Cutting off / starting problems - need help.

#1

My 1987 325is (Spec E30 prepped race car) started having a problem back a few months ago - while on track she would suddenly just shut off on me and would not restart. Then after sitting for a while she would start back up and seem to run fine, until later, she would do it again. SOOO, I’ve replaced plug wires, crank position (Impulse pulse) sensor, alternator (which I found was not putting out juice), checked coil (which seems to have slightly high resistance), checked main and fuel relays, checked / cleaned ground on oil pan, can’t remember what else - but I’m getting nowhere.

Any suggestions?

Tom

Winston Salem, NC

#2

Check the owner-added to meet safety regulations cut off switch.
RP

#3

Do they develop a short or something? And wouldn’t it cause there to be no battery power at all? Mine will turn over but just not start.

#4

When this happens and you are cranking over the engine does the fuel pump run and is spark present? If neither condition is true the possibilities are:

  1. No timing data to the DME (CPS problem)
  2. No power to the DME (main relay, fusible link, kill switch)
  3. Bad DME

Since the pump is controlled by the DME and only runs if timing data is presented to the DME, if the pump runs but there’s no spark the problem will likely be:

  1. No power to the coil
  2. Fault in the ground side wiring between coil & DME
  3. Bad coil
  4. Bad DME

If there is spark when cranking over the engine you need to determine whether normal fuel pressure is present at the rail and whether the injectors are firing. Unless the engine harness has been changed, this car won’t have the infamous C191 connector that can take out the injector signals.

#5

We had this happen to our 24Hrs of LeMons car. jlevie has a good guide to debugging it.

In our case, we had spark and fuel pressure but no start til the car cooled off for a while. We gunked up the C191 connector with magic electrical goo and replaced the DME, and it hasn’t come back since. Though only time will tell. We only got the failure during half hour stints of yellow flag laps, which are hard to replicate without taking the car out in a novice HPDE session.

Ah, also check your main relay. And zip tie down your relays while you’re in there.

#6

No fuel and no spark. Switched out both DME and crank position sensor with known good units - no luck. Must be a short in the kill switch - will be looking at that tonight.

#7

What I wrote up for the LeMons car is:

For the engine to run the following conditions must be met:

Power on DME pins:
27 Start Input
18 Unswitched Power input
37 Power Input from Main Relay

Ground on DME pins 2, 14, 19, 24

Timing data from the CPS on DME pins 47 & 48 from a rotating engine

To have spark power must be present at the coil positive and ground pulses
from the DME’s pin 1 must reach the coil negative. Power to the coil is
controlled by the ignition switch via C101. When checking for spark, use the
output lead fronm the coil to eliminate the distributor, rotor and plug wires.

To have injector firing power must be present at each injector and ground
pulses from the DME’s pin 16 (Bank1) and pin 17 (Bank2) must reach the
respective injector bank. Note that the injectors are wired as two banks of
three. With cylinder 1,3,5 being Bank 1 and 2,4,6 being Bank 2. Power to the
injectors is controlled by the main relay.

The fuel pump relay must have power on pin 86 (relay coil) from the main relay
output (pin 87) and power on pin 30. The DME will ground pin 85 to turn on the
relay and power the pump(s) via pin 87. Of the above, only the fuel pump power
is fused. So if the there’s power at pin 87, but not at the pump, check fuse
11.

The main relay and DME pin 18 receive power from the smaller of the to wires
that connect to the battery’s positive terminal. That wire incorporates an
in-line fuse. When the DME is presented with a start signal, it grounds the
main relay pin 85 and furnishes power to the fuel pump relay, injectors, and
DME.

Troubleshooting:

Disconnect the battery and the DME cable. Then:

  1. Disconnect the coil negative and check continuity from that connector to
    DME pin 1. Also verify that from DME pin 1 to ground is an open circuit.

  2. Check the resistance across DME 47 & 47, which should be 500-560
    ohms. If the CPS is dismounted, the resistance can be seen to change
    from about 500 to 540-540 when a ferrous object is brought to the face of the
    sensor. Neither pin should be grounded.

  3. Check for continuity from DME 36 to main relay 85 and from DME 3 to fuel
    pump relay 85.

Reconnect the coil, remount the CPS (air gap should be 0.040"), plug the
relays back in, reconnect the DME, and connect the battery. Then do the
following checks:

  1. With the key off, verify that power is present at DME pin 18 and main relay
    86 & 30.

  2. With the key on, verify that power is present at DME pin 27 and pin

  1. Power to pin 18 is from the main relay and there should be power to the
    injectors and fuel pump relay.

  1. With the key on, verify that no voltage is present at the DME grounds (2,
    14, 19, 24).

  2. Verify that power is present at the coil positive and at fuel pump relay
    pin 30. Those get switched power from the ignition switch via C101.

The engine will start and run (if poorly) with only those connections to the
DME in place. The other signals from Cylinder ID, AFM, temp sensor, etc., are
necessary for proper operation. But they won’t prevent the engine from firing.

IMPORTANT:

A power check means seeing a voltage within about a tenth of a volt of what
you measure across the battery terminals, which should be at least 12.6v on a
charged battery.

A continuity check means seeing less than 1 ohm of resistance.

An open circuit means seeing a resistance of at least 100k ohms.

A good quality auto-ranging Digital MultiMeter will make these tests much
easier.

