K series MOT fail - emissions

[mudguard]

Any advice on what to check/change?

K-Series 1.8 VVC EU3, since last MOT new HG and head skimming, replacement cat.

Car (and cat) were warm, straight from a 20 mile run. But the results were very similar on a previous attempt where it wasn't exercised much before the test. So perhaps less likely to be "cold cat syndrome".

Fast idle test

Engine Speed 2450-3050 rpm 

CO Max 0.3%         Reading  2.00%          FAIL

HC Max 200 ppm   Reading 155ppm      PASS

Lambda 0.95-1.09  Reading 0.94            FAIL

 

Second fast idle test

CO 2.38%

HC 100ppm

Lambda 0.93

 

Natural idle test

CO Max 0.5%  Reading 1.17%   FAIL

 

Further information: Lambda sensor has welded itself to the primary so cannot be immediately checked visually. But perhaps enough heat to weld it is enough to cook it?

 

[SM25T]

YHM

[revilla]

It is running very rich for some reason.

Do you have access to an OBDII scanner with live data? It would be interesting to see what the lambda sensor is telling the ECU.

Likely possibilities I can think of:

1) The lambda sensor is telling the ECU it isn't running rich. For example if the lambda sensor falsely reports the engine running lean, the ECU will enrich to compensate.

2) The lambda sensor is working fine but is being fooled into thinking it is running lean. This can happen if there's a leak in the exhaust ahead of the lambda sensor. Where is your sensor located? If it's in No 4 manifold header this is less likely but if you have something like a 4-2-1 exhaust with the sensor further down the system, any leaks in the joints tend to pull in air (they actually suck rather than blow!), which means the lambda sensor senses excess oxygen and thinks the engine is running lean.

3) You've got a faulty fuel injector. Two examples;

a) One injector leaking a bit leads to excess fuelling on one cylinder.

b) (Probably more likely if your lambda sensor is just one the header from one cylinder) The injector on that one cylinder is a bit clogged up. The ECU therefore sees it running lean and enriches to make that one cylinder run correctly. That leaves the other three running rich but the ECU can't see that because the lambda sensor is only monitoring one cylinder.

4) You've got a temperature sensor fault. The ECU thinks the engine is cold so runs it richer. Again this would show up on OBDII live data.

5) (Probably less likely on an ECU and lambda controller engine but worth checking) Your air filter is clogged up. I don't think this will be it, if the air flow was restricted the ECU would see the lower manifold pressure and compensate but worth giving it a clean anyway.

6) Did the car feel the same after the HG change? Sometimes cam timing errors can mess with the combustion and emissions, for example getting the cam belt a tooth out. Did you notice any change in performance? If nothing else shows up a belt timing check is easy enough to do.

[mudguard]

Thanks Andrew, that's very helpful.

Lambda is on no.4 primary in the engine bay, it's 4 primaries out to side of car then a collector there into 1 ahead of the cat.

The ECU could be thinking the engine is too cold. The thermostat was changed to the summer type due to the winter type not opening due to surge of coolant into the expansion tank due to historic removal of the ball valve in the inlet manifold - as previously discussed on a thread here. The thermostat now opens 100% reliably, but the temperature sticks at around 65, which is quite a bit lower than its previous normal running temperature of 80. But I believe plenty of Ks are on summer thermostats with no issues. The cat gets very hot as it should, but the old damaged one stayed cool.

Air filter was cleaned this year - seems to be in good condition.

The car felt considerably better after the HG work - it became able to pull in 4th gear from low speeds - and the overheating is 100% cured. However I think the improvement is mostly due to the old cat being in such a poor state that a lot of energy was being wasted in forcing the exhaust through it. Surprisingly, it did get through the 2019 MOT, a couple of months prior to the HG work, in this state.

The idle became quite a bit rougher post-HG, it is not so easy to start, is more prone to stalling, and occasionally stumbles and stalls on throttle release. I have cleaned the IACV, which was full of soot, and this is now improved, if not completely eliminated. I have kind of ignored these idle issues as TATDS after major head work, as the results were so good otherwise. But maybe it's a clue as to what's gone wrong.

About 6 weeks ago I had an occurence of misfiring. I checked the HT leads and found mild blemishes on the spring to no.1. That completely resolved the misfiring problem.

I'm not aware of the injectors ever having been cleaned - so that sounds like a good thing to do.

[revilla]

Have you actually had the head significantly modified (e.g. ported)? Or was a just a skim following the HG failure? If the second then it shouldn't be a case of TADTS.

With those idle and starting symptoms coming on immediately following the HG change I would double-check the cam timing for sure. May be fine but there's enough evidence to warrant a check.

You could also unplug the lambda sensor and see if the symptoms improve when it reverts to the base map, however this can be a bit misleading with idle symptoms as the ECU will probably detect the lambda disconnection as a fault and raise the idle speed slightly (it does this to prevent a possibly faulty engine from stalling) which falsely appears to "cure" the symptoms. Might be worth swapping out the lambda sensor to see what happens? I've got (used) spares if you want to try one.

[mudguard]

It was just a skim.

I don't have the car as I didn't want to drive it the long distance home following the failed test. It's now sitting in CC's workshop (they presented it for MOT) and whilst they're trying very hard indeed to be helpful they don't seem to be completely sure what to do next - K series emissions seem not to be a core part of workshop knowledge these days, which is understandable to an extent.

