HELP I've posted this in the 129 roadster section, but that site doesn't get much traffic.My son in law was using my wife's 97 SL320 while we were on vacation. He was driving along and the car died. He pulled to the roadside to restart. The car starts, runs a few second, and dies. This is what the car does now all the time. The car starts , runs a few seconds, and dies. I'm looking for someone who has had the same problem and solved the problem. By the way a 97 E320 has the same system.

If you are a professional MB mechanic, I look forward to your input. I've discussed this problem with some of the best in the profession to no avail. Here's a list of facts.

No codes (checked with snap-on, Baum, SDS)
Fuel pressure in specs
Delivers more than liter in 40 seconds
Grounds and power to Engine Control Module checked
No problems indicated with DAS
Tried all 3 keys including valet
CPS replaced with new unit just for s*its and giggles
Same symptoms when fuel pump made to run with Snap-On scanner
Yr model too early for rusty flywheel
Car runs on carb cleaner
While running on carb cleaner no MIL and attached noid light flashes but appears dim

UPDATE: latest checks
Unplugged MAF no change
Shorted engine coolant temp
Smoked the intake

This one is kicking my butt.
Thanks in advance for any new light, because I'm running out of straws.

 

If it runs on carb cleaner then it should just be a fuel issue.
Have you had the fuel pressure gauge on while the car was running?

If fuel pressure ok, I would also check fuel injector spray.

 

If your mechanic has an oscilloscope, check for the injector pulse at the injectors and see if it disappears when the engine shuts off. If it does, follow that circuit back to the ecu and see if it disappears there, if it does, the computer probably thinks the key is being turned off, that's why your not getting any codes. If that's the case, I would check the ignition switch or car alarm circuits. Monitor all of the power inputs to the ecu to see which one is cutting out when the car dies.

 

Thanks for the feedback. We have a basic handheld mechanic's scope. I've been looking for a good digital Tektronix lab scope, because it is going to be necessary on the newer vehicles with SAMs and CAN buses. We've been thinking in the direction of injector pulse width but haven't checked it yet. I have checked power and grounds at the ECU according to an Alldata printout, but don't remember if I did it live. The alarms are not an issue, because the infrared door locks have nothing to do with with DAS X which is the driver auth on this car. It is the latest upgrade from DAS 1 and DAS2. When I run the scan test with SDS, it passes DAS everytime. Others have suggested ignition switch. I'm trying to figure how to prove that. Maybe I'll get a chance to work on it today. It has just been 90 and 100 degrees for the last several weeks here in the bay area. I'm semi retired; Iwork when I feel like it now.

 

If you remember the car's way back in the 60's and 70's that had points ignitions. If the ballast resister went bad the car would do the same thing. It would start in the "start" position but would die as soon as the key was in the "run" position. I don't have a schematic for your switch. I would check into the "start" and "run" circuits from the switch

 

Bump. Has no one experienced this problem? I recheked the wire from the key to the engine control unit. It is a violet .5 mil wire and goes to terminal 40 on plug C. With key on I thought I found someyhing with zero volts, but got 12 volts in start position. Rechecked the prints and discovered that this is a start bus and not a run bus. Will check pulse width at injectors tomorrow with engine running on carb cleaner. My partner who could be an MM if he ever applied wants to blow false air across the MAS to see if pulse width changes. The theory is that pulse width is decreasing and starving the engine for fuel, but what is causing it?

 

You should use a spray bottle with gas, not carb cleaner, if you are going to run it for any amount of time. That's not good for the o2 sensors. I just noticed you said the noid light appears dim. If that's the case, you need to check the voltage at the injectors, preferably with a scope. I'm not sure how that system works, I don't have a schematic for it. Whether there is a constant voltage and the computer grounds the injector to activate it or the injectors are grounded and the computer supplies the voltage to activate. One of these has a problem. If the computer supplies the voltage to activate the injector, check the voltage level at the injector, then at the computer, make sure there isn't a wiring problem, if it's low coming out of the ecu, there's your problem. If the computer grounds the injectors, check the constant voltage at the injectors.
Hope this helps.

 

As they say on IATN" closed with fix." I'm posting in hopes of helping anyone else who experiences this problem. It turned out to be a failed Mass Air Flow Sensor, but hold on before you say you figured that. It wasn't that simple, because you could have installed a new sensor in the beginning of throwing parts at a car and the car would still start and die. I took the ECM over to another shop where a friend was having a barely running problem with a 97 S320. We scanned his car with my SDS and got no MAF codes on a running vehicle. Vacuum was low, so we hooked up a back pressure gauge to check the cats. We installed my ECM in hopes the car would recognize it and sort of run, but it didn't recognize it. I left my ECM there for a week, because I forgot it. When I got time to work on my car again, I retrieved it and installed it. My partner and I were preparing to run the car on propane to see if we could get codes while monitoring with SDS. Before we started, I wanted to make sure we had the same conditions, so I tried to start the car, and when I did it started and ran. It ran for about 10 minutes and died. We played with the MAF which made a difference plugged and unpluged. We installed a failed MAF from a C280 that had thrown lean trim codes. That seemed to solve the problem although the car didn't run right, because of the lean trim problem. I got a new MAF the next day. I gave my MAF to my friend working on the S320. He called me right away to say his car now starts and dies. I gave him the lean trim code MAF, and his car then ran.

My observation in all this is that the ECMs must store some deep memory that died during the week the unit was out of the car. Even though the MAF was defective, it wasn't shutting down the car, because unplugging it made no difference. It was the computer protecting the car like HAL. Remember HAL? I can now unplug the MAF and the car still runs, which is normal. I don't like to demonstrate that too often, because I gotta go back in and clear MAF sensor and throttle position implausible codes. A suggestion to you fellow MB mechanics who read this; this is one of those few times it is ok to throw a part at a car. I hope this benifits someone.

 

Good job staying with it and thanx for posting !