kenamond
DSM Wiseman
- 3,225
- 67
- Feb 15, 2006
-
Los Alamos,
New Mexico
My son and I are trying to get my 1997 TSi-AWD back to "perfect running condition" and have hit an issue that has me a bit stumped. I'm pretty sure that this issue is what led me to park the car years ago and buy a new DD.
Cutting to the chase, the car runs perfectly or it doesn't. Schizo DSM. When it's misbehaving, it will start sputtering/missing at all conditions: idle, cruise and boosting. It's not a constant lack of power, it's bursts of working and not working, like it's cutting out for a handful of revs, then firing for a handful, etc. It might start up cold with or without the issue, start up warmed-up with or without and it can switch back and forth from good to bad while driving. It will usually be great or suck for minutes on end or for the whole drive. All day today it was misbehaving. Then we went to the grocery store tonight and it ran fine to and from. My current guess is a non-obvious sensor issue.
The only hint I have so far is that the aftermarket mechanical boost gauge would occasionally read hard zero in the middle of a drive for an extended period (many seconds) and then start reading manifold pressure again. It's not supposed to do that, but that didn't match the good/bad running behavior. We'd see this behavior with the boost gauge both when it was misbehaving and not. I've had that gauge, tied into the system the way it is, for 14 years and when it was my DD, it never did that. My boost gauge is mechanical and its vacuum line is tee'd into the line from FPR solenoid to the FPR, so this meant the solenoid was activating, exposing my boost gauge (and FPR) to atmosphere instead of IM pressure. Although it wasn't throwing an FPR solenoid code, we just replaced with a Mitsu (presumably OEM) solenoid today but the solenoid is still activating (middle of a drive for about 30 seconds but the car was running fine throughout the drive). We can bypass the solenoid, but like I said, the car's stumbling and the FPR solenoid don't happen at the same time, so it isn't just that. I realize that the car runs sub-optimally under load if the solenoid activates (running lean because fuel pressure would be too low without the IM pressure reference), but it doesn't correlate to the stumbling/main problem. There's gotta be something else.
My understanding of the FPR solenoid is that it helps with either warm or cold starts. I've read both versions and don't know which, but that makes me wonder, "what tells the ECU that the engine has just started?" Is it the cam or crank sensor? I recall that the cam or crank sensor signal tells the ECU to turn on the fuel pump, for example. I'm wondering if some other bad sensor is causing the ECU to trigger the FPR solenoid and that other sensor has something to do with the real issue.
So my main question is, what conditions cause the ECU to activate the FPR solenoid? What sensors, either directly or indirectly would lead to this?
Obviously, my other question is, what's wrong? Might not be related.
If you're still reading, some other background on what my son and I have done in the process of trying to figure this out:
We had started out with BLT and fixed a bunch of leaks. The car threw a knock sensor code and we saw that the knock sensor wire broke and it was oozing brown, so we changed that and TB gaskets, injector seals/o-rings while we were at it. No real change other than better BLT results (not perfect but pretty good - we have bigger fish to fry right now than getting it to hold 30psi forever). We did end up making a new high-end boost leak tester that has a built-in gauge, built in pressure regulator and a direct connector to the air compressor
The only other codes we read were boost control solenoid (forgot to plug it back in after doing some work) and the rear O2 sensor code (I have catless exhaust with no rear O2) which it never threw when this was my DD and which stopped throwing since we cleared it a week ago. There are no new codes after driving all week.
We've changed a few parts as well: coils and plug wires (plugs are relatively new NGK BPR7ES gapped to 0.028"), exhaust manifold gasket/studs/nuts and mani-turbo gasket (had an exhaust and stud oil leak), JB-welded a intercooler pinhole leak, better-fitting reducer-couplers at some intake plumbing joints and MAP sensor (the tab for the bolt on one side broke). My son temporarily fixed this by fabricating a metal strap that held the broken sensor body onto the IM - I was a proud papa.
