Members | Sign In
ScannerDanner Forum > Post your repair questions here
avatar

2003 Chevy Tracker 2.5l Idle Misfire/Hard starting when hot

posted May 09, 2016 02:55:37 by Jerry Hardway
I have a 2003 Chevy Tracker with the 2.5L V6 and auto transmission, the guys on the suzuki forum sent me here. Just got it from a buddy who has spent about $2000 trying to fix the problem between the dealership and another garage. Just about every sensor on the engine has been replaced as well as fuel pump and fuel regulator. The only codes it has consistently thrown are lean codes for both banks. Yesterday it finally stored a random misfire code in pending codes.

When it goes closed loop it has short term fuel trims at 19.5% for both banks and long term fuel trims around 15% for both banks. When in closed loop it idles smooth as glass, but as soon as it tries to idle in open loop it starts loping and missing, it sounds almost exactly like a motor with a performance camshaft.

Also, after it gets warm and you shut it off for 10 minutes or so it is very hard to start. I have been probing around with my oscilloscope for the past couple of days and figured out that during the no start condition the computer isn't sending trigger signal to the coil packs.

After some more testing on the Crank and Cam sensors I believe my Cam sensor is faulty causing the no start condition. The cam sensor on this car has two channels, one with 6 pulses per cam revolution and varied pulse width to signal ignition timing and the second with 360 pulses per cam revolution with constant pulse width to give a time reference to the computer.

When I crank the motor over with the injectors unhooked cold the 360 and 6 pulse signals are all there and clean, when it is warm the 6 pulse signal is still good but the 360 pulse signal is very choppy. I checked the voltage and ground to the sensor when it was cranking as well and they were both good.

The strange thing is the cam sensor was just replaced and is still producing the seemingly bad signal. I guess the new sensor could be bad, but I don't want to be too willing to just throw money at it. I've attached snapshots of the scope view during both cases. The 6 pulse signal is in blue and the 360 pulse signal is in red for both pictures.

Just wanted to get another opinion on it and see what you thought.

Thanks.

page   1 2 3 next last
33 replies
avatar
Tyler said May 10, 2016 01:03:04
Hey Jerry! Sounds like you've got a fun one on your hands here.

My initial thought is that the lean condition and the hot crank no start are not related, and are likely separate issues. Could be off on that, but I don't see a reason to connected the dots just yet.

First thing I'd suggest is a little fuel trim analysis, mainly revving the engine to 2500 RPM and rechecking the scan data. If you see a significant improvement, you're looking for a vacuum leak. If they stay the same or get worse, then we're thinking about a measurement or fuel delivery issue. This'll give us some general direction to go in.

If you've not seen them already, there's quite a few ScannerDanner videos that deal with lean diagnostics. Couple examples:

https://youtu.be/9gJoH5_Thaw

https://youtu.be/1Ea0Sjx_ZXE

Also, when the vehicle goes into open loop, what do the upstream O2's start doing? I'm thinking they're going lean and staying there.

The dual signals out of the single cam sensor is a new one to me, NEVER seen that before. But I looked up the diagram, and sure enough, there's four wires!

I definitely don't like the behavior of the 360 signal when warmed up. If this sensor is a 5V pull down design, then that pretty much confirms a bad (new) sensor. Any chance of oil contamination inside the sensor itself?
[Last edited May 10, 2016 01:14:54]
avatar
Jerry Hardway said May 10, 2016 03:14:40
I did do a fuel trim analysis while driving it a few days ago. Whether it is at 3500 or 700 rpm the fuel trims stay right around the same spot.

Thanks for the video inks, ill have to check them out.

As for the O2 sensors they read a low voltage once it goes into open loop they go lean, i've never seen them read a high enough voltage to indicate a rich mixture since i've been fooling with it.

As for the dual cam sensor, from what I understand from the manual the camshaft position sensor serves the purpose of both the crank and cam sensors on the common vehicle. The crank sensor is only there for a reference signal for the computer diagnosis. The cam sensor isn't even a hall effect sensor, its a dual channel optical encoder.

