My1989 E30 will not start when its cold. Help?

[Cloudbase]

For the last two years my 1989 250K mile E30 would have trouble starting when it's cold out. It has gradually becoming worse to the point that now it will not start at all when cold.

The specific conditions are when it is cold soaked at temps below 25 degrees F.

When I first crank it, it stumbles for a few seconds like it is going to start but then it just turns over. The only way to get it started is to hook up a tow strap and tow it, drop the clutch and motor the engine for about 3-5 seconds. After that it runs and starts perfect until the next time it gets cold soaked. If I wait for the temp to warm up considerably, it will start again after not starting when cold without towing and motoring the engine.

In the last 50K miles, it's had the head rebuilt by the PO including Bav Auto wires, rotor, cap, coil, BMW fuel filter and motorman injectors. In my time it's also had a new OEM Throttle switch, ICV valve, AFM, and ECU temp sender. I've tested the fuel pump delivery volume in the warm and cold temperatures. It's 875 ml in 30 secs, per the Bentley Manual.

It's still cold enough tomorrow to continue troubleshooting. I'm thinking pulling the plugs and checking spark to start, but am doubtful that is the problem (although I generally don't trust Bav Auto parts).

My mechanic thought fuel pump (which I think is good after testing delivery volume in the cold today), ECU, or ECU temp sender. His theory on the ECU or ECU temp sender is that it's not recognizing how cold the ambient temperature is and not richening the fuel adequately.

I've also read water freezing on the spark plug diodes. This seems very odd, as the symptoms will present after long highway drives the day before. It seems like any water condensed in the oil / engine would have burnt off. But, it would explain why motoring it does eventually allow it to start because the compression would increase the temp inside of the cylinder.

I'm about to pull it all apart to do the timing belt / water pump / etc. It's a good time to replace anything suspect.

Any thoughts would be greatly appreciated. Thanks!

[mpowerful]

Ohm out the ecu temp sensor. I dont remember the exact values but its in the 2000-3000 range cold and 200-300 range hot. Its in the bentley manual.

[jlevie]

Before doing anything else, have a smoke test run. Intake leaks can this problem.

[Cloudbase]

Jim, wouldn't an intake leak present itself at other times than just cold soaked? Unless it was some component that was shrinking due to cold.

[jlevie]

The colder it gets, the greater affect an intake leak has. The DME can compensate to point, which might be enough for warmer conditions.

[Cloudbase]

Rog. I'm going to do the timing belt first, that way any further intake leaks I might cause by disturbing everything could be found by a smoke test.

[bmoores]

Any links to how to perform the smoke test?

TIA

[Earendil]

It has gradually becoming worse...

The specific conditions are when it is cold...

...stumbles for a few seconds like it is going to start but then it just turns over. The only way to get it started is to hook up a tow strap and tow it, drop the clutch and motor the engine for about 3-5 seconds. After that it runs and starts perfect until the next time it gets cold soaked.

This sounds like an electrical problem to me, specifically a bad battery. While not a common way for a battery to die, they can fail in a state where they can store enough charge but can't supply the voltage. This method of battery failure will result in an engine that can crank and crank but doesn't produce the voltage required to run the ECU and light up the spark plugs.

How do I know? I had an E30 in Spokane (@#$%ing cold) Washington that always cranked over a few times and would "almost catch" before finally firing up. As the temperatures dropped, the amount of cranking took longer before it would finally "catch" though it often times would "sorta"' catch. Finally it got into a state exactly like yours. Note that the starter never sounded like a dead battery, but when I plugged a voltmeter up to the battery and tried to start the car, I saw the system voltage drop to around 9.8v :down:

It's easy to prove me wrong, if indeed I am. Put a multimeter on the battery terminals and make sure you get good voltage (12.8 or so) when the car is off. Then, while watching the voltage, try and start the car. You should see some drop, but with a good battery and electrical system it shouldn't drop below 12v, and if it drops below 11v I would expect the car to have trouble starting.

Alternatively, try and "jump the car", and see if it starts up any faster. It could of course be that you have something else wrong with the electrical system that is sucking juice, but that becomes less common than battery failure. Also note that when a battery is being charged/discharged that it produces heat. Lots of times failing batteries will heat and cool just enough that they produce different symptoms at seemingly different engine temps.

[Cloudbase]

Update:

I pulled the wiring harness off the blue temp sender. I'm getting 3300oHms off the spade and valve cover as a ground.

The battery was replaced last year. I just checked it, +12V at the terminals.

I checked the ECU loom end of the ECU temp sender, plugs 24 and 45 for Motronic 1.3 per this write up.

http://www.e30zone.net/e30zonewiki/index.php/Blue_Plug

No resistance at the ECU loom, nothing. I even made little spades to make sure I was getting a solid contact. I'm thinking it could be the dreaded C191 connector. Perhaps water is freezing in there and separating the connection.

I need to get it running so I can move it in my garage though.

Further Update:

Well I used a hair dryer and an aviation engine insulating blanket to heat under the manifold. I never could get an Ohm reading at the ECU loom, but it did start right up once I pre-heated the engine. Wonder if my testing method is bad.

[E30NJ]

just a guess, what about the cold start valve.

[Cloudbase]

Post facelift cars don't have a cold start valve IIRC

It was too warm today to duplicate the issue so I spent all day working on a friend's IX.

I should have a chance to dig into it next week when I do my timing belt. I'll report my findings.