Back again. GMT-400 5.7 @ cruise hiccup

[dougg01]

Here we go...

'99 Tahoe 2dr 4wd 5.7 A4 w/ 140k.

 

Had lots of ignition-type problems show up in September.  Nearly all have been resolved.

Replaced:

Spider w/ a Delphi unit

Coil

Cap&rotor twice (now running accel)

Wires

Plugs

Cats are gone (suspected clogged cats)

New distributor

 

The majority of the problem ended up being a bad cap replaced with another bad cap.

 

Now I am left with a hiccup at cruise speed (low throttle, constant or near constant speed).

 

This isn't setting a SES.

 

Logs from Torque Pro seem to show these events, but I'm still clueless as to a cause.

Torque pro exports these as a csv that automagically and cleanly opens in Excel.  Diacom+ logs never went this easy back in the day.

 

The truck seems to shut off for a fraction of a second.  This is not repeatable on demand.  I can drive 10 miles or more without this happening.  The Tach does not seen to respond to this (GM tachs are notoriously slow).

 

I have not found the correct PID for torque pro to show me the TPS voltage or actual throttle position so I can correlate all the readings which kinda sucks, but I can look at the speed and rpm to gather whether I am holding steady.

 

If I am detecting the event correctly in the logs the duration is generally less than 3 frames (about 2 seconds).  With speed and rpm steady the MAF drops significantly (as much as 40g/s), MAP drops (as much as 10psi), and timing goes up a small amount.

That's faster than I can close/reopen the throttle while holding a constant speed.

 

In my testing - this condition only exists with the vehicle in closed-loop.

 

Some have suggested a vacuum leak.  Even cold, I cannot find one.  Certainly its hard to replicate the vacuum signal at cruise rpm with the hood open traveling at least 45mph while spraying carb cleaner.  I also have a real hard time believing a vacuum leak suddenly appears and disappears in 2 seconds at cruise load.  I have not yet taped the vacuum gauge to the windshield to watch for a fluctuation at cruise.

 

Converter engagement is predictable and smooth.

 

If it was consistent I might have more to go on.  Atmospheric conditions do not seem to play a role.

 

I'm likely chasing my tail on this and good gravy it's frustrating.

 

The knock sensor is NOT connected.  When this is resolved I will reconnect.  This is not my first rodeo, but it is my first time having this much trouble diagnosing an obd2 GM.

 

The truck pulls hard and smooth from idle up in any gear regardless of load.  No detectable hesitation.

 

I have no idea how accurate the misfire PID is on torque pro.  It does see a tiny amount of very sparse misfires (several minutes between).

 

It's been suggested the ECM needs to complete a 'WOT relearn' via a tech tool.  I've also been told a drive cycle will not accomplish the same affect.  Yes, I've pulled the ECM fuses for long periods several times and disconnected the battery.

 

I have swapped the MAF and Ign module for units off running/driving vehicles without the issue and the condition has not changed.

 

I don't want to resort to hosting a bonfire with my truck at the center, however, it's not looking like a bad idea right now.

 

Help if you can.

 

Oh....and the engineers that routed the harness and such over the engine need to be dragged into the street, shot, and left there to rot.  :-D

 

[Jsdirt]

Isn't this a GMT-800? I wasn't aware of any 400's having the spider injectors.

 

I think most auto engineers that designed stuff built in the past 22 years, across all brands, would qualify for that punishment ...

 

You're going to need a decent scan tool that can graph the o2's, MAF, MAP, & misfire data - you'll need a snapshot of what happens to all of that when the engine stumbles. Sounds to me like there's a wire that's rubbed through somewhere in the harness, that only hits ground when the engine is under a certain load to place it in the exact position for when you hit the next bump in the road. It could be the power or ground to the ECM, or to the injectors, or to the ignition module. If everything drops out equally, you know to focus on powers and grounds to the engine computer.

 

Diagnosing intermittents are alot of fun, as you've found out. Best thing to do, is do a thorough visual of all wiring harnesses related to engine or transmission management, paying close attention to areas of heat and vibration. That should get you in the correct area.

[dougg01]

Spiders are on all post-TBI 5.7s and 4.3s.  Can't speak to a 5.0, but I'd image it's the same. (EDIT - that does not include marine applications).

 

I have had and have still several 2nd-gen S-series stuff with the spiders and more similarities with the 5.7.  This is my first non-TBI 5.7 in a truck.  I also own a pair of OBD1 LT1 firebirds.  I rarely sell my cars.  Most sit in various states of not being driven.