I pulled the sensor off tonight and looked it over real good. I couldn't see anything wrong with it so I ordered a re-manufactured one and it will be here sometime this week i hope. However, when I opened the sensor up I could see oily fingerprint marks on the encoder wheel, maybe that had something to do with it. I just dont know.
avatar
Tyler said May 11, 2016 00:00:29
I did do a fuel trim analysis while driving it a few days ago. Whether it is at 3500 or 700 rpm the fuel trims stay right around the same spot.


Ahhhhh, gotcha. This tends to point away from a vacuum leak, then. I'm now more suspicious of a MAF sensor misreporting, or possibly low fuel pressure.

As for the O2 sensors they read a low voltage once it goes into open loop they go lean, i've never seen them read a high enough voltage to indicate a rich mixture since i've been fooling with it.


Right on, this explains the rough idle issue in open loop, as it's running out of fuel.

As for the dual cam sensor, from what I understand from the manual the camshaft position sensor serves the purpose of both the crank and cam sensors on the common vehicle. The crank sensor is only there for a reference signal for the computer diagnosis. The cam sensor isn't even a hall effect sensor, its a dual channel optical encoder.

I pulled the sensor off tonight and looked it over real good. I couldn't see anything wrong with it so I ordered a re-manufactured one and it will be here sometime this week i hope. However, when I opened the sensor up I could see oily fingerprint marks on the encoder wheel, maybe that had something to do with it. I just dont know.


Ohhhhhhhhh, well that explains the lack of coil trigger then!

I wondered if it was some kind of optical system, which would struggle to work if there was oil contamination. The fingerprints wouldn't damage it (unless they touched one of the LEDs), but I wonder why they'd disassemble the sensor in the first place?
avatar
Jerry Hardway said May 11, 2016 21:37:01
Ahhhhh, gotcha. This tends to point away from a vacuum leak, then. I'm now more suspicious of a MAF sensor misreporting, or possibly low fuel pressure.


The new MAF sensor voltage is withing the appropriate range as specified from the manual, but it still could be problem i guess. The fuel pressure is good with around 35 psi while idling, the lower range in the manual is 30 psi.
avatar
RobBrown said May 12, 2016 19:32:58
Are we certain the O2 sensors are functioning properly?
As for the O2 sensors they read a low voltage once it goes into open loop they go lean, i've never seen them read a high enough voltage to indicate a rich mixture since i've been fooling with it.


If you've never seen them indicate rich, I'm concerned that they're fouled/biased low. Could you shoot some propane into the intake and see if they jump up high?
Just a dude who found a sudden and unexpected passion to bless folks who can't afford auto repair with free help. I never, ever imagined this is how'd I'd be spending my free time:-)
avatar
Jerry Hardway said May 15, 2016 03:03:38
I did some more testing today and the O2 sensors will indicate rich under full throttle conditions, but that is the only time they will. All other times it is flat lined lean around .01 volts. The new camshaft position sensor fixed my no start when hot issue, but I still have the lean condition.

I decided to pull the injectors and test them per the service manual. They failed both the flow test and the resistance test, so I ordered a new set of injectors. Fingers crossed this will get me fixed up.
avatar
Tyler said May 16, 2016 00:17:14
The new camshaft position sensor fixed my no start when hot issue, but I still have the lean condition.


Great to hear about the camshaft sensor, sir! Kinda a tough call when a part has already been changed, I know.

I decided to pull the injectors and test them per the service manual. They failed both the flow test and the resistance test, so I ordered a new set of injectors. Fingers crossed this will get me fixed up.


Genuinely curious, how did the service info say how to test for flow? Injector balance test? Observe the spray pattern?

Thanks for keeping us updated on your progress!
avatar
Jerry Hardway said May 16, 2016 05:35:16
Well, the exact method for flow testing was to use their special apparatus to hold the injector onto a special fuel rail designed for the test so you could test one injector at a time. It hooked up to the feed line and fuel pressure regulator from the car and you shorted across the fuel pump relay to apply pressure to the injector. Then you hooked the injector up to a 12v battery for 15 seconds and measured the output.

As for how I did it, I used the fuel rail from the car and just taped the 5 injectors that I wasn't testing at the time to hold them on. I would have somebody hold the one injector in the fuel rail we were testing while I activated the pump and shot the fuel into a cup. Afterward I would measure it with a syringe.