They want to change the lambda, and they have a new one available - which seems a good idea to me especially as the old one is partly melted. They also want to change the cat, which is going to be a long delay as none in stock, which I'm resistant to especially as it seems not to be a cat problem.

I think I'll ask them to clean the injectors as well as change the lambda, and then see where we are on emissions. I'll also ask them if they know how to check the cam timing on a K.

I don't mind spending a bit on trying to get it through, because if it can't pass, I really don't know what's going to happen to the car. It's quite sad really.

[garybee]

K series emissions are no different to any other modern petrol engine.  As yours is EU3 it'd be easiest to just get a fault code reader and work it out yourself.  There is a lot of information in generic OBD data and looking at fuel trims/lambda is often very revealing.  You shouldn't have to but good mechanics are so rare that that's the reality of the situation.

[opplock]

There's a thread devoted to this subject. I copied the link to favourites as my very friendly local garage had problems getting my k series through MOT last year. I had the car serviced 2 months later and asked them to investigate. No problem was found. 

https://www.lotus7.club/forum/techtalk/k-series-mot-lambda-limit

[mudguard]

Thanks for that Opplock.

From my reading of that thread, the "correct" wider lambda limit (0.95-1.09) was applied and the car failed even that by a whisker (0.94).

However, the big problem appears to be CO rather than lamdba; the car is measuring at > 6x the permitted CO maximum. 

Which would seem not to be a case of a slightly wrong limit applied?

[Markharrison2]

if you are running throttle bodies, make sure they are balanced. This caused my car to fail on emissions last year

[Andrew Gilbert]

I'd get it booked into a specialist who can have a look to see whats going on and maybe making any adjustments on a rolling road, years back mine did the same turned out to be an air leak and a slightly off map.

[Wrightpayne]

My '98 pre SVA K series self build is visual only. Are you sure it must have an emissions test?

Ian

[mudguard]

No TBs, just standard VVC (silver) ally inlet with single throttle for all 4.

Would anybody be able to recommend a specialist to do the sort of work being suggested?

[mudguard]

It's a 2002 kit build, first registered Jan 2003. As far as I can tell, it's always been subject to emissions testing. Certainly all MOTs in my ownership (4 years) have. And it's always passed emissions, until now.

[garybee]

What ECU do you have fitted?  Makes a big difference to who you need.

[mudguard]

Standard MEMS3.

[7_Malc]

Where are you based?

My 1800-K Roadsport was serviced and MOT'd at Ratrace Motorsport in Potters Bar earlier this year. Rob is an old hand at Sevens - and can steer most through the tests. He also rebuilds engines so is used to fault finding and rectification. 

http://www.ratrace-motorsport.co.uk

[mudguard]

Lambda sensor has been changed, and spark plugs. Old lambda was badly sooted, believed to be caused by a spark plug related misfire. The outgoing plugs were 2-year-old NGKs, replacing the - believed - original Rover plugs, which in 2018 were still in good condition after 16 years.

Runs better, but emissions are worse, and apparently this means the soot has gone into the cat and destroyed it, so needs to be replaced. The car got through an MOT with the old near-destroyed cat last year.

No mention of checking ECU temperature sensor, ODB2 diags, injector cleanliness, injector leakage, air filter health, cam timings, exhaust leakage.

To their credit they're willing to back up their convictions and try to borrow another cat as a test. If they're right I'll happily cough up for a new one.

 

[revilla]

I've a feeling the root cause may be being missed here ...

The lambda could have been not working properly because it was sooted up, leading to an incorrect mixture. But that begs the question, why was it sooted up so much in the first place? Sounds like it was already running rich and that sooted up the exhaust system and probably damaged the cat.

Would be a real shame to put on a new cat and have that get eaten.

I guess if they are going to source a new cat they should be able to see if that has cured the hydrocarbons issue, but make sure that mixture indications (lambda value) are sensible or it will be a short-term fix.

A standard K on standard management should easily get through an emissions test with lambda values very close to 1.0. The ECU has plenty of trim leverage to adjust the mixture, if it can't hit the target something needs addressing.

[garybee]

"Standard MEMS3"

No specialist needed then.  Although a car mechanic that knows how to do more than replace components until the fault goes away is almost rare enough to be referred to as a specialist  

[mudguard]

That's my feeling too Andrew. 

And as for easily passing the emissions test; I found the 2017 report - lambda 1.016, CO 0.07%, HC 29ppm. 

I don't recall ever seeing the 2018 numbers, and I don't have the 2019 numbers to hand (2019 was about 4 months pre HG work), but I remember thinking at the time that it had just scraped through on the lower boundary of the lambda, but not understanding the significance of that. So if I'm remembering correctly, something started to go wrong in 2018 or early 2019, which would make it completely unrelated to the HG work.

Perhaps, even, the cat destruction that was addressed during that work was nothing to do with the overheating, but this other, earlier problem, which has remained unresolved ever since, and will keep eating cats.

It's like a detective puzzle!

[aerobod - near CYYC]

The cat overheating is probably a consequence of overfueling, which will cause the cat to have to oxidise more fuel than it was designed to, causing it to heat up and eventually melt and clog.

[mudguard]

Like this:

[revilla]

i had almost identical cat damage from an oil control issue. Worn pistons led to glazed cylinders led to oil consumption and the oil was burning in the cat leading to that. The cat was too hot to touch two hours after shutting down.

[garybee]

Yep, misfire will do the same too eventually.  And you already know that you had that issue.