#8

Excellent checklist! Thanks Jim!

#9

Excellent checklist, thanks for sharing it.

I had a similar situation where corrosion inside the kill switch was preventing full voltage from reaching the main relay. Quick test is to measure voltage on both sides of the switch. If I understand correctly, the DME and main relay expect to see at least 8v (I was getting 3v) which prevented the relay from firing the fuel pump.

#10

First of all, Jim this check list is golden, thanks for that. I have been having the same issue as the OP on my 1987 325is. Will run fine but then all of a sudden it will cut off. Will crank but not start and then after some time will work again fine. I followed all your tests and it passed all of them, so I am starting to change parts. Will start with ECU and have also spare coil, CPS and relays. One question though: I performed resistance tests to the coil. The primary was within the expected range, but the secondary test resulted in 6,000 ohm vs. an expected result of 5,000. Could this be the culprit?

Thanks!

#11

Since the coil apparently has less resistance then it should, might as well replace it with a known good one. That’s probably not the culprit tho. Who knows, maybe it’s crapping out when it gets hot. Stranger things have happened.

I’m more inclined to test, then blindly replace stuff and hope for the best.

I’d be curious what fuel pressure and the Fuel/Air ratio is doing. You can get a fuel pressure gauge at Harbor Freight for beer money. Attach the gauge to a windshield wiper and plumb it into the line between pump and fuel rail. Then replicate the symptom and see if the gauge tells you anything interesting.

A charm of testing the fuel pressure is that the gauge is cheap, and it will come in handy again in the future. If this is a race car, I’d put a fuel pressure gauge on your dash.

The F/A ratio meter is less cheap. But it’s entirely possible that something is going on that is causing your engine to become so lean that it dies and won’t start. If this is a race car, I’d put that on your dash too. That will cost you prob $200 so it’s not beer money. The local exhaust shop will happily put a bung in your exhaust downtube for the O2 sensor.

Harbor Freight will happily sell you a test device that will confirm that your plugs are firing. I always get shocked when I try tho hold a plug with pliers, and I got tired of getting shocked.

HF will also sell you a gizmo that confirms that your injectors are getting pulse signals. I don’t remember what it’s called tho.

Some years ago I was in a death struggle with engine management problems. One thing I did was pull my fuel rail out, cover the engine compartment with a cardboard sheet, then up-end the fuel rail on top. This let me actually see my injectors fire.

My point is that this shouldn’t be complicated, altho it may take some tenacity. The car should run if it’s getting air, fuel, and spark, altho it’s helpful to get them in the right quantities. There are ways to test each one of those. To the greatest extent you can, test, not blindly replace stuff.

#12

Thanks Ranger. You are probably right as to changing stuff vs. testing. Part of the challenge here is replicating the issue as it happens on track and is not predictable. One thing I did not mention before is that many (I do not recall if all) of the times this happened it was after hitting a kerb or a bump. This weekend while doing the tests I removed the DME and noticed that the previous owner had secured it in a way to the car that a stud could hit it with any relevant vibration. He had put a piece of foam on top of the DME casing to protect it from the stud, but it had already made a hole through the foam and had already dinged the casing. That led me to think that maybe the stud would hit the DME and break an internal circuit or something creating the issue. I replaced the DME with a spare one and made a proper bracket and setup to secure the DME in a way that it does not get hit by anything or vibrate. We’ll see if how it behaves this weekend.

#13

Intermittent problems are the worst. Put a fuel pressure gauge and a F/A meter on your dash. That way when you next have problems, you’ll get good clues as to what the problem is or is not.

When I was struggling with engine management problems for a brutal 8months, years ago, it was a total fight to the death. I was tearing my hair out trying to figure out wtf was going on. I went after a lot of unlikely possibilities, like pulling apart every ground in the car, and every connector in the engine bay, abrading the contacts and putting them back together with conductive grease. I eliminated the connector to the injector harness and hard-wired it. I installed the fuel pressure gauge and F/A meter.

I couldn’t get the 1st F/A meter to work, but the second one seemed to. Except that it always read crazy lean so I ignored it.

I shouldn’t have ignored it. That second F/A meter was telling me what the problem was. The whole problem turned out to be an AFM that a couple well-meaning buddies had adjusted by changing the location of the spring dial. My mixture really was way lean. Back then, our understanding of how that spring dial worked was flawed.

The best lessons are painful.

#14

I would vote for the coil. Gets hot, quits working…cools, works. :evil: :evil: :evil:

#15

[quote=“cwbaader” post=82930]I would vote for the coil. Gets hot, quits working…cools, works. :evil: :evil: :evil:[/quote] I’m with Chuck. His idea fits the symptoms nicely.

#16

I chased very similar symptoms for most of the 2015 season - intermittent difficulty starting, eventually followed by cutting off on track. Was never able to troubleshoot effectively until it finally stayed dead in the paddock. When dead, the Jim Levine tests were all good except that voltage at the positive coil terminal was low (about 6 V, as I recall). Eventually discovered a connector that was not fully seated, so it would sometimes, but not always, make adequate contact. In my case, the connector was to a relay that I believe is referred to as the OBD relay. On my car, the relay is under the dash near the left side. Once the connector was properly seated, everything was golden. I subsequently bypassed the relay entirely so I’ll never have to worry about the relay failing.

Cliff