They were supposed to output 64-70 cc of fuel during this test, but the best injector only produced 59 cc.
[Last edited May 16, 2016 05:36:08]
avatar
Jerry Hardway said May 20, 2016 02:40:38
Alright guys, I got the injectors in today and i can say that it helped things tremendously. Going down the road it has much better throttle response and gobs more power. However, it is still severely lean at idle although it hits stoich and the fuel trims come way down around 1500 RPM. I'd say with having the intake manifold apart so many times between myself and the previous mechanics there is a vacuum leak somewhere now. Hopefully I will get time to smoke the manifold tomorrow and see if that is the case.
avatar
Jerry Hardway said May 21, 2016 02:22:04
Quick question....

Does the fuel trims adjust the pulse width of the injector signal?
avatar
Tyler said May 22, 2016 00:28:36
Well, the exact method for flow testing was to use their special apparatus to hold the injector onto a special fuel rail designed for the test so you could test one injector at a time. It hooked up to the feed line and fuel pressure regulator from the car and you shorted across the fuel pump relay to apply pressure to the injector. Then you hooked the injector up to a 12v battery for 15 seconds and measured the output.

As for how I did it, I used the fuel rail from the car and just taped the 5 injectors that I wasn't testing at the time to hold them on. I would have somebody hold the one injector in the fuel rail we were testing while I activated the pump and shot the fuel into a cup. Afterward I would measure it with a syringe.

They were supposed to output 64-70 cc of fuel during this test, but the best injector only produced 59 cc.


Interesting! I like how you adapted the test to work without the special equipment, well done. Having a spec is super handy, but I suppose someone could do this this test without one and look for a flow difference.

So that's another problem down then. I think your fuel trim analysis is right on, and that you're probably looking for a vacuum leak. Any chance you could post some trim numbers now that you've got the injectors changed?

Quick question....

Does the fuel trims adjust the pulse width of the injector signal?


Exactly!
avatar
Jerry Hardway said May 22, 2016 03:22:56
Total fuel trim was hanging around 15% as long as the engine was above 1500 RPM during the test drive. With the engine idling it doesn't like to go into closed loop, but if I blip the throttle it will go into closed loop with a STFT of 0 and a LTFT of around 15%. The STFT will then take about 15 seconds to climb to 20%, stay there for a while and then drop back into open loop.
avatar
Jerry Hardway said May 24, 2016 00:44:05
Okay, expanding on the function of the fuel trims...

From my last post:
With the engine idling it doesn't like to go into closed loop, but if I blip the throttle it will go into closed loop with a STFT of 0 and a LTFT of around 15%. The STFT will then take about 15 seconds to climb to 20%, stay there for a while and then drop back into open loop.

During this whole procedure the injector pulse width doesn't vary. It stays the same whether in open loop or closed loop no matter what the fuel trims are.

My thinking is that the pulse width should be significantly greater with a STFT of 20% (total of 35%) than with a STFT of 0% (total of 15%) with all other things being equal. Or is it not as a pronounced of an effect as I am expecting?
avatar
Tyler said May 25, 2016 01:53:23
My thinking is that the pulse width should be significantly greater with a STFT of 20% (total of 35%) than with a STFT of 0% (total of 15%) with all other things being equal. Or is it not as a pronounced of an effect as I am expecting?


Yeah, you're totally thinking about it correctly. You're reading the pulse width off the scan data?

The amount of change may not be what you expect, but it should be there. I scrounged around, found a decent capture off a '97 Chevy pickup:



On this one, the difference in pulse width between 0% and -5% was only around .2 to .3ms of pulse width. Sorry, it's the best example I could find quickly.

Does the pulse width PID *never* change? If so, chalk that one up as a trash PID. If you can get the O2's to start switching off idle, then the PCM IS applying some kind of pulse width change.
[Last edited May 25, 2016 01:55:26]
avatar
Jerry Hardway said May 25, 2016 13:34:52
I'm measuring it with with a scope probed into injector # 1. The pulsewidth never changes that I can tell. I will take a video of the scope output during this when I get home thursday to show you how it looks.

I've never seen the O2's switch at idle yet.
[Last edited May 25, 2016 13:35:29]
Login below to reply: