Has fuel, has spark, cranks but doesn't start.

[jerrit]

This is a very long story. The car is a 1997 900 turbo we picked up for $500 last year to race in lemons. It blew up an engine in its first race, due to oil starvation caused likely in part by a Boost Mobile SIM card and what I think was a relish packet stuck in the oil pickup screen. We found SIM cards all over the interior of this car when we stripped it out, I have no idea what the story is and I'll probably never know.

We got an engine from a junkyard that was supposedly known good, checked the oil pickup this time, did all the installation and fired it up. It spewed smoke from everywhere and emitted flammable liquids from the exhaust pipe, a mixture of gas and oil. It also managed to overfill the crankcase with gasoline. Compression was very slow to build up on the #1 and #4 cylinders, so after exhausting other possibilities I concluded it had bad rings and we got a THIRD engine.

Engine 3 exhibited similar symptoms with some differences, like the liquids from the exhaust weren't as flammable. We swapped the turbo which seemed to fix the billows of smoke, but it developed a stalling problem that eventually worsened to the point where it won't start, which is where we are now. An engine and turbo that seem to be ok in the little time we've seen them run.

Diagnosing an engine that cranks but doesn't start is supposedly straightforward. You need five things for the engine to fire: air, fuel, spark, timing, stoichiometric mixture.

1) We disconnected the idle air solenoid and attempted to start without touching the throttle, confirming a strong vacuum at the hose, occurring in time with the engine sounds.  This verifies that cam timing matches crank timing (not that we had any doubt), and also verifies that we are getting air into the cylinder (which may not have been happening if the solenoid were malfunctioning).

2) We traced all of the pins in the DI cassette connector.  There are 9 total.  1 is true ground, 7 go to pins in the ECU connector, and 1 hits an under-dash connector and eventually makes its way to alternator ground.  We traced them on the spare harness and verified continuity of all of the connections on our actual harness.  This rules out broken wires in the harness preventing spark.

3) We used the clear-plastic-hose test (pull out spark plug, add hose, hit starter, watch for "cloud") to determine that all four injectors are firing when expected.  Moreover, the hose bulged visibly when the cloud appeared, which indicates that we're getting fuel during the compression stroke, as expected.  We verified that the hose also smelled like fuel after each attempt.  This means that we have fuel and timing.

4) While being exceedingly careful, we hit the starter with the DI cassette off the car, and connected to a single plug at a time using a spark tester.  The bulb in the tester lit for every cylinder.  Given what I know about the t5 startup procedure, we'd expect to see four sparks per crankshaft rotation, which we did.


We have previously tried swapping these things on this engine:
ECU
DI Cassette
Spark plugs
Crankshaft position sensor

So, judging by this, it seems like the main thing remaining to check is A/F mixture, and we have no clue how we might accomplish that.

I'm really out of ideas at this point.

[sgould]

Welcome!

You may be getting fuel, but are you getting enough?  Is the fuel filter clear?

I'm not 100% up on the Trionic 5 cars.  But I'm not sure that the air mass is measured at startup.  I thought that standard assumptions were made on air and fuel.

Have you got a temperature sensor problem that is  saying things are hot when they are not?  So under fuelling.

It's all a bit of guesswork by me, and probably unreliable.

[phoenix]

The "safe" method of checking the DI and spark is to lie the DI cartridge upside down, with the plugs in. Wrap a good ground wire around the body of each plug and earth it. Then (with fuel pump fuse out) crank.

T5 takes it's timing signal initially from the CPS, but that can be 180 deg out so it fires all plugs together to begin with then works out the correct order from the the ionisation resistance across the plug gap. You need to have the correct plugs- NGK BCPR7ES11, gapped to 1mm.

Assuming the sparks are OK it sounds like you may have either too much or too little fuel. Too little could be caused by a failing fuel pump. Too much by the vac hose being adrift from the MAP sensor. If you disconnect the temp sensors it assumes default values, which should be enough to start it.

[jerrit]

I believe the temp sensor works properly. Before it began stalling it would read cold and gradually warm up to expected value. Also, when I plug my in to CAN and fire up t5 car pc I get reasonable values in degrees C.

Vacuum lines are all new and connected securely.

I have the right plugs, gapped correctly. Between spark and fuel, I'm more confident that the spark is correct. I've only determined that _some_ fuel is getting in the cylinders, not necessarily _enough_ fuel. Is there a good way to determine fuel pressure at the rail?

[phoenix]

It's difficult to determine absolute fuel rail pressure- you'd need an adaptor to fit inbetween the regulator and the rail. However if you could fit a pressure gauge in the supply line (opposite end to the regulator), that would be a good start. I would expect to see at least 5 bar.