First of all some background. When i bought the P38, it came off ebay, with one listed fault, a bad crank sensor. Apparently the engine would run for about 15mins, then die. The bloke told me the rear seal was leaking and that had contaminated the sensor. Sounded fishy but the cutting-out-when-warm i'd seen before with crank sensors, so i figured it probably just needed a new one.

When i got the truck home, i put a new crank sensor in it, but discovered the one i took out really wasnt all that old. Hmmmm!

So, take the P38 in for its MOT today, Drove it up and down the dual carridgeway probably 6 or 7 miles to make sure it was all running fine, then dropped it in for its test. Went back to collect it and the boy tells me he couldnt test it as it wouldnt run, kept cutting out. Great.

So i drove it home, just. Last half mile it was acting very strangely, missing and cutting out and not wanting to idle, before finally stopping right outside the house. I managed to coax it onto the drive, and sat there pondering what was going on. Tried ot start it a few times and it would start, rev up a bit then you could hear the engine had simply stopped firing. Adding more throttle would sometimes help it pickup but it would fart about for a few seconds then die.

First thing i did, somewhat randomly, was pop the petrol cap off, to be greeted with a loud and pronounced hiss. Hmm, vacuum in the tank perhaps? I started it, and it fired up perfectly and dropped to a nice idle, revved fine, ran fine, all seemed fine. OK i thaught, somethings wrong with the tank breather. But lets be sure, i'll take it for a drive and see if i can repeat the same fault.

Drove to the petrol station, no dramas, removed cap and there was a small hiss. Put a tenner in and drove home. About a quarter mile away this time, it just dies. I coast into a side road and try to restart, but nope it starts revs abit, pops farts and cuts out. OK i thaught if its a fuel delivery issue, i'll see if it'll switch over onto gas. Started it, revved it a bit and hit the gas button. Once gas came on it suddenly sounded much nicer, revved up nicely, but it was short lived, as a few moments later it starts missing on gas as well before stalling.

Scratching my head and wondering how i'm going to get it home, i sat about a bit, tried it a few times, then after some pondering, unplugged the MAF... It fired right up. Waited a moment to see if it was shortlived, but nope, kept ticking over. But hmm! its done exactly this before with the petrol cap.... So set off home and it drove home just fine. Much like the trip to the petrol station, it drove with no issues and all worked perfectly.

Back at the house i ran it on the drive for 10 minutes with no issues.

Sooo, whats going on? MAF failure? Fuel tank cap a red herring?

Single point or sequential gas Aragorn?

Sequential system with a Bigas ECU.

Went out just now as its been sat a while and decided to see if it would start. With MAF plugged in, it catches then immediately stalls. Tried a couple times, same thing. Unplugged the MAF again, this time it catched, revved a bit, almost died, revved a bit, almost died then shot up to 2500rpm and ran fine. Put the MAF back on, same thing as the first time, an initial catch then just dies. Unplugged it again, and again, it repeated the catch-rev-almost-stall-rev-almost-stall-running fine cycle that it did the first time. So Maybe it is the MAF? Or maybe the ECU's just getting confused with sensors appearing and disappearing?

I've taken the MAF off and cleaned it. Trying to diagnose this stuff with no MOT is particularly annoying as i cant really drive it around and see whats happening. At least today i had an excuse as i was driving to and from the MOT station etc.

With a sequential if it won't run right on petrol, it won't run right on gas either. My SE did just the same, it would start then immediately die. I found that if I kept pumping the throttle to keep the revs up, eventually it would idle normally. Swapped the MAF sensor with the one from the ex-plod and it was back to normal. The GEMS isn't as fussy about pattern MAF sensors but ideally you want a genuine one.

Cleaned the MAF last night, not that doing that ever actually works but it was pretty grotty. This morning i refitted it, plugged it back in and started the engine. Fired right up and idled and revved completely happily. So whatevers going on is very clearly temperature related.

Coolant sensor checks out fine on the live data, as does intake air temp, so its gorra be something that doesnt like getting hot and is stopping working...

so its gorra be something that doesnt like getting hot and is stopping working...<br>
Or a crappy corroded connector somewhere in the MAF circuit perhaps?

