Here's a video of exactly what's happening....
starts by barely hitting key (ONLY WHEN COLD START).....
Runs as smooth as you could ask with idle at 1500 rpm...
AS SOON AS it hits 1k rpm it's like I switched key off.....
I can get it to restart with fluid but I have to keep throttle to it and it runs rough after that initial start....
UNLESS I run it long enough for motor to warm up good then from that point on it runs as good as you could ask and better than it ever has since I've owned it.....
WHAT IN THE WORLD is causing this? It is about to drive me INSANE!!!
xc_hide_links_from_guests_guests_error_hide_media
Rustjunky67,
Your problem has a specific/unique footprint, and is intriguing. Instead of trying to weave the following together
into some grand unified theory, instead I'm just going to list a few observations, with the hope that the
subject matter experts can chime in.
* FWIW, in a previous life I used to teach troubleshooting, both analog, digital, and closed loop electromechanical.
A point I used to stress over & over was that you couldn't fix an intermittent until you could first figure out how
to
break it at will.
Congratulations! - as you demo'd in your video, you have just figured out the hardest part of the problem.
* I think that an important point to keep in mind is that once the engine is fully warmed up, you are happy with the
way it runs. (!) This is good for a couple of reasons. First of all, this makes the problem a bounded one, in that
the system starts fine. Then it fails. But once it's fully warmed up, it again runs fine.
This is much easier to troubleshoot than an engine that works fine cold, but always runs rough when warmed up.
Could be anything: mechanical (compression), or a thermal fault in the ignition (spark), or a weak fuel pump. (fuel)
So let's keep this important clue in mind. How can an engine start fine, stall suddenly right at 1,000 rpm -- but if you
continue to nurse it/warm it up to normal operating temps it will run fine for the rest of the day?
* You don't have the GMT400 community scan tool yet. But this is in the works? Well, instead of just relying solely
on the scan tool to show you the way, I propose that we troubleshoot this old school first, come up with a theory based
upon our observations, and then when the scan tool shows up, we can use it to confirm/deny our theory?
For example, if I was your neighbor, since you have figured out how to break this at will, I would want to try to 'see'
what's going on as much as possible. We could get not one but 2 important clues by going to the cold engine, and
connecting a good vacuum gauge and a timing light to it first. Then start the engine, and let that timing light flash
away while watching that vacuum gauge closely.
By definition, in general terms the more vacuum that the engine can pull against the closed throttle plate the healthier it is.
(stock cam & tune.) If you have 2 identical engines running side by side, and the first one is giving you 20 inches
of steady vacuum at 1200 rpm, while the second engine is giving me a shaky 12 inches of vacuum at the same
1200 rpm, then I already know that one engine is right with the world, and the other one ain't.
So, with nothing more than a vacuum gauge, I want to see what your engine gives me right after startup? Is it
dropping off slowly as the rpm drops, until it's much lower than it should be and the engine finally shudders to a halt?
Or does the vacuum stay nice & high, all the way down to 1000 rpm, and then it is lost suddenly, as if you turned the
key to the off position? (Think sick engine as perpetrator of stopping? Or is it a healthy engine that is the victim of a
supporting function suddenly disappearing?)
And this is where the timing light comes in. I'd be watching for a steady flash flash flash right from start up. And as the
engine gets closer to the magic 1000 rpm, I want to see if A) the flashes continue, but slow down in time with the
engine rpm? Or B), do the timing light flashes suddenly quit, and
then the engine comes to a stop?
So, with just a vacuum gauge & a timing light, I think we can collect some old school solid troubleshooting data while
waiting for the scan tool to show up and put it's laser focus on the broken area, whether a marginal part during open
loop operation is being compensated for once the engine warms up & the system transitions into closed loop mode.
Or whatever. But, the less guessing = the more champagne we are going to spray around when we get to perform a
premeditated fix based upon our careful observations. :0)
Could we do even better? A: Yes. I'm not a TBI guru, but supposedly there's some sort of in-line adapter tee thingy
that gives you a test port to monitor fuel pressure while the engine is running? If this is true, then imagine
monitoring the fuel pressure + timing light + vacuum gauge as the engine goes from fine at 1200 to 1000 to stall?
Spark + Fuel + Compression = it has to run. Even at 1000 rpm during the warm up. (!)
*
Important history of operation question: Since you have owned this truck, was there
ever a time that it was able
to go from a cold start to fully warmed up without this failure occurring? If so, can you tell me when it went from
good to bad? (Before some new tune up parts were installed last spring, a valve job a year ago...or maybe it was last
August 3rd at 4pm, right before a fender bender in the Piggly Wiggly parking lot? And it was never the same since?
I'm being serious. You mentioned that the truck is running better than it has in 'the last 10 years'? The key to tightening
up the problem description is to figure out as close as you can to when it was the last time the machine was good versus
the very specific symptom that it has now? Once we have the moment in time where the problem started nailed down, now
we can try to figure out what related changes occurred during this state transition/window in time that we need to focus on
first.
