Fuel pressure sensor replacement guide with pics
07-31-2023, 01:59 PM
#61
Race Car
Both my 6TT's idle perfectly stable like the tacho needle is glued to the gauge. Hunting idle is not normal.
Things to check,
1. Throttle body sensor can get dirty or defective and the ecu can't see the throttle plate position which means the throttle plate jumps around. Can check this with durametric or some other scanner, just watch throttle position (not gas pedal position) while it's idling.
2. Vacuum leaks, although the shop you brought it to allegedly didn't find any they also lied to you about idle characteristics.
3. Intake/boost leaks. These are a part of life with a turbo car, a boost leak check should be mandatory when troubleshooting any rough running condition.
4. These cars eat MAF's for breakfast. Usually a bad MAF will throw the trifecta of trouble codes: check engine, PSM failure, ABS failure.
5. Plugs/coil packs.
I think you need to do some datalogging to figure out what is going on.
Things to check,
1. Throttle body sensor can get dirty or defective and the ecu can't see the throttle plate position which means the throttle plate jumps around. Can check this with durametric or some other scanner, just watch throttle position (not gas pedal position) while it's idling.
2. Vacuum leaks, although the shop you brought it to allegedly didn't find any they also lied to you about idle characteristics.
3. Intake/boost leaks. These are a part of life with a turbo car, a boost leak check should be mandatory when troubleshooting any rough running condition.
4. These cars eat MAF's for breakfast. Usually a bad MAF will throw the trifecta of trouble codes: check engine, PSM failure, ABS failure.
5. Plugs/coil packs.
I think you need to do some datalogging to figure out what is going on.
The following users liked this post:
saxonthebeach (07-31-2023)
07-31-2023, 02:23 PM
#62
Both my 6TT's idle perfectly stable like the tacho needle is glued to the gauge. Hunting idle is not normal.
Things to check,
1. Throttle body sensor can get dirty or defective and the ecu can't see the throttle plate position which means the throttle plate jumps around. Can check this with durametric or some other scanner, just watch throttle position (not gas pedal position) while it's idling.
2. Vacuum leaks, although the shop you brought it to allegedly didn't find any they also lied to you about idle characteristics.
3. Intake/boost leaks. These are a part of life with a turbo car, a boost leak check should be mandatory when troubleshooting any rough running condition.
4. These cars eat MAF's for breakfast. Usually a bad MAF will throw the trifecta of trouble codes: check engine, PSM failure, ABS failure.
5. Plugs/coil packs.
I think you need to do some datalogging to figure out what is going on.
Things to check,
1. Throttle body sensor can get dirty or defective and the ecu can't see the throttle plate position which means the throttle plate jumps around. Can check this with durametric or some other scanner, just watch throttle position (not gas pedal position) while it's idling.
2. Vacuum leaks, although the shop you brought it to allegedly didn't find any they also lied to you about idle characteristics.
3. Intake/boost leaks. These are a part of life with a turbo car, a boost leak check should be mandatory when troubleshooting any rough running condition.
4. These cars eat MAF's for breakfast. Usually a bad MAF will throw the trifecta of trouble codes: check engine, PSM failure, ABS failure.
5. Plugs/coil packs.
I think you need to do some datalogging to figure out what is going on.
1. Will check throttle body sensor with Durametric later today- will note I cleaned the throttle body itself ~2 yrs / <10k miles ago.
Edit: I just checked - there are multiple values on durametric related to throttle position (Spec. throttle, actual throttle, throttle potentiometer 1, throttle potentiometer 2, etc). I assume Spec. throttle = ecu specified throttle position? If so, this value moves around a bit (constantly changing anywhere from 2.3 to 2.8%). "Actual throttle" is much more steady, but does move even at idle (jumps from 2.3 to 2.7 and sometimes 2.5). All the other values (potentiometers) move constantly, similar to the spec throttle value. My idle RPM moves a bunch as discussed, mostly between 800-900 RPM, and my spec RPM at idle (AC off) is 840 RPM. Is "spec RPM" specified RPM? If so 840 is high, no? My understanding is idle RPM should be 740 +/- 40 RPM.
2. I misspoke slightly on this one earlier- The shop that checked for vacuum leaks stated they did actually find one small one (on diverter valve) and repaired it. Fairly confident there are no longer any leaks, but no way to be 100% certain here as I don't have personal access to a smoke machine.
3. Shop claims no boost leaks either, although I imagine this is something I can theoretically check myself.
4. Car is not throwing any codes at all, so inclined to think MAF is not the issue
5. Plugs/coil packs were also done 2 yrs / <10k miles ago so likewise disinclined to suspect that here
I keep coming back to the fuel system just given the change in idle behavior when I replaced that check valve with a good one. The couple starts immediately after replacing the valve were the only time in recent memory I've seen the idle behave as it should (glued to tach as you say). Is there anything else that could be at issue in the fuel system? I'm wondering about the pump itself, but I also hesitate to throw a $600 part at the car on a whim.
