for my next exhaust!

Thanks to all for the advice offered.