Ok just took it for its MOT test again.

Failed on emissions. YEY!

1st Fast idle was 1.51% CO, but with 69ppm HC and lambda at 0.976 - FAIL on CO

2nd Fast Idle was 1.22% CO, but with 53ppm HC and lambda at 0.982 - FAIL on CO

Natural Idle Test was 0.16% - PASS.

So its running fine at idle, but getting rich with revs, which does suggest to me a MAF problem?

Looking at the STFT's on Torque, at idle the STFTs are both slightly positive, 5-6%, however at 2500rpm, bank one drops to -10% (with its lambda switching sensibly), and bank 2 drops to -25% (with its lambda stuck at 1v) So perhaps a bank 2 lambda issue?

Bank 2 lambda does switch normally at idle though...

What are you doing testing the emissions on petrol? It should be tested on the fuel it is presented for test running on so if it is running on LPG it gets tested to the LPG limits which are 3.5% CO, 1200 ppm HC and no lambda check.

What Gilbertd says! Get it tested on LPG... <br>
Re your STFT and Lambda readings, I suspect you're using a generic device (ELM or something) to get your info (going by the lambda stuck at 1v.) <br>
You've got Titania sensors on your 94 (0-5v). Generic devices usually assume a Zirconia sensor so return a 1v peak reading. <br>
I never managed to get sensible lambda data using generic readers on mine (upside down graphs and that sort of thing) and it wasn't until I got first the Hawkeye and then the Nanocom that I got data that actually made sense!

Hmm didnt know that about LPG? I've been running it on petrol because i'm not sure the LPG works properly, and common wisdom seems to be to get things working right on petrol before trying to get LPG working. It needs fixed either way, as it clearly should be perfectly capable of passing the test!

Can you find me a definitive source for those readings? I've looked in the MOT manual and it doesnt mention different limits for LPG systems, only reference it makes is that hydrocarbon limits on LPG need to be converted to account for propane vs hexane.

Orangebean, yeh i realise its a generic reader, but lambda on bank one cycles nicely between ~0.2 and 1v, the other bank is just stuck at 1v, but only while revving it at 2500rpm. Drop to idle and they both drop to 0.2v and the STFTs swing from -10 and -25 to +3-4ish and then start switching normally. So the readings make some sense.

I've realised since posting above that i've put it in for the bloody emissions test without the MAF connected. I mustve forgotten to connect it this morning :( However i plugged it in and it wouldnt run at all, so there would defo seem to be something up with the MAF.

Could it be a faulty injector simulator, leaving the fuel and gas injectors on?

Flowchart from the testers manual is here http://www.motuk.co.uk/images/730_newnoncat.pdf . Also, if it is presented running on LPG, it does not need a full CAT test so it doesn't need to have cats fitted (see http://www.motuk.co.uk/manual_710.htm ). It may be that as it has an LPG conversion on it, the cats have been gutted so are only empty boxes.

However, running it with no MAF would make it drop into a failsafe default which will be running rich. The Titania sensors show 5V when lean and 0V when rich but some generic readers invert the lambda sensor results so they display a comparable output to a Zirconia sensor with 0V being lean, while others just change the voltage. So, depending on the reader, 0V might be lean but might be rich. Better to hang a multimeter on the output so you can see what it is actually doing. Sounds like you've got one lazy sensor, or the heater isn't doing it's job, so only gives a reading when it gets heated enough by the exhaust gases.

gordonjcp wrote:

Could it be a faulty injector simulator, leaving the fuel and gas injectors on?

Unlikely, it would be running so stupidly rich you'd smell it and the emissions would be through the roof.

Dug out my LPG emissions result printout from my MOT in August <br>
CO % Vol..... Min=0 Max=3500 Actual 0.599 <br>
LPG ppm Vol.. Min=0 Max=1200 Actual 58 <br>
Confirms Gilbertd's info above <br>
EDIT- which he's already done above. I'm typing slowly tonight!

fair enough, thats good to know. Clearly i want to fix it "properly", but not being able to properly test drive it gets quite annoying and makes it difficult to figure out whats going on.