 

I have logs.  It's not a bump in the road causing it.  It is some set of conditions.  Now that I am also logging throttle position I am not finding the event.  I'm going to add the cylinder misfires to the logging and see what I come up with there.  Last logging session was this past evening on my way home from work.  I was specifically trying to drive at a steady throttle and speed and it did happen a handful of times during the drive, however, it did not happen when/where I expected it to.  Generally it's nearing or over the crest of a shallow hill.  I was ensuring the clutch was engaged and no rpm increase or drop is noted when the throttle and speed are constant (ruling out disengage/re-engage lockup).  I don't really care what I find in the logs as long as I find something conclusive that leads me down a trail.  If a sensor stopped recording that would be something pointing in a direction.

 

Grounds for the ECM or sensors...?  I am not ruling it out, but I have a long, bumpy driveway and 1/3 of my commute is narrow, winding, hilly roads that I tend to drive like I'm in one of my firebirds.  If a ground or power lead was suspect I'd expect this hesitation/hiccup to expose itself then too.  Only at cruise.  Only in closed loop.

 

I'll start piling up some kindling.

 

As for scanning I have a Bluetooth OBD2 adapter and I'm using Torque Pro on my phone.  In my opinion this isn't as good as Diacom+ (what I used to use for tuning my PCMs), but it gives a fair amount of data and gives you a live dashboard to monitor things.  My wife and I ride to work together so I have her on the hook to watch the dials on my phone as it's logging.

 

Right now my data points are every .5sec.  If I don't find some distinct indicator soon I will try it in .25sec intervals.

[Jsdirt]

I am aware of that - what I was getting at is I believe you have a GMT-800 - no GMT-800's ever had TB injection. The 400's did. '94 & '95 were pre-OBD-II crossover years that could have one injection system, or the other.

 

I have no experience with those tools, but it sounds to me like your missing some important data due to the speed of the tool, possibly. Only way to know for sure is to get a buddy, or anyone with a Snap-On Verus, Zeus, or Ethos if they have extremely deep pockets, or an Autel Maxisys or similar, and watch your o2, & MAF data pids, and also watch and see what the fuel trim is doing at cold idle, warm idle, cruise, and full throttle. Even better would be to scope the crank sensor, ignition, MAF, and o2s while driving - that would reveal the problem pretty quickly. Not many DIY'ers are going to have any of this stuff, however. 

[dougg01]

Time for a bonfire.

 

I know a guy with a snap on, but not sure what version.  My father gave his oscilloscope away about a decade ago.  Seriously.  Tells you how I rate.

 

My 99 is the last of the GMT-400s.  It was assembled late in the run too....in Mexico.

 

In open loop it has stumbled and I noticed an unordinary amount of misfires, but that condition did not last and I noted no other sensors acting up.

 

I appreciate your time.

[Jsdirt]

That's good to know - I wasn't aware the 400 ran that long. Thought it was only on 1-tons after '95.

 

Ok, now that gives you a little direction. If it happened in open loop, then you can take the o2s & fuel trims out of the picture. This could be related to the problem you're having at cruise. If this were to show up here with what you've described so far, first thing I'd do is throw a fuel pressure gauge on it to see where it's at. Those spiders are extremely picky about fuel pressure - 1-2 psi can make a dramatic difference, so if a pump is getting weak, or has a bad segment in the armature, that will lower the pressure enough to cause issues. That said, normally you'd see problems at WOT with any kind of fuel delivery issue, so that would steer me toward a fuel QUALITY check first, before doing anything else. It's really starting to sound like some bad gasoline, or there's a little bit of water sloshing around at the bottom that only gets picked up occasionally. Small amounts you wouldn't notice at full throttle, but you'd definitely feel a a stumble at cruise. Not sure if you have ethanol in the fuel where you live,  but old ethanol tends to separate out of old fuel over time, and sit right where the fuel pump can suck it up. If the vehicle sits outside for long periods of time, especially in warm summer weather, or if stale fuel from around the yard gets dumped in it,  this could be your issue.

 

SO, first thing I would do is forget the pressure for now, and remove a fuel line and place it into a clear container ( preferably a large, glass one), cycle the key a few times (or command pump on with scan tool), and see what you've got sitting at the bottom, if you've got a consistent amber color throughout with no wavy-looking streams, or dark, or OJ colored stuff in there. Smell it first - alot of the time that'll tell you all you need to know. This also presents another test - pay attention to how fast your container fills. If cycling key, you're only going to get about a 2 second burst at a time, so keep that in mind, and count how many cycles got you "X" amount of fuel, just to be sure fuel VOLUME is good.