We've been trying to log, but we've had some problems with the OBD2 reader working. It worked this afternoon and then didn't work this evening. If we can get that sorted out, maybe we can see something useful like weird STFTs when its stumbling or weird temps.
Thanks!
Cutting to the chase, the car runs perfectly or it doesn't. Schizo DSM. When it's misbehaving, it will start sputtering/missing at all conditions: idle, cruise and boosting. It's not a constant lack of power, it's bursts of working and not working, like it's cutting out for a handful of revs, then firing for a handful, etc. It might start up cold with or without the issue, start up warmed-up with or without and it can switch back and forth from good to bad while driving. It will usually be great or suck for minutes on end or for the whole drive. All day today it was misbehaving. Then we went to the grocery store tonight and it ran fine to and from. My current guess is a non-obvious sensor issue.
The only hint I have so far is that the aftermarket mechanical boost gauge would occasionally read hard zero in the middle of a drive for an extended period (many seconds) and then start reading manifold pressure again. It's not supposed to do that, but that didn't match the good/bad running behavior. We'd see this behavior with the boost gauge both when it was misbehaving and not. I've had that gauge, tied into the system the way it is, for 14 years and when it was my DD, it never did that. My boost gauge is mechanical and its vacuum line is tee'd into the line from FPR solenoid to the FPR, so this meant the solenoid was activating, exposing my boost gauge (and FPR) to atmosphere instead of IM pressure. Although it wasn't throwing an FPR solenoid code, we just replaced with a Mitsu (presumably OEM) solenoid today but the solenoid is still activating (middle of a drive for about 30 seconds but the car was running fine throughout the drive). We can bypass the solenoid, but like I said, the car's stumbling and the FPR solenoid don't happen at the same time, so it isn't just that. I realize that the car runs sub-optimally under load if the solenoid activates (running lean because fuel pressure would be too low without the IM pressure reference), but it doesn't correlate to the stumbling/main problem. There's gotta be something else.
My understanding of the FPR solenoid is that it helps with either warm or cold starts. I've read both versions and don't know which, but that makes me wonder, "what tells the ECU that the engine has just started?" Is it the cam or crank sensor? I recall that the cam or crank sensor signal tells the ECU to turn on the fuel pump, for example. I'm wondering if some other bad sensor is causing the ECU to trigger the FPR solenoid and that other sensor has something to do with the real issue.
So my main question is, what conditions cause the ECU to activate the FPR solenoid? What sensors, either directly or indirectly would lead to this?
Obviously, my other question is, what's wrong? Might not be related.
If you're still reading, some other background on what my son and I have done in the process of trying to figure this out:
We had started out with BLT and fixed a bunch of leaks. The car threw a knock sensor code and we saw that the knock sensor wire broke and it was oozing brown, so we changed that and TB gaskets, injector seals/o-rings while we were at it. No real change other than better BLT results (not perfect but pretty good - we have bigger fish to fry right now than getting it to hold 30psi forever). We did end up making a new high-end boost leak tester that has a built-in gauge, built in pressure regulator and a direct connector to the air compressor
The only other codes we read were boost control solenoid (forgot to plug it back in after doing some work) and the rear O2 sensor code (I have catless exhaust with no rear O2) which it never threw when this was my DD and which stopped throwing since we cleared it a week ago. There are no new codes after driving all week.
We've changed a few parts as well: coils and plug wires (plugs are relatively new NGK BPR7ES gapped to 0.028"), exhaust manifold gasket/studs/nuts and mani-turbo gasket (had an exhaust and stud oil leak), JB-welded a intercooler pinhole leak, better-fitting reducer-couplers at some intake plumbing joints and MAP sensor (the tab for the bolt on one side broke). My son temporarily fixed this by fabricating a metal strap that held the broken sensor body onto the IM - I was a proud papa.
We've been trying to log, but we've had some problems with the OBD2 reader working. It worked this afternoon and then didn't work this evening. If we can get that sorted out, maybe we can see something useful like weird STFTs when its stumbling or weird temps.
Thanks!