I think the cats are fine, as otherwise it wouldnt have managed the 0.16% CO at tickover. Typically a decatted car will run around 0.5%. Its usually impossible to get below the 0.3% cat limit without actually having a cat.

I'll get this MAF swapped out (ordered a genuine used one from ebay) and see what happens. If its still being a shit, i'll tell them to switch it onto gas and run the test on that just to get it thru while i figure out whats going on.

Getting pissed off with this bloody car now!

Took it in today for another emissions test having replaced the MAF sensor. Looking with Torque, the STFT's are within a normal range (they looked a bit wonky with the old MAF)

He said "its getting better" but couldnt give me a print out, as apparently the machine "choked" and couldnt complete the test. I asked what the numbers were, and all he could tell me was 0.35% CO, then the machine choked out and aborted the test. I asked what would cause it to choke, and he said usually missfiring or overfuelling. Hmmph. I've sat revving it on the drive way and cant detect any hint of overfuelling or missfiring. the exhaust smells clean and not rich etc etc.

Not sure what to do now? Do i stick some new lambdas in it? anybody got any ideas?

Oh, and i asked them to test it on LPG, and they refused saying LPG messes their machine up or some nonsense. Not sure that would help anyway as the STFT's on LPG are all over the shop, clearly the LPG system needs some TLC. Yes, i could go somewhere else, but clearly, the car should be able to pass a petrol emissions test, and the other MOT places in the area are worse than this place.

STFT's may be all over the shop but I bet it will be below 3.5% CO no matter how bad it is. My tester didn't bother testing mine last time as he said that if the emissions were far enough out for it to fail it wouldn't be running!

I agree, it should be able to pass the petrol test but with the LPG kit installed it shouldn't need to. If it's been mostly run on gas you've no idea on the state of the injectors or fuel pressure regulator if they've been dormant for a long time. I spent ages getting mine to run reasonably well on petrol just in case I ever needed to use it (although the petrol that had been in the tank for over a year didn't help).

Theres a note in the history from the last LPG service (5 years ago...) about a fault with one of the LPG injectors "fluttering", but seems it was passed over with an "ignore it until it gets worse" type comment.

When running on petrol the short term trims look sensible enough. Switch to LPG and one bank goes to -25% and the other bank to +25% and it just sits there... Certainly not happy! In my mind, if i know its running properly on petrol, and the readings etc can be trusted, then i know i can trust the info to sort out the LPG issues.

I've noticed that even on petrol the short term trims are perhaps acting oddly, at idle one bank shows mildly positive, and the other bank is slightly more negative. Not a huge amount though. say -5% and +3%. However if you raise the engine RPM's, it seems to get wider. at 3000rpm, the positive bank has increased to maybe +9% and the negative bank is at -11%ish.

I've ordered a pair of new lambdas for it. Also thinking about pulling the fuel injectors and having them cleaned. I'm also going to re-check the spark plugs condition, and pull the ignition coils out and ensure they've not "burst".

ok, faught rediculous winds and pulled the spark plugs out just now. Certainly something odd going on!

https://www.dropbox.com/s/azls39kcog9xjx4/2016-09-27%2017.45.00.jpg?dl=0

These plugs were new, probably less than 10 miles ago. Albeit most of that distance done with what would seem to have been a fualty MAF sensor.

The odd bit, is they're all quite different.

Starting with bank1 they all look "okish" with the caveat that 5 and 7 both have soot covering exactly half of their plug ceramic.
Bank 2 is where it gets all weird. Cylinder 2, is the same as bank 1. Cylinder 3 is dark and sooty. Cylinder 5 is darker and sootier, plus also looks oily. Cylinder 7 is VERY sooty and oily with weird deposits on it (i presume burnt oil).

Thoughts? I'm thinking perhaps faulty injectors, and/or some sort of mechanical issue? The oil could simply be valve stem seals, or something worse.

On my first P38 which was a GEMS, I found they had put the LPG inlets for the front and rear cylinders backwards, as they placed them far too high up on the inlet manifold before where the tracts swap over underneath. So both the front two and rear two cylinders had the injectors reversed. It ran horribly until rewiring them to get them injecting gas into the right inlet tract at the right time.

Whereabouts are you?