Re: datalogging- are there any specific parameters in Durametric I should log over a drive / at idle to help diagnose?
Last edited by saxonthebeach; 07-31-2023 at 02:44 PM. Reason: Checked with durametric
07-31-2023, 03:17 PM
#63
Race Car
You are correct about Spec RPM etc, that's what the ECU wants. The throttle pot 1 and throttle pot 2 are the readings off the throttle body potentiometers, if they are changing in a similar way to each other then they are most likely OK. What can happen is one pot develops a dead spot around idle and this causes the idle to hunt as the pot crosses that dead spot.
I can't believe the fuel pump would cause idle hunting unless it was on the verge of death. Even with a huge split in the fuel hose in the tank my car idled fine, it just would not take more than a tiny bit of throttle before it started bucking badly. When the fuel line completely let go it would not run at all.
When was the last time the fuel filter was changed?
Data logging probably won't find a vacuum leak, it can find the presence of a large boost leak although not the exact location. However 90% of the vacuum system is pressurized with boost when under boost so a boost leak check should find 90% of vacuum leaks. Using a smoke machine to check for vacuum leaks on this car is a waste of time because of this.
Basically to boost leak check disconnect both turbo to intercooler hoses at the turbo, cap one off, use the other to pressurize the intake system. The shut off the pressurized air. It should hold 16-20 PSI for several seconds (mine holds for 30+ seconds before slowly decreasing).
The factory manual says to boost leak check by pressuring the system from where the MAF is. This is worthless because of the oil tank vents etc and because this section never sees pressure.
Dura has trouble logging more than 7 or 8 at a time. I stopped using it and now use my Cobb Accessport. I have not datalogged to find idle problems so don't know where to start there.
In realtime just sitting there idling I would look at each cylinder's misfire count (should be 0) and ignition timing. I can't remember what Dura calls this ignition timing, it's what the ecu uses to control knock.
Check each bank's lambda may tell you if the fuel pump is working properly at idle.
Using your fuel pressure test gauge while idling can show if the idle fluctuations are related to fuel pressure.
Do note that when I had a dead injector on #5 the ECU recorded misfires on every cylinder except the one that had the dead injector so I am not sure how valuable the misfire counter is. I found the dead injector by touching the exhaust pipes looking for the cold one.
When I was looking for boost leaks I logged:
engine RPM
injector duty cycle
setpoint manifold pressure
actual manifold pressure
N75 valve duty cycle
Lambda bank 1 (need to convert to AFR)
Lambda bank 2
then ran some 3rd gear pulls. This would also test fuel delivery as Lambda ought to be 10-ish AFR at full boost and full load. If not suspect fuel delivery somewhere.
If I have time tonight I'll load up dura and see what my car says when idling.
I can't believe the fuel pump would cause idle hunting unless it was on the verge of death. Even with a huge split in the fuel hose in the tank my car idled fine, it just would not take more than a tiny bit of throttle before it started bucking badly. When the fuel line completely let go it would not run at all.
When was the last time the fuel filter was changed?
Data logging probably won't find a vacuum leak, it can find the presence of a large boost leak although not the exact location. However 90% of the vacuum system is pressurized with boost when under boost so a boost leak check should find 90% of vacuum leaks. Using a smoke machine to check for vacuum leaks on this car is a waste of time because of this.
Basically to boost leak check disconnect both turbo to intercooler hoses at the turbo, cap one off, use the other to pressurize the intake system. The shut off the pressurized air. It should hold 16-20 PSI for several seconds (mine holds for 30+ seconds before slowly decreasing).
The factory manual says to boost leak check by pressuring the system from where the MAF is. This is worthless because of the oil tank vents etc and because this section never sees pressure.
Dura has trouble logging more than 7 or 8 at a time. I stopped using it and now use my Cobb Accessport. I have not datalogged to find idle problems so don't know where to start there.
In realtime just sitting there idling I would look at each cylinder's misfire count (should be 0) and ignition timing. I can't remember what Dura calls this ignition timing, it's what the ecu uses to control knock.
Check each bank's lambda may tell you if the fuel pump is working properly at idle.
Using your fuel pressure test gauge while idling can show if the idle fluctuations are related to fuel pressure.
Do note that when I had a dead injector on #5 the ECU recorded misfires on every cylinder except the one that had the dead injector so I am not sure how valuable the misfire counter is. I found the dead injector by touching the exhaust pipes looking for the cold one.
When I was looking for boost leaks I logged:
engine RPM
injector duty cycle
setpoint manifold pressure
actual manifold pressure
N75 valve duty cycle
Lambda bank 1 (need to convert to AFR)
Lambda bank 2
then ran some 3rd gear pulls. This would also test fuel delivery as Lambda ought to be 10-ish AFR at full boost and full load. If not suspect fuel delivery somewhere.
If I have time tonight I'll load up dura and see what my car says when idling.
07-31-2023, 03:45 PM
#64
You are correct about Spec RPM etc, that's what the ECU wants. The throttle pot 1 and throttle pot 2 are the readings off the throttle body potentiometers, if they are changing in a similar way to each other then they are most likely OK. What can happen is one pot develops a dead spot around idle and this causes the idle to hunt as the pot crosses that dead spot.
I can't believe the fuel pump would cause idle hunting unless it was on the verge of death. Even with a huge split in the fuel hose in the tank my car idled fine, it just would not take more than a tiny bit of throttle before it started bucking badly. When the fuel line completely let go it would not run at all.
When was the last time the fuel filter was changed?
Data logging probably won't find a vacuum leak, it can find the presence of a large boost leak although not the exact location. However 90% of the vacuum system is pressurized with boost when under boost so a boost leak check should find 90% of vacuum leaks. Using a smoke machine to check for vacuum leaks on this car is a waste of time because of this.
Basically to boost leak check disconnect both turbo to intercooler hoses at the turbo, cap one off, use the other to pressurize the intake system. The shut off the pressurized air. It should hold 16-20 PSI for several seconds (mine holds for 30+ seconds before slowly decreasing).
The factory manual says to boost leak check by pressuring the system from where the MAF is. This is worthless because of the oil tank vents etc and because this section never sees pressure.
Dura has trouble logging more than 7 or 8 at a time. I stopped using it and now use my Cobb Accessport. I have not datalogged to find idle problems so don't know where to start there.
In realtime just sitting there idling I would look at each cylinder's misfire count (should be 0) and ignition timing. I can't remember what Dura calls this ignition timing, it's what the ecu uses to control knock.
Check each bank's lambda may tell you if the fuel pump is working properly at idle.
Using your fuel pressure test gauge while idling can show if the idle fluctuations are related to fuel pressure.
Do note that when I had a dead injector on #5 the ECU recorded misfires on every cylinder except the one that had the dead injector so I am not sure how valuable the misfire counter is. I found the dead injector by touching the exhaust pipes looking for the cold one.
When I was looking for boost leaks I logged:
engine RPM
injector duty cycle
setpoint manifold pressure
actual manifold pressure
N75 valve duty cycle
Lambda bank 1 (need to convert to AFR)
Lambda bank 2
then ran some 3rd gear pulls. This would also test fuel delivery as Lambda ought to be 10-ish AFR at full boost and full load. If not suspect fuel delivery somewhere.
If I have time tonight I'll load up dura and see what my car says when idling.
I can't believe the fuel pump would cause idle hunting unless it was on the verge of death. Even with a huge split in the fuel hose in the tank my car idled fine, it just would not take more than a tiny bit of throttle before it started bucking badly. When the fuel line completely let go it would not run at all.
When was the last time the fuel filter was changed?
Data logging probably won't find a vacuum leak, it can find the presence of a large boost leak although not the exact location. However 90% of the vacuum system is pressurized with boost when under boost so a boost leak check should find 90% of vacuum leaks. Using a smoke machine to check for vacuum leaks on this car is a waste of time because of this.
Basically to boost leak check disconnect both turbo to intercooler hoses at the turbo, cap one off, use the other to pressurize the intake system. The shut off the pressurized air. It should hold 16-20 PSI for several seconds (mine holds for 30+ seconds before slowly decreasing).
The factory manual says to boost leak check by pressuring the system from where the MAF is. This is worthless because of the oil tank vents etc and because this section never sees pressure.
Dura has trouble logging more than 7 or 8 at a time. I stopped using it and now use my Cobb Accessport. I have not datalogged to find idle problems so don't know where to start there.
In realtime just sitting there idling I would look at each cylinder's misfire count (should be 0) and ignition timing. I can't remember what Dura calls this ignition timing, it's what the ecu uses to control knock.
Check each bank's lambda may tell you if the fuel pump is working properly at idle.
Using your fuel pressure test gauge while idling can show if the idle fluctuations are related to fuel pressure.
Do note that when I had a dead injector on #5 the ECU recorded misfires on every cylinder except the one that had the dead injector so I am not sure how valuable the misfire counter is. I found the dead injector by touching the exhaust pipes looking for the cold one.
When I was looking for boost leaks I logged:
engine RPM
injector duty cycle
setpoint manifold pressure
actual manifold pressure
N75 valve duty cycle
Lambda bank 1 (need to convert to AFR)
Lambda bank 2
then ran some 3rd gear pulls. This would also test fuel delivery as Lambda ought to be 10-ish AFR at full boost and full load. If not suspect fuel delivery somewhere.
If I have time tonight I'll load up dura and see what my car says when idling.
One thing I just realized on the fuel point- the other big variable that hasn’t been mentioned is that the battery is disconnected for a solid 30-45min every time I go in the fuel tank, which disconnect I think causes a soft ECU reset? So maybe not fuel system related as you say; could just be coincidental that it ran well for a bit after a reset.
08-02-2023, 05:26 PM
#65
Thanks for all the help- I am still working to sort out this idle issue, but since this has gone way far afield from this threads original topic, I made a new thread here: https://rennlist.com/forums/996-turb...l#post18940687