Ideas on a random no-start condition?

[LateFan]

(EDIT - This is an 86 eta engine)

I have more homework and testing to do per The Jim Levie Basics List, but we're a bit stumped by the random no-start situation we have. My son & I have it fairly close to useable as a DD. It seems to be following an odd/even day model for starting.

We've replaced every piece of rubber on this engine, including the brake booster check valve and grommet. All new service items. Plugs are cleaned and gapped. Injectors refurbished by member MEPEH. Tested the TPS, ICV, etc. Replaced the valve cover gasket and plugs where we found a massive air leak.

It starts instantly and runs smoothly, although there is a high idle and some cycling. It hasn't done the cycling in a couple of weeks - just the high creeping idle. I pulled the oil cap off while idling to see what a major air leak does - it drops to 500 and runs roughly, tries to quit.

On odd days, it will not start. With no pedal, it just cranks (fairly fast - good starter and new battery / alt.) With some pedal, it will try to kick but stumble and die. With WOT, it will start and rev for an instant, then stumble and die. No amount of throttle feathering will keep it running.

To me it feels like flooding. After it stumbles a few times, there is gas smell. It spins fast, like maybe the cylinders are washed down (?).

Could I be fooled by symptoms? Could we have random low fuel pressure from a failing pump? It does some random whining you can hear above idle (the pump in front of the axle). But it never quits or stumbles like running out of gas.

It feels like fuel, not spark to me - it runs really well and idles for an hour without problems. We have really improved how our $1000 project car runs, but there is some system or component we haven't gotten to that continues to do this.

We are doing another follow-up smoke test tonight. We'll test the cold start valve per Bentley to see if it's leaking gas into the manifold (it's unplugged now - no effect). I bought a used fuel pressure regulator from a member and we'll try that when it arrives.

I know vac leaks are the biggest problem, but if it will run with the oil cap off, would a tiny leak we haven't found somewhere cause it to not start like this?

Thanks for your time and thoughts -

[LateFan]

We did another smoke test last night. No signs of leaking, until we bumped up the pressure a bit and one of the new rocker shaft plugs at the VC hissed a bit, then our intake plug blew out. Tightened the VC a bit, but I think we've found most of the leaks.

Today we tested the cold start valve per Bentley. Engine cold - sat all night. It did not spray ...at all. Bone dry. Testing the connector pins showed about 10.2V from one of them to ground while starting.

Then we went to the thermo-time switch. As far as I can decipher the Bentley instructions, we tested the HARNESS connector pins to ground, not the switch pins (which have no power to them unhooked, right?). One of these showed the same 10.2V. I got no signal from the actual sender pins.

Does all that mean the cold start injector is faulty?

Then we hooked everything back up and started it. Ran well, nice & smooth. Let it idle for some time while it heated up. The idle crept up steadily to about 1200. No idle cycling. I felt the coolant hose at the thermostat go hot when it opened. Temp gauge got to about 1/2.

At about 9 minutes, the engine instantly shut off. Like a switch. No stumble, no misfire, no fuel starvation, just boom - off.

It won't re-start, and is acting the same as the non-start days. Cranks and stumbles. But it's hot now and the cold start shouldn't be functioning anyway....so........??

oh...EDIT..we installed a used functioning fuel pressure regulator as well. On our old one, the vacuum pipe was spinning in the case.

[bddog]

I had all kinds of trouble until I replaced my bad Idle Control Valve. The Oxygen Sensor only has so many miles in it before needing replacement.

[Cephas]

Crank position sensor on the bellhousing. They're cheap so replace both if you can. My eta did the same thing you describe in post two - start, run fine till warm then...dead. It was that sensor.

[NV325i]

^ I agree crankshaft sensor failed on mine with similar symptoms.

[Swanny]

Would they even get the car to stumble or start at all with a failed cps? Granted when mine failed it was from an 87is with just the single sensor on the crank but I had 0 fuel running through the lines because the ecu was refusing to run spark OR fuel.

Easy enough to test both. Could be an intermittent failure....

[Cephas]

Would they even get the car to stumble or start at all with a failed cps?

Yes. Mine was intermittent. Usually died out when the car was warm and worked again after it cooled completely down.

[LateFan]

Interesting... CPS...I wouldn't have thought of that. Good idea.

To answer others questions, the O2 sensor is new, and the ICV has been cleaned and tested as OK. I keep trying to convince myself to replace it, but it seems to be working fine.

Another thought - what all happens over those 9 minutes...other than temperature? Could it be one of the two temp sensors that sends signal to the ECU? Once at full warm, it either sends or doesn't send the correct signal, and the computer screws up?

[Cephas]

I figure if one of those goes bad the most likely result would be a rich or lean running engine. Not a no start.

[LateFan]

Is the slowly creeping high idle related to any of this? Or do we have overlapping problems?

[bddog]

A creeping idle that goes to 1200 is not normal.

[LateFan]

Crank position sensor on the bellhousing. They're cheap so replace both if you can. My eta did the same thing you describe in post two - start, run fine till warm then...dead. It was that sensor.

BUMP

Welp... that wasn't it. Changed both out with new OEM BMW. Still random no-starts. Or, it seems it starts for an instant but won't keep running.

It almost acts like it's out of gas. We're getting some intermittent fuel pump whining - could it be not enough fuel pressure?

Does a fuel pump relay fail all at once, or can it act intermittently?

How do you test the fuel pump vs the in-tank suction pump? Is this in the Bentley?

[JoeyZ]

Any news in this. Im running into the same issue now. It seems as though its limited to days where its cold out.

[LateFan]

Changed the crank position sensors (2). No change, but they are due for replacement anyway. I've replaced the coolant temp senders (all but the dash gauge sender, which is apparently NLA).

I replaced the fuel pump relay and ECU / DME relay - cheap and eliminates those possibilities.

We are going to check the cold start injector again with the new temp-time sender, but that will have nothing to do with warm starting problems.

I think the symptoms are pointing to fuel pumps, while I've been chasing other things. I'll check the two pumps next, but I'm not hearing any hum whatever from the in-tank pump. The under-floor pressure pump is whining intermittently (maybe strained from pulling from the tank). The tank is maybe 2/3 full, and I've learned it can run on the pressure pump alone if it's not low. We may be just on the edge of it not starting all the time.

The creeping idle still needs to be solved. No vacuum leaks anymore. It may be those water senders, but we haven't re-tested with the new ones yet. I hope it doesn't come down to the idle control module in the glove box - another expensive gadget to replace!

[bddog]

Fuel pumps are often intermittent when they fail. Kinda miss mechanical fuel pumps that would leak when they went bad, similar to a water pump.
Sounds like you need at least one fuel pump. If it was for my kid, I would replace them both and a new filter. Would hate for him to get stranded in some strange place.
The rising idle is more than likely your ICV. You can ruin them by cleaning them. The best test is if they are working right.
You can drill a quarter inch hole in a penny. Pull the air meter side of the hose at the ICV. Put the penny there and re connect. It should idle at about 600rpm.
Good until you buy a new one.