Possible sticky injector simulator?

[Gordonjcp]

I'm having a weird one with my P38 with a Millenium controller and a pair of Pitagora simulators.

Although it's set up to start on gas and I pretty much never use petrol, I've noticed that the petrol gauge is slowly dropping - and the MPG reading on the trip computer often heads up over the 30mpg mark! Now I expect the trip computer to misread on gas, but that's ridiculous. Although every time I stick the diagnostics on I can see the lamba readings flipping up and down on both petrol and gas, the fuel trims are hard over towards lean which suggests to me that it's rich.

Could it be that the simulators are not actually dropping the injectors off, possibly intermittently? What would be a good way to test? One idea I had was to unplug the injectors from the simulator loom and see if it threw a code for open circuit injectors after a drive.

[lukie]

Had something very similar to this myself the other week same kit but on a disco, the odd thing when the LPG kit was flicking to has I could even here a solenoid click I was 100% certain I was on LPG although I was running pure petrol...

First thing to do is be 100% sure what fuel you are running, easiest way to do this is clamp the LPG feed pipe hose off to stop any gas going through, if the engine dies your on gas other checks needed, if the engine doesn't die your running petrol although your LPG control thinks your on gas, this is what the disco did, in the end I had lost an earth from the inner wing, if you car doesn't die with feed pipe clamped chance are you have a lose wire somewhere

[Gordonjcp]

I've already tried unplugging the solenoid, and the engine cuts out. That only tells me that it's not doing it on my driveway though

[lukie]

Ok so you know your running LPG that's good, driving down the road isn't going switch you from gas to petrol unless you have a loose wire in which case I would suspect to see the switch lights going on and off, my next suggestion being a green oval thing... Check the fuel filter for leaks first (hardly get changed during maintenance) if that's ok check the rest of the fuel line for leaks

[Gilbertd]

Pull the petrol pump relay. It's a yellow one at the front of the fusebox (nearest the battery) around the middle. You could pull that and poke the schrader valve in the fuel rail to get rid of any pressure then it definitely can't use any petrol. It could be that one channel on the emulator has gone down so you are running petrol on one cylinder.

Weird that your trip computer is showing 30 mpg though. That would suggest you are running rich on LPG so the petrol ECU is shortening the injector pulses to get the fuel trims right but with the Millennium it shouldn't be (as long as the lambda is flipping properly when on LPG). My trip computer showed me 14.9 mpg coming up to Scotlandshire this morning but that is because one of my lambda sensors is dead so it tries to run a default rich map.

[rossko]

It's not possible to extract any sense from the petrol management fuel trims while running a Millennium. The Millennium does its own thing, and that might produce a long term average of 50-50 high/low from one of the oxygen sensors. If the petrol management would like to see a 60/40 average ratio at this load and a 45/55 ratio at that load, it'll never be happy. Trims will wander.

Then we don't know how your system has been wired; its not that common but not exceptional either, to find the petrol management is being fed a simulated oxygen sensor signal whilst on gas.

If there is any misbehaviour - leaky petrol injector, misfiring, even the wonky emulator you suspect - that will mess about with trim values too, but its not much use as a gas diagnostic.
It would be interesting if you noted a significant difference between banks behaviour - but be wary of that simulated wiring thing.
It would be interesting if you clarified what the petrol trims settle down to when actually driving on petrol.

Usual emulator failure modes:
Poor connection or failure on the petrol injector side, causing petrol management to flag a single injector fault.
Poor supply/earth connection or total emu failure, causing 4 cylinders to dual fuel (you'll have 2 x 4-way emus) You'd be suprised how well it drives like that but the stinking exhaust is a big clue.
But both emus failing (due wiring) runs really rough.

Unplugging emulators will throw a full house on injector faults and tell you nothing new. (the circuit is still completed through the petrol injectors even whan active/inhibiting)

A wonky power/earth/RPM to the Millennium could cause it switch off from time to time, without being noticed. How in-yer-face is the switch/lights? Story time - a system that switched briefly to petrol only at one spot, bend at top of hill. Cause - dangling gas tank
feed wire melted on hot exhaust, but only swung against it to short out just there.

Don't forget the petrol vapour emission controls are still active - in this hot UK weather I would expect that to consume it slowly.

[Gilbertd]

in this hot UK weather I would expect that to consume it slowly.

Ross, he's in Glasgow, as am I at the moment, and it's bloody freezing up here! Well, I think it's freezing but the locals are telling me it's really warm at 20 degrees......

[Gordonjcp]

I must admit, I didn't think about the evap system. It's been really hot (>20°C) recently, so that could be it. I must admit I'm not totally clear on how the mixture adjustment works - surely if I screw the mixture screw in, the actuator will just open up to compensate?

[Gordonjcp]

Well, there's something far wrong with it. If the long term fuel trims are to be believed - and I can't entirely trust the software I used (EasyOBDII - crashy, half-complete and amateurish) then they stick at 24.22% rich on both petrol and gas. The short term fuel trims can be anything from 25% rich to 25% lean on both petrol and gas, with no real pattern. One bank can be fully rich and one can be fully lean. Again, I don't trust the software.

There's a severe flat spot and lack of power above 2000rpm, very slightly better on gas.

[rossko]

Are we talking GEMS? It's not-quite OBDII and most readers cock up reading long term FT, not to be believed at all. You might also note misplaced decimal point in airflow reading, and so on.

I would suggest a good drive on petrol only is in order, to see if the (more credible) short trims settle down. If they do, you'll be able to see/interpret what happens when you switch to gas.

[Gordonjcp]

The short term trims are all over the place on both petrol or gas. I'm currently looking around for intake leaks but not finding any.

<edit> Just went a couple of miles and back, absolutely no acceleration above 2000rpm, nothing until I get to kickdown. Idle fuel trims now staying steadish at +5% rich, if the diagnostic software can be believed, on petrol.

[Gilbertd]

Gordon, I'm home now (got back last night) but will be back up for 4 days from Tuesday. Not sure what shifts I'll be doing but when doing a late I won't be starting until 3pm so will have all morning free. We could have a joint poke and hope session if you've got time. Would also be interesting to check what your software (and my U581 code reader) makes of the readings on yours and mine as they are the same motor with the same LPG system.

[Gilbertd]

As I've got known good MAF and TPS on my spare car, I'll whip them off and bung them in the boot of mine so we have alternatives to try.

[Gordonjcp]

Richard - sounds great, I'm not sure what I'll be able to do with mornings but I'm about in the evenings.

[Gilbertd]

All I Know at the moment is that Wednesday I'll be driving up in the morning and then doing a late shift at the opening ceremony at Celtic Park, so that's out. Saturday I'll be doing an early but setting off to drive back as soon as I finish at 3pm. What I'm doing Thursday and Friday though I won't know until the day before.

[rossko]

Well we still don't know if it is GEMS. But it would seem to run closed loop to circa 200rpm, but refuse to go any further? I would suggest there is a Tesco bag in the air filter box, or a jammed backfire stopper flap in the mixer.

[Gordonjcp]

It's '97, so GEMS

The air filter has been replaced by a cone filter by a previous owner, but it seems clean enough. Removing it makes no real difference.

The MAF output voltage is about 1.4V at idle, 2.1V at 2000RPM, 2.4V at 3000rpm and about 3V at 4000rpm, but I can't find any details beyond the idle voltage of what it ought to be.

[Gordonjcp]

Since Gilbertd poked the ECU things have been much improved. I gave the MAF a bloody good clean and taped over about half the cone filter to "improve" the airflow by making it worse, if you see what I mean.

Then it started running rough and missing, and I said a rude word. Turns out that one of the injector simulators is indeed a bit intermittent, but I tracked it down to dry joints on the connector at the bottom and possibly on one of the plate-through holes for the earth lead. The major clue was that it would run just about as badly with the gas solenoid plugged in as not...

Next up, some sort of funnel arrangement to duct cold air towards the cone filter - well, at least until I can find a proper airbox in a scrappy.

[Gilbertd]

Bugger, totally forgot about that. There's a diesel P38 in a scrappy local to me and I could have nipped down there today and seen if the airbox is the same. I could have got it for you if it is but my brain thinks it's Sunday so I didn't think they would be open. No chance tomorrow as I'm setting off at 7:30 to head back up for another 4 days.

[Gordonjcp]

No worries, I'll pop along to Muirhead at the weekend and see if a diseasel one is the same. Now I need to work out why my alternator is growling. One thing after another...