Technical Temperature-sensitive dying at idle

[MikeSalt]

Been a bit of a rough few months for my car and it's developed another problem, which is very odd. From cold, or at opertaing temperature, the car runs perfectly. However, at intermediate temperatures, it keeps dying when I let it idle. It is usually accompanied by an odd shake, which I presume to be a misfire. No EML lit, and the spark plugs are only a couple of months old. This evening, I'll clean up all the electrical connections (coil on plug since it's an HGT) and inspect the plugs but does anybody else recognise these symptoms?

 

[fingers99]

I'd see if there's a code anyway.

It's almost as though one of the temp sensors is lying to the ECU which is cutting back on the fuelling.

Live readings may give a clue.

 

[MikeSalt]

Fingers, do you know which temperature sensors I should be looking at? I find the one that controls the temperature dial on my dashboard to be quite believable. Is there an intake air temperature sensor, or even a seperate coolant sensor that the ECU talks to? I've got MultiECUScan, so I'll take a look at the live readings. Now might be a good time to buy the registered version so I don't have the 20 minute limit.

 

[fingers99]

There's an air temp sensor combined with the MAP sensor on the inlet manifold (possibly mucky from blow by?). AFAIK, the coolant temp sensor (green one, on the thermostat side of the inlet mani) sends signal both to the ECU and to the gauge on the Mk2s (but I may be wrong regarding the HGT).

 

[MikeSalt]

Does the petrol HGT have a MAP sensor? Anyway, I'll bunk off work early today, hook it up to diagnostics and take it from there. Thanks fingers.

 

[Ziggy122]

Had a quick look on a 2004 1.8 16v (so that makes HGT) It shows no MAP, but only a MAF (mass air flow) Instead of MAP (Manifold Absoulte pressure)

Ziggy

 

[MikeSalt]

Thanks Ziggy, looks like I might have to just hook up MES and see what parameters are available to me. I've got a digital multmeter with temperature probe, so I can check the air temperature at the filter casing inlet (allowing a little flexibility to account for air being heated slightly through the action of traversing the filter).

I did have a problem similar to this before, which went away after changing the spark plugs. It was never this severe though, and didn't come on as quickly. It was all fine Friday, then coming back from helping a mate to do some gardening it started playing up. It was a warm day that day but my temperatures according the dial was just dandy.

 

[Ziggy122]

You could forwell try disconnecting the MAF
this will put the EML ON

But it should run on a base program, and you should see it sort of return to a near like how it should be

If its still dying, something else is a fibbin

Since you have MES - you can check the Coolant temperature sensor as its working

Ziggy

 

[MikeSalt]

Ziggy, it is indeed an Air Flow Meter that is fitted to the HGT (Hitachi Part Number AFH60M-15). Unfortunately, I'm having a nightmare of a time to try to find this part in isolation. Might have to get a whole new (well, used) throttle body and transfer the bits over.

 

[MikeSalt]

Right, I have a plan of action now. First of all, I'll get home. Thankfully, I drop straight onto the motorway from my place of work, so the engine will have fully heated by the time I have to let it drop to idle. Next up is to let it cool in the garage for a bit and unplug the MAF. When the coolant temperature needle is halfway down from operating temperature, I'll start it up again and see if it dies or is improved. If improved, I'll inspect the MAF and attempt a clean with electrical contact cleaner, or replace if that doesn't solve it. If it doesn't improve things, I'll hook it up to MES and start the detective work.

 

[s and b]

your symtoms suggest coolant temp sender, if this is reading as cold to ecu then fron cold it will start and run fine
once engine warmed up it will still be telling ecu engine cold so it will overfuel rather than underfuel
this would cause wooly running once hot as co will be at near 10
this wont usually bring on an eml until cat killed and second lambda sends light on

a co read with a 4 gas analyser the type used at mot stations would prove or disprovewhat im saying

or even a good gunsons digital works fine here and is much quicker thasn firing up the monster in the corner

 

[MikeSalt]

Hi s and b,

I have the coolant gauge on my instrument panel, and that is showing a normal warm-up and stops exactly in the middle of the dial as it should. Also, once up to operating temperature, the car runs just fine.

Cheers,

Mick

 

[s and b]

Hi s and b,

I have the coolant gauge on my instrument panel, and that is showing a normal warm-up and stops exactly in the middle of the dial as it should. Also, once up to operating temperature, the car runs just fine.

Cheers,

Mick

no thats nothing to do with the cts
its usually on the inlet manifold into the water jacket
sorry no idea on a hgt

 

[MikeSalt]

no thats nothing to do with the cts
its usually on the inlet manifold into the water jacket
sorry no idea on a hgt

I think that both the ECU and the dial take their coolant temperature readings from the same sensor. However, if I hook it up to MES, I'll be able to confirm this for sure.

 

[MikeSalt]

Am I right in saying that the lambda sensors don't switch to closed-loop operation until the engine is at operating temperature? I wonder if the reason the problem goes away once up to operating temperature is that it relies more heavily on the lambda sensors rather than the MAF at that point, so a lying MAF gets ignored? No error codes in MES, and only the two parameters that mean anything in this context: engine temperature and air mass. Got a big 24" fan blowing on it to cool it all down so I can try disconnecting the MAF.

Assuming this article is correct, I think I've just answered my question. MAF is about to have the big finger of suspicion pointed at it.

Extra edit, got the temperature sensor against the thermostat housing, next to the coolant temperature sensor, and the values are within 9 degrees of each other (73 vs 64). This seems reasonable since I expect a sensor touching the casing to be a touch cooler than a sensor immersed in the coolant.

 

[MikeSalt]

Disconnecting MAF made it worse if anything. Quite a bad shake too, so I'm going to investigate the spark systems. Odd that they aren't triggering an EML, although I suppose if the engine dies quickly enough, it doesn't detect the fault.

 

[MikeSalt]

Plugs were quite sooty actually, possibly a by-product of the blowing exhaust last month? Cleaned then up anyway and it has improved things a bit, but fault still there. I might just chuck in a fresh set of spark plugs to see what happens because otherwise I'm completely stumped! By the way, MAF reading increased with engine speed so I think that's working, and the platinum wire looks clean too.

 

[s and b]

you could try cleaning the maf off the car with some carb cleaner as a short term fix (the little bionic wire bit)
do not buy a chinese copy oft tinternet
you need to go to a proper motor factor like pages and buy a proper bosch one
beware theres a million copies out there so heed my advice if you do end up needing one
any car ive ever had always put eml light on when maf went duff mind
one other thing many cars go to default mode factory settings if maf disconnected so if you have the facilities to clear any fault codes then its free to try this innit

 

[MikeSalt]

I am completely stumped on this one. One thing I forgot to say was that I observed the lambda sensor status during one run to see if switching to closed loop solved the issue. Unfortunately, it does not, it seems that the only fix is to get it right up to operating temperature. I tried disconnecting the MAP, that made things worse. I've verified (within 15%) that the coolant temperature sensor is reporting the right values to the ECU via a digital thermometer and MES. I cannot think of anything else that would affect just the warming up period (not immediately after cold start, and not once up to operating temperature). The thing that's really baffling me is that there is no EML.

 

[s and b]

connect an earth wire to the exhaust manifold bolt and connect to battery earth
does lambda work better then

in fact swap lambdas round too just for the fun if they undo