95 BMW M3: Warm Engine = Intermittent Start


I have a real trooper of a '95 M3 with 276k/mi. on the original block (reman. head, 3rd used ZF trans. just went in two weeks ago, and updated suspension, but otherwise stock).

The Problem:
Basically, the car doesn’t want to crank sometimes when it’s warmed up. A few years ago a problem started (or didn’t start) where the car would intermittently fail to turn over if the engine was brought up to temp and I parked somewhere for 15-30min. then came back to start up and drive off. Being a manual, I took to parking on inclines so I could bump start it when this happened. Living in Kansas, those inclines are not always easy to find.

Data Points:
• When attempting to crank, the display dims a little as if current is being drawn
• There is sometimes a click like the solenoid is engaging, but after a few attempts it will stop clicking, or it won’t click at all from the outset
• Even without the clicking, the display still dims as with current draw.
• The problem is inconsistent, sometimes there, other times fortunately not, and happens summer or winter
• It has occasionally happened when the engine was cold (first start of the day), but almost always starts cold (99%) or after the car has sat for an hour or more

What I’ve Done:
• Replaced starter (twice) with new units (crank was never weak, it was always all or nothing)
• New battery (with no difference in cranking ‘authority’), old one worked as well as the new
• Replaced engine ground strap and added another ground connection directly to starter body (starter bolts to Al trans. housing so ground jumps back to block then ground strap)
• Cleaned all positive lead connections in the engine compartment from the jump terminal (battery is in the trunk) to the fuse block and starter/solenoid
• Replaced battery ground strap in trunk
• Soldered a bridge to bypass the EWS II anti-theft module*

So, my mind won’t leave the premise that there is a weak contact somewhere that is irritated by elevated temperatures and weakens, then contracts when it cools and restores enough connectivity to crank the starter, leading to the symptoms I’m seeing. I have NOT examined any connections past the battery positive terminal to the jump terminal in the engine compartment, but I understand there is a fuseable link inline.

I’m hopeful that the symptoms (dimming display) resonate with someone who has had similar problems and can point me in the right direction of how to diagnose/troubleshoot this or what to look for.

Thanks in advance for all shared wisdom or witty non-sequiters,

(BigCat because I help care for a few. Check out the links below if interested.)

*EWS II is an anti-theft system that uses an antenna ring around the ignition cylinder to talk to the RFID chip in the key and will interrupt either starter or spark. I bought the car in 2001 with 96k/mi. on the odo. The PO apparently had issues with the unit as it has a sticker post '95, and the ‘fix’ of jumping the starter wires was already done when I examined it, but with the aluminium press-fit connectors you can buy at a hardware store, so I removed it and soldered a proper jumper in its place.


I would worry about the main battery cables. At the very least I would peel some insulation off of the existing cables to look for hidden corrosion down underneath. It does just sound like a classic power connection issue. You mention having cleaned positive terminals and adding some extra grounds. What have you done with the existing ground connections in terms of the main power cables?


Thanks, Cig-
I’ve re-worked all the main grounds in the engine bay and at the battery-to-body in the trunk.

The positive lead is the last one I haven’t explored as it runs passenger side under the interior trim (also where the fuseable link is located).

The fact that it’s almost always when the engine is warm makes me think it’s something in the engine bay getting heat soaked, but you’ve given me the idea to check resistance on the power cables while cold as well as in the hot, no-start situation to see if I can isolate which lead may be the culprit.

Thanks again for the feedback and if I had a brother, I’d suggest no one drive like him!


Better than checking resistance, check the voltage drop: http://www.aa1car.com/library/voltage_drop_testing.htm


Here is your schematic. Battery goes to power distribution before the starter.


Great resource, Cig. I will try that. Thanks again!

And thank you, knfenimore. I’ve cleaned the power distribution block (same pickup as the engine bay jumping lugs) on all connections, but I’ll double check the power in to the block as I may have overlooked that one last time I cleaned them since it’s below the sheet metal. I’ve got the Bentley manual (well worn and bookmarked) and gone over the circuit diagrams to verify I’m not overlooking something.

Many thanks again!


Ask your mechanic to measure the voltage at both starter terminals, from the terminal to the starter case. If either is less than 10.5 volts during attempted cranking, then work backwards from there to find out where the voltage drop is occurring. Toyota starters for my Corolla at least will crank at 9.6 volts, but you need some safety margin. So shoot for 10.5 volts. If both are above 10.5 volts and it doesn’t crank then the problem is the starter motor or the engine is locking up or the starter motor isn’t making contact with the flywheel. Starter motors can be intermittently bad and still crank fast when they actually crank. This can be caused by bad (high resistance) spots on the solenoid contacts, or bad spots on the commutator.


Will do, George. I’m the ‘mechanic’, so I’ll be doing the voltage drop test as outlined by Cigroller following your order of operations.