You see, there are numerous
Possibilities for the root cause of this misbehavior, but instead of random shotgunning, we want
to organize all of these Possibilites into a list, from Highest to Lowest
Probability. Once we establish that, *then* we should
be able to come up with some structured testing that will allow us to further tighten up the problem definition before ordering
a part. Or 2. But not 2 dozen parts. :0)
For example, if everything was
fine until the fuel pump had to be replaced, and subsequently we have this strange symptom, I'm going to
put everything related to that maintenance as a higher probability than, say, investigating the charging system or performing a
compression check. (Not removing any Possibilities from the table, but just trying to assign relative Probability to each individual link in the
internal combustion success chain.)
Now, it's quite possible that the problem came on slowly over time, instead of a step function. Not only is that a valid observation, but
it also can help us eliminate the stuff that is known to work perfectly right up to the moment that it fails completely. (Like a light bulb.)
And I've also worked on stuff that the new owner had never personally seen work properly. Ever. Bought the thing as a 'basket case' on purpose.
Or bought a real cream puff in good faith, but the pretty (yet unreliable) vehicle with a beautiful facade had stayed in the PO's garage all
those years for a reason...but this detail wasn't disclosed by the seller?
Normally these old lemons resist repair due to nested, overlapping issues. Think of a chain where, instead of a single bad link, there's 3
or 4 weak ones that you have to weed out, one failure after the other, before you finally end up with a reliable chain with hundreds of good
links left to do your bidding. Pretty rare, and unfair, but it can & does happen.
Subject Matter Expert time:
* Let's say that all the normal stuff (grounds, wiring harness, normal tune up bits, TBI operation) is verified good with the scan tool.
Is it possible for there to be 1 or more corrupted memory locations in the ECU's PROM that have goofed the values in a small number of
specific cells that are only mapped to with the right combination of engine coolant temps, MAP, and rpm? Is this memory parity checked or
ECC protected, so that if there was a failure it would be detectable as a parity failure during power-up self-test and throw a Pxxxx code with
a MIL/SES light? Or there's ECC that allows the system to fix single bit failures on the fly? (See attachment for visual representation of
discrete cells in the lookup tables.)
* Or, would a higher probability be that the ECU in the OP's truck was a Reman unit, and the PROM is correctly holding onto
one or more cells filled with bad data (from an latent corrupted image) written into it during the remanufacturing process?
To me, this would both explain a lack of a power-up self-test failure/Pxxxx code, yet at the same time giving the engine something to starve/choke
on during warm up?
I ask this because the FSM was big on the dealer tech checking for/updating the ECUs with the latest version of the software? Is there an old
technical bulletin that describes a driveability issue like this fixed with version Foo of the code?
EDIT:
* Is this a case where the aging capacitors are weakened, and now we've got all kinds of RFI and/or ripply AC on top of the DC
that we're trying to run this computer with? Got junk leaking into one or more of the +5.0 volt reference voltages that feeds into
the sensor, and therefore could cause the resulting noise-modulated output to mislead the ECU after the A/D conversion step?
NOTE: I used to think of these computers as always the infallible victim of bad data from the sensors, but decades after they were initially built
maybe we have to consider the possibility that the ECU
can be the perpetrator, not the victim? (Last resort of course.) But I'm just trying
to reason through why it seems like the computer is stepping into a corrupted cell, and that is what we are smelling?
* What ratio is the 'authority' of each of the input variables during Open Loop operation? ECT vs MAP, vs TPS? In most systems
like this some inputs make for small changes in the output, whereas other inputs can make a small change & cause a major
change in the output? Hint: Given that the TBI is a wet intake manifold with no exhaust crossover heating, I'd assume that
calibrating fueling vs engine temp during warm up is touchy/critical to both meeting ever-tightening emissions during the
cat warm-up phase, while at the same time delivering reasonable driveability for the owner? A bit of a calibration tightrope,
where if the engine bay is just a little off it will show up here first?
* Last but not least, the OP mentioned having to sometimes spray a little starting fluid to get it running again once it suddenly stalled
out?
****
Rustjunky67, as you can see, if you can provide an answer to when the truck went from good to bad with this particular issue,
the more we can tighten up the problem statement, moving from exotic wide-ranging theories towards concrete, minimal-cost
focused fixing. No doubt that what I wrote above will generate more questions than answers, but that's not a bad thing when
troubleshooting a stubborn issue like this.
But I wrote all this because if the engine starts fine, always stops at 1000 rpm, and then warms up & runs right the rest of the
day, then we aren't looking at gross mechanical issues, are we? I'm thinking that by watching a vacuum gauge, a timing light,
and (hopefully) watching a fuel pressure gauge while the engine stalls we can come up with some ideas to prove/disprove with
the scan tool.
SBC, TBI, & ECU gurus, please chime in with what you think.
Thanks for your time & attention to this troubleshooting opportunity...