Diesel Place banner

Trying to avoid DS4 pump replacement

15K views 84 replies 12 participants last post by  OkDually 
#1 ·
Hi, a newbie here. However, I've posted stuff on many a forum about other stuff in the past and people are always super helpful, so thanks in advance.

My skill level is "okay" amateur mechanic and trouble shooter.

I'm posting maybe a little prematurely, but hoping it might yield an couple easy answers before I dive in too far.

I have a ~'94 K3500,about 80,000 miles on it, a fair bit of customization from previous owners. For years it's been started usually with a squirt of starter fluid on the air filter--probably just bad glow plugs since I just checked the controller. (I have a new set ready to install). Once it starts, it's always run pretty well. Until....

A couple weekends ago I started it after not having run for prob 6 mo. and checked everything to go do a haul run. However, in the process I ran it out of fuel out in the sticks and had to put some red-dye in it to get back to civilization. I did get it started okay after bleeding, etc. it ran fine, and it seemed ready to go. Fueled up.

On the run, maybe 50 miles in, at some point it started acting weird, surging, basically, as it turned out, full on and full off depending on whether I touched the pedal. This was on the highway mostly, no stopping, so the surging was a little subdued and I'd say manageable. (Later, looking around at diagnostic websites, that appears to be a sign that the fuel injection pump has a problem.) Anyway, I pull into a gas station a bit later and take it out of gear and foot off the pedal, and it goes full throttle. I had to turn the key off to make it stop. I fueled it up, and it managed to start again after a little fussing, got it in gear, and continued, with surging and fingers crossed that I could limp home.

So I keep going since I'm pretty close to my destination, and a few miles short the engine quits. Some cops happened by and pushed me out of the intersection, and I let it sit for a while, waiting for a buddy. When he shows up we got it to start with a lot of starter fluid, but the best it would do is idle really rough and inconsistently. I had a spare PMD (it's remote mounted) and that didn't do anything either, so I declared it dead and had it towed back to a local shop to diagnose. Oh, and my girlfriend, who was along, was not pleased, needless to say.....

A week later they apparently threw a code reader on it and all they told me was that it had a "bad injection pump" and "bad lift pump", quoted north of $3500, and charged me $130. I'm going to throw a reader on it again myself as soon as I get my hands on one.

Anyway, it's supposed to have a DS4 on it and I got a set of the parts including a reman at my disposal. I proceeded to watch some pump installation videos and whatnot, and I feel pretty comfortable with doing the job. However, and here's where my questions lie:

I saw snippets of repairs that suggest it could be the optical sensor in the pump (replaceable, though some don't recommend it), which would be a lot easier since the pump can stay installed and more or less correctly timed. Also, the master fuel solenoid could be faulty. So I pulled off the top cover off the pump, and I found that the fuel was not all the way to the top in the chamber. I'm used to seeing that on other pumps I've had apart. A footnote: the lift pump seems to be fine since the fuel bleeder knob spits ample fuel out when I open it and crank. There is also a brass "T" valve installed in a fuel line somewhere and opening that yields fuel (I don't know if this is a customization on this vehicle).

So, is the fact the fuel chamber wasn't full indicative of anything?

Anything else that jumps out?

Secondarily, the turbo is spitting relatively a lot of oil into the intake, since it's basically "damp" through and through. It's a reman turbo that's in there, with less than 10k miles on it. Worn bushings I'm guessing? Is it a critical thing to have a little oil coming in, or do I put it on the todo list?

Okay, thanks!! I'll post more tomorrow.
 
See less See more
#43 · (Edited)
Alright folks, I have more knowledge, but maybe not anymore wisdom (yet)

(thanks in the interim for additional great comments--I guess I'm obsessing with figuring this out without pulling the lower intake manifold...)

I threw a klunky USB 2 channel scope on the PMD wires, and here's the report.

A good primer on what pins are what (found elsewhere here on this forum):

635475


So I'm looking at

B, which is the high powered fuel metering solenoid command signal (pulls the solenoid coil low, and the high side is apparently D, connected to Batt+ via keyswitch relay, as is the fuel shutoff solenoid)
B is also monitored by the PMD and provides some feedback to the ECU via A I believe, and also I think terminates the high powered B pulldown transistor (i.e. the ECU doesn't terminate B, it's terminated by some mechanics in the pump and a circuit in the PMD). If anyone can confirm this, that'd be great.

E, which is the pump-on command from the ECU, supposedly timed to crank angle.

A, which I think is the feedback for what B is doing to the ECU.

Note: when I turned the key on, and then cranked after the glowplugs, I would get these following traces, which fired the pump 3-4 times, no more and no less. Then nothing. Even after cranking a bit more. However, I could turn the key off and then do it again, so something is "saving state" at 3-4 firings and causing the ECU not to command any further fuel pulses.

Additionally, if I disconnected the optical sensor, which is supposed to put the thing into limp mode, I get NOTHING.

{WAIT--something wrong in my notes with the scope traces--getting new ones and will update}

So that leaves a few questions:

1) is the PMD doing the right thing? It appears to be, except for the delay.
2) what is the delay about? Now that I'm thinking about it, it could be the pump's stepper motor adjustable delay, though the ECU sets this.
3) what's happening after 3-4 firings? Why is the ECU (apparently) thinking things aren't right?

After all this I looked at the codes a few times, and it would throw a 13 and 62.
13: "Engine shutoff solenoid circuit fault"

hmmm.... I wonder if 13 is causing the 3 pulse shutdown.
 

Attachments

#45 ·
Okay.... got it squared away.

There are three signals the PMD is involved with.

A: "inject" signal from ECU to the PMD. The PMD takes this and energizes the fuel metering solendoid via B.

B: fuel metering solenoid on: PMD to Pump, pulls one side of the solenoid to gnd.

C: solenoid control: feedback to ECU from the "poppett valve" by way of the PMD. Apparently the fuel pressure and the fuel metering solenoid work to move this valve around and it's state is sensed (by way of B I think), and that is conditioned and send back to the ECU. In other words, this pulse is "on" when fuel is being injected.

The ECU according to other doc's I've recently seen, then turns of A to shut down the fuel metering solenoid. Previously I suspected the PMD could do this internally but this appears not to be the case.

Note: The ECU for some reason is only commanding 2-4 "inject" cycles before stopping. Key off-on resets it.

Here's some scope traces. These are from crank on to nothing.

Yellow is B (solenoid).
Blue is E, feedback
You can see there is some lag on the feedback turn-on, through the turnoff is pretty simultaneous. I suspect that the internal ECU controlled stepper-motor "advance" is the reason for that.

635482




This one is:
Yellow A: ECU's "inject" signal
Blue E: feedback

The lag in the fuel injection is again visible.

635481



Lastly, this one is
Yellow: A (ECU "Inject")
Blue: B (Solenoid)

There doesn't seem to be any delay there, and not surprisingly, B is A "amplified" (i.e. the fuel solenoid is activated as soon as the signal comes in from the ECU.

635483



So this all looks good as far as the PMD is concerned. Questions are:

1) Code 13--is the fuel shutoff solenoid bad, as has been suggested?

2) Is the fuel metering solenoid doing anything? Although there are only 3-4 inject cycles coming from the ECU, there seems to be some delay in the feedback and that's as I understand it, supposedly because the poppet valve is doing something. i.e. it suggest the pump is actually pumping, just 3-4 cycles at at time.

So what's causing the ECU to send ever shorter signals and ultimately stopping the fuel pulses?
 
#46 ·
If you're getting DTC13, I know of no other way to verify the operation of the FSO solenoid than to swap it out or gut it. And even swapping isn't 100%, as there's no way to be positive that the new unit is opening correctly once installed. Unless you've tested it on another engine maybe. That's why I chose to gut mine. I wanted to be damn sure that there was no further possibility that it was the culprit. It's pretty easy to do. Just note that you will always find a DTC13 present if you are running a gutted solenoid. But once the underlying problem(s) are resolved, installing a new FSO unit is quick and easy.

I cant recall, but is there any IP inlet pressure sensing input signal to the ECM? If so, maybe a no-fuel condition might be why the fuel metering command would cease after a few pulses?
 
#48 ·
If you're getting DTC13, I know of no other way to verify the operation of the FSO solenoid than to swap it out or gut it. And even swapping isn't 100%, as there's no way to be positive that the new unit is opening correctly once installed. Unless you've tested it on another engine maybe. That's why I chose to gut mine. I wanted to be damn sure that there was no further possibility that it was the culprit. It's pretty easy to do. Just note that you will always find a DTC13 present if you are running a gutted solenoid. But once the underlying problem(s) are resolved, installing a new FSO unit is quick and easy.
635512


This is the page I found on the the FSO in a GM Tech Manual which explains a lot but not quite everything...:

!
Also this one is good to have:

635513

Unfortunately this schematic doesn't distinguish between the various electrical components of the pump.

Anyway...

So the ESO (aka FSO/fuel shutoff solenoid) is supposed to be "hot in run" on one side (it is), and the PCM (been calling it the ECU, might as well get with the lingo...) pulls the other side to ground under normal conditions, energizing the coil. In my truck I pulled the ESO/FSO out (allen wrench plus some gentle persuasion from a channel lock), turned the key on, and when I plug it in, it persuasively retracts. The coil reads 17 ohms or so.

However, what I'm not understanding is what terminal E16 is in the PCM, that is supposed to read ignition voltage or it throws a code 13. If it's "BC16" in the first diag, or "C16" in the second, and if the PCM is pulling that to ground to energize the coil, what voltage is it supposed to be reading, unless there's an internal resistance to make a resistor divider in there somewhere. I see about 0.3V at the pulled down side of the coil at the coil, but who knows what's downstream.

I cant recall, but is there any IP inlet pressure sensing input signal to the ECM? If so, maybe a no-fuel condition might be why the fuel metering command would cease after a few pulses?
Here's what the first manual has for those subsystems:

635514
635515


It doesn't look like the fuel control is looking at inlet pressure, which I guess would be the "Boost Sensor". The boost sensor is only used by the PCM to control the wastegate. Nevertheless, I have it unplugged since I've got the top manifold off. I suppose I can try plugging it in and seeing if it gives more pulses.

If the crankshaft sensor was bad, that would throw a code (same manual). Same with the optical, though as noted before, if I unplug the optical, I get NO fuel pulses even though it's supposed to go into a limp mode (though now that I think about it, I read it's supposed to crank for 10-15 sec in that mode before it does anything, so maybe that's why I'm not seeing it.)


Sooo.... anyone have the PCM source code????!!!!
 
#47 ·
Tip> The shut off solenoid can be gutted by taking a small snap ring out and removing the plunger.. The solenoid is basically a useless device that can kill your truck at any time. The IP will automatically shut the engine off when the key is turned off, without the solenoid.
Some will say it's needed but it really serves no purpose. Please don't confuse the shut off solenoid with the fuel solenoid. Shut off is the tall round cylinder thingamabob on the front top of the pump.
 
#49 ·
There are manuals listed in the 6.5L FAQ page that you can download. You can browse to see if those might have the source info you need.
 
#52 ·
#55 · (Edited)
Found this out of some manual:

Concerning the FSO:
CIRCUIT DESCRIPTION
When the ignition switch is in the "OFF" position, the engine shutoff solenoid is in the "No Fuel" position. By providing a ground path, the Powertrain Control Module (PCM) energizes the solenoid which then allows fuel to pass into the injection pump.

DTC 13 WILL SET WHEN
Ignition voltage on terminal "E16" when PCM is requesting engine shutoff solenoid "ON."

ACTION TAKEN (PCM will default to)
A current and history DTC 13 will be stored.

DTC 13 WILL CLEAR WHEN
The fault condition(s) no longer exist.

DTC CHART TEST DESCRIPTION
Number(s) below refer to circled number(s) on the diagnostic chart.

Check for open circuit from ignition switch to the solenoid.
Check CKT 981 from solenoid to PCM for open.
DIAGNOSTIC AIDS
An open in CKT 981 or 339 will cause a DTC 13
Also a no start condition will exist. <-----
Hmmm.... might be on to something here.

So E16, which I believe is also the pin pulling the solenoid low, is showing voltage when it's not supposed to be, resulting in a "no start" condition.

I don't know how E16 can be high if it's energizing the solenoid, unless the solenoid is actually shorted to ground and the wire is broken somewhere, leaving E16 floating. It does however sound like E16 is both pulling down that pin and sensing it at the same time. It must have some sort of pull-up on it though.

I think what I'll do is with the FSO unplugged, short wire 984 (supposedly connected to E16) to ground to see if Code 13 goes away with the key on. Stay tuned.

More (from some other forum):

File In Section: 6E - Engine Fuel & Emission

Bulletin No: 87-63-04

Date: April, 1998

Subject:
Hard Start, No Start, DTC 13 or P0215, Service Engine Soon Light On (Replace Electronic Shutoff Solenoid on Vehicle)

Models:
1994-99 Chevrolet and GMC C, K, G Models
with 6.5L Diesel Engine (VINs F, P, S, Y - RPOs L65, L49, L56, L57)

Condition

Some owners may comment that their vehicle is difficult to start or will not start. In addition, the Service Engine Soon light may be illuminated.

Correction

If DTC 13 (1994-95 model years) or PO215 (1996-99 model years) diagnostics lead to injection pump replacement, a new on-vehicle replacement procedure is available for the ESO (Electronic Shutoff Solenoid. This procedure eliminates the need to replace the entire injection pump.

Parts Information

Description
Solenoid, Electronic Shut Off P/N 12561591
Does anyone have a copy of Bulletin No: 87-63-04?
 
#56 ·
Found this out of some manual:



Hmmm.... might be on to something here.

So E16, which I believe is also the pin pulling the solenoid low, is showing voltage when it's not supposed to be, resulting in a "no start" condition.

I don't know how E16 can be high if it's energizing the solenoid, unless the solenoid is actually shorted to ground and the wire is broken somewhere, leaving E16 floating. It does however sound like E16 is both pulling down that pin and sensing it at the same time. It must have some sort of pull-up on it though.

I think what I'll do is with the FSO unplugged, short wire 984 (supposedly connected to E16) to ground to see if Code 13 goes away with the key on. Stay tuned.

More (from some other forum):



Does anyone have a copy of Bulletin No: 87-63-04?
What you posted is Bulletin 87-63-04 (y)
 
#57 ·
I guess I can't help you here. Bypassing the shut off solenoid takes maybe 5 minutes. Diagnosing it through schematics, wiring diagrams, ECU functionality, electrical probability, shorts, crossed wires, and then asking the forum for suggestions takes up hours and hours of time. Not very efficient.
 
#58 ·
I agree it might not be efficient in some ways, but the fuel solenoid appears to be working. As I said, it retracted with authority when I took it out, hooked it to the connector, and turned the key on. When it's installed, there's nothing stopping that same retraction, and I can hear it happening too.

I could just start replacing everything, but that's not efficient either, in other ways.

But what is a problem is that it's throwing a code 13, and also that the PCM is only commanding 3-4 injection cycles on the fuel metering solenoid before giving up for some reason (which could be related to the code 13). Code 13 is that it thinks the FSO is not functioning because it should read low voltage at pin E16 and it's not.

I can imagine the PCM is shutting down the pump on account of an FSO error because it doesn't want the pump to run dry for very long if the FSO is indeed not letting fuel flow. Does this make sense?

E16 either it isn't low because of some bad connections (open or high resistance circuit most likely), or the PCM internally has a problem and isn't reading the voltage on E16 properly.

So I have two experiments in mind to determine this:

1) is to short E!6 at the FSO and see if it clears code 13. If it doesn't then I can pull out the PCM and check the voltage at the PCM. If that is low, then there is a PCM problem.

2) Get another solenoid (looks like $80-$100 will do it), hot wire the installed one, plug in the new one, and see if that clears code 13. If so, then I can install the new one. If not, return it I suppose.

However, I'm also getting the understanding that people are able to start even with code 13, so maybe there is a whole different issue causing the PCM to stop firing.

Anyway, I hope you see the method in my madness. In any case, thanks again for all your input.
 
#59 ·
You did an amazing job the trouble is we don't have any reference point on those signal to compare it to.
In other words, we don't know how a good signal looks like.

Is the truck running at all?
Sometimes, code is just there.
If the truck runs, I will drive it around for at least a couple of days see if the code clears.
 
#60 ·
Thanks, Trying....

Right now the truck will not start, period. However, I promise that once it does, for the greater good of all 6.5 mechanics, I will post scope traces!!

I'm waiting on replacement for some warranty/bad Bosche glow plugs (wth?! 3 of 8 tabs broke off to a light touch), so I'm not trying to fire it up today unless I put in some bad ones temporarily.

Should I clear the PMC with a hard/battery disconnect to get rid of the Code 13? Is there a better way?

I just got done testing the PMC connections to the FSO, and they are good. With key on and the FSO unplugged I get 0.030V at the pin (D16 in this case as it turns out), and 0.3V plugged in, both of which are what I measure at the FSO as well.

Interestingly, when I pulled the glove box there was a mouse nest on top of what appears to be some sort of relay bank on top of the vents. Wonder if it pee'd on something and it's causing a short of sorts.... or chewed a wire. Good grief.

I also case grounded the PMC to battery ground directly briefly, and that gives a 0.030V (30mv) difference in the measurements, but that's not too surprising. I wouldn't think that would be enough to upset anything. Anyway, no difference. Still getting code 13 (and 62 as well, BTW).

Next: I'm going to grab a spare FSO and try plugging it in, and also do a hard/battery reset.
 
#61 · (Edited)
If it’s a ‘94 (or any OBD1), codes can be cleared by holding both the brake and gas pedals to the floor, then turning the ignition key to the run/on position and waiting for 10-15 seconds. Then key off, release pedals, key back to on to recheck codes. You should get the 12-12-12 result if all cleared.

Oem FSO solenoids were on eBay for $50 recently.

If you’re willing to spend that $50 on the troubleshooting process, and given your results so far, I would strongly recommend installing a gutted FSO to cross that off your list as a potential root cause. You’ll then definitely get a constant DTC13, but you could then confidently proceed to look elsewhere for the source of the no-fuel issue.

If a gutted solenoid allows the truck to run, you can install your new solenoid and keep the gutted unit as an invaluable troubleshooting tool.

Despite the apparent design redundancy of this solenoid, there’s been some interesting past discussion here that suggests that the FSO can be commanded to close by the ECM in certain dire circumstances in order to protect the engine. No definitive proof to offer, but that certainly makes good sense.

The truck will run all day with a constant DTC13 (ask me how I know!), so that error code is NOT the source of your problem. To be clear, the ECM is not defueling the IP due to seeing the DTC13 present.

Also, the quality of products from the wiring harness supplier for 94 (and 95, iirc) is known to be problematic. If push comes to shove, high quality replacement harnesses are available through members/vendors on this site. Not cheap but reasonable, and that sure ended my occasional “dying in the intersection due to no-fuel until I wiggled the harness just right” problem. And this after EXTENSIVE probing with a meter to try to find the intermittent short/open that was causing the issue. (I also tried a new ECM to no avail. But another very useful spare to have for troubleshooting purposes.) Truck has actually been rock solid dependable for 20k since the engine harness replacement. Nice to be able to spend time on upgrades rather than constantly troubleshooting.
 
#63 ·
If it’s a ‘94 (or any OBD1) ... <snip>
Thanks RedShift, very useful. I should add that on this truck an exhaust gasket failed and was blowing very hot exhaust up behind the engine for a bit, enough certainly to smoke some insulation here and there. It was some time ago, but maybe that has something to do with it now.
 
#62 ·
All of these started because of using starting fluid and run out of fuel.

Did you check if the fuel is there?
You said, you replaced the OPS or not?
OPS controls the LP when engine is running on a 94 and it has caused a lot of issue in 93-95 model year.
I would replace with ACDelco OPS ONLY.

Yes, OPS can show oil pressure but the part that control the LP can still be bad.

Are you sure the IP in there is good?
The PMD is good?
 
  • Like
Reactions: Sirleitet
#64 ·
All of these started because of using starting fluid and run out of fuel. Could be... but why?

Did you check if the fuel is there? Yes
You said, you replaced the OPS or not? Did not
OPS controls the LP when engine is running on a 94 and it has caused a lot of issue in 93-95 model year.
I would replace with ACDelco OPS ONLY.

Yes, OPS can show oil pressure but the part that control the LP can still be bad.

Are you sure the IP in there is good? No, but I have no reason to believe at this point that it's not.
The PMD is good? Appears to be.
Are you saying the OPS controls the LP or the IP? I am sure the LP is good because I get ample pressurized fuel coming out of the return line at the IP, T valve, and fuel filter bleeder valve when it's running. I can also hear it, since it runs for a second or two when done cranking while the oil pressure signal is still over the threshold.

What is definitively THE problem is that the PCM is only sending 3-4 fuel pulses to the PMD-->IP, and then shutting it down. The question is why is it doing that? A quick look at the PCM source code would probably answer that... sigh. Has anyone ever seen that published anywhere?

Tangentially, shouldn't the OPS also throw codes if it's working not right? I can see that the PCM could be using the OPS for some sort of feedback, but it's supposed to be the case that if you simply disconnect the OPS, things should run in a limp mode. When I do disconnect it, I get no fuel pulses, but I read somewhere I might have to crank for 10-15 sec for them to start if it's in "no OCS" limp mode (I'd be there's a software timeout).
 
#65 ·
OPS controls the LP (Lift Pump, it is an "L" - ELL not "I) when engine is running.
There is no fuel going during cranking in a 94, unless, it is mod with an additional relay.
So, if OPS is bad (the one that controls the Lift Pump), the LP will not start running as soon as the engine catch.
That is why we put 12V direct into the Lift Pump with an "L" not "I" when testing the Lift Pump with an "L" not "I".
Then we bleed the air and start cranking.
Everything have to happen at the same time, now you have fuel to the Injection Pump with an "I" not "L".
Then the Injection Pump electricity needs to be there.

Did you ever replace the Glow Plugs?
They may already burnt due to the starting fluid.
 
#66 · (Edited)
OPS completes the circuit to the LP only. The ECM bypasses the OPS to run the pump during cranking to start, then the OPS takes over providing current to the LP when the key is released to the run position. This allows a low oil pressure condition to cut power to the LP in the event of an accident or oil supply failure, potentially saving the engine and/or minimizing the chance of fire.

As previously noted, if you have fuel flowing well at the filter drain t-valve during cranking, your LP is working. If the engine were firing during cranking but dying as soon as you release the key to the run position, that would point to a possible OPS failure. Since you’re getting fuel to the IP during cranking, but no fire, it seems unlikely that the OPS has anything to do with this. If ever in doubt, you could simply jumper it’s contactor terminals to confirm.

Afaik, the source code has not been published. But I believe there has been some work on that front by members here. Ask Buddy or Quadstar.
 
#67 ·
The OP has a 1994 with OBD1.
LP runs WHILE CRANKING via the CRANK FUSE and then when oil pressure has built up.
The PCM has no control or input from the LP.
96 and up OBD2 went to PCM control and there is a code for low voltage @ the LP.
That code still has no effect on the engine running.
The OPS on an OBD2 is a backup/redundant part and is really not needed.
In fact the 6.5 diesel Vans after 1999 don't even have one.
 
  • Like
Reactions: OkDually
#68 ·
Update:

The both-pedals-down-clear-PCM-OBD1-codes procedure didn't work unfortunately. Tried it a few times and with some minor variations. Perhaps this was something they added in a later software update?

To attempt to clear PMC codes, I disconnected the battery for a while, swore at it for a bit while I had to put the GM cr-p side post stripped battery bolts back in, and indeed, got 12's.

Cranked it a few times with some of the glow plugs still not in (thanks, Rock, still waiting), and still got 12's.

Not sure why the code 13 wouldn't clear on its own, as indicated it would in the GM service bulletin. Perhaps a software update?

I'm not sure what all that means yet, but the next step is to throw the scope back on it.

I'm going to speculate that DieselPro's diagnosis some time back of a broken armature on the fuel metering solenoid is looking like where this is going, but the 3-4 fuel pulses and out PCM response is one I don't want to be defeated by.

I haven't read up on PCM software updates, but it this something I don't want to touch? I'm sure I can do it, but I don't want to create a bunch of needless unrelated headaches.
 
#69 ·
With the scope attached, I would try different "known working" PMD's w/resistor installed, and see if you get any different readings on the pulses.
Also, verify that the ground wire for the OEM PMD harness is attached to the top of the IP and is tight
 
#70 · (Edited)
Might as well clean and check the ECM grounds atop the #8 cylinder stud while you’re at it. (If you havent already.).

Very strange that the DTC reset procedure didn’t work. Here is where I should probably mention that when I bought my ‘94 it was not running due to a no fuel condition. After troubleshooting for a month worth of evenings I gambled on trying a non-returnable replacement ECM. Fired right up afterwards. If you can find a known good spare to swap in as a test, I’d sure give this a shot. Wrecking yard, possibly?

AFAIK, there have never been any updates for the ECM/PCM firmware offered by GM. As mentioned earlier, Buddy and Quadstar have done some heavy lifting in this arena, and are able to provide retuned versions of the firmware on the OBD1 chip.
 
#73 · (Edited)
AH!!! Thar's Yer PRAHBLAHM!!!

And the prize goes to.... DieselPro! Ding! Ding! Ding! Ding! Ding!

Well, let's not get carried away.

So I took off the lower manifold and disassembled the fuel metering solenoid end cap to examine the armature, as suggested by DieselPro and referencing this post by quantam mechanic:


Sure enough, after looking closely, there is a small hex cap screw that attaches the armature shaft to a steel plate. As the solenoid is energized, this plate I believe closes a magnetic circuit, and that is somehow sensed by the PMD and fed back to the PMC.

(Theorizing here on the functionality FWIW: In physics terms, magnetic fields like to travel through certain materials, like iron