Hello, long time reader of these forums, finally had to stop piggybacking and contribute to the conversations. So here's the deal: I purchased a new project with a 2003 5.7 350 mag mpi serial# 0m663930 (I'll confirm serial when I get back to the boat) bravo 2 drive. Runs like a champ for a random amount of time anywhere from 30 seconds to 15 minutes and then dies as if I turned off the key. It does this no matter the rpm idle to 2000 ( max I'll turn it on the muffs).
The guy I got it from said he had it scanned multiple times with no codes showing....I'm currently having trouble getting a hold of a Merc mechanic to scan it for me but I'm working on it. I tend to believe him on this one thing however, because....
I have went through service bulletin 36 to the best of my capabilities along the long road to where I'm at. Along the way I have replaced the oil pressure sensor, the iac, the cps and the alternator. All of which did not fix the problem. I have been running on an external tank and all grounds are 100%. I even cut open the harness to look at splice 104 and a couple others all are good tested with a 5 amp pull while monitoring voltage for drop ecm was disconnected for these tests. I had my friend a long time Chevy mechanic come by and he checked fuel pressure it held at 38 psi at idle...a bit low but like I said this thing runs like a champ when it runs. He also checked the injectors and they are all firing again runs great.
So after a few weeks of dealing with this I decided to monitor the voltage to the sensor inputs on ecm plug b. At this point am 99.5% sure it's the ecm, so unconventional or not I wanted to see if I was loosing a sensor feed signal. Longer story short I got good readings on every wire....but....here's what I'm stuck at. I plugged my probe into plug b18 (I believe) the wake up wire (purple wire) I had 13-14 volts without interruption on this wire while it was running and after it died. With the key still on I touched the probe with my hand( I was sweating it out in the 95-100 degrees out here and I was acting like a ground by leaning on the exhaust manifold) and bam the computer fires up( the fuel pumps cycle the initial 2 seconds then turn off) until I let go of the wire then off again. So I pull out my test light and hook one wire to the battery+ and the neg wire I touched to the ecm casing. Over and over the computer turned on and off when I touched the lights negative wire to the case. So with that I tried grounding the ecm case strait to the battery ground...tried cranking and it turned but no fire...removed the ground and it fires up runs great for a few minutes then off. I didn't run the engine while touching the light to the case because I am somewhat afraid of toasting the ecm completely. Also let me add that I checked the connections on the cannon plug and at the key for the purple wire, all good.
Is there anyone out there that could make some rhyme or reason for this? My best guess is that there is something shorting out inside the ecm to the case probably a ground because the case shows ground on a volt meter but can't power a light. And when I touch the light it overloads the short causing it to actually un-short and "wake up the ecm". My next questions would be: is there anything else that can cause this to happen? Can I get my ecm fixed or flashed or just buy a new one. Can I dismantle the ecm and attack it with a solder gun and try my luck? What would the part number be for a new computer for the engine? I'm having trouble finding the match it seems they have changed part numbers. Also before I go down the road of a new ecm I am planning on taking my current one to a mechanic to see if they can "field test it" for me, then at least I'll know for sure.
Before I release this one to the masters let me say thank you for past and present help!
The guy I got it from said he had it scanned multiple times with no codes showing....I'm currently having trouble getting a hold of a Merc mechanic to scan it for me but I'm working on it. I tend to believe him on this one thing however, because....
I have went through service bulletin 36 to the best of my capabilities along the long road to where I'm at. Along the way I have replaced the oil pressure sensor, the iac, the cps and the alternator. All of which did not fix the problem. I have been running on an external tank and all grounds are 100%. I even cut open the harness to look at splice 104 and a couple others all are good tested with a 5 amp pull while monitoring voltage for drop ecm was disconnected for these tests. I had my friend a long time Chevy mechanic come by and he checked fuel pressure it held at 38 psi at idle...a bit low but like I said this thing runs like a champ when it runs. He also checked the injectors and they are all firing again runs great.
So after a few weeks of dealing with this I decided to monitor the voltage to the sensor inputs on ecm plug b. At this point am 99.5% sure it's the ecm, so unconventional or not I wanted to see if I was loosing a sensor feed signal. Longer story short I got good readings on every wire....but....here's what I'm stuck at. I plugged my probe into plug b18 (I believe) the wake up wire (purple wire) I had 13-14 volts without interruption on this wire while it was running and after it died. With the key still on I touched the probe with my hand( I was sweating it out in the 95-100 degrees out here and I was acting like a ground by leaning on the exhaust manifold) and bam the computer fires up( the fuel pumps cycle the initial 2 seconds then turn off) until I let go of the wire then off again. So I pull out my test light and hook one wire to the battery+ and the neg wire I touched to the ecm casing. Over and over the computer turned on and off when I touched the lights negative wire to the case. So with that I tried grounding the ecm case strait to the battery ground...tried cranking and it turned but no fire...removed the ground and it fires up runs great for a few minutes then off. I didn't run the engine while touching the light to the case because I am somewhat afraid of toasting the ecm completely. Also let me add that I checked the connections on the cannon plug and at the key for the purple wire, all good.
Is there anyone out there that could make some rhyme or reason for this? My best guess is that there is something shorting out inside the ecm to the case probably a ground because the case shows ground on a volt meter but can't power a light. And when I touch the light it overloads the short causing it to actually un-short and "wake up the ecm". My next questions would be: is there anything else that can cause this to happen? Can I get my ecm fixed or flashed or just buy a new one. Can I dismantle the ecm and attack it with a solder gun and try my luck? What would the part number be for a new computer for the engine? I'm having trouble finding the match it seems they have changed part numbers. Also before I go down the road of a new ecm I am planning on taking my current one to a mechanic to see if they can "field test it" for me, then at least I'll know for sure.
Before I release this one to the masters let me say thank you for past and present help!
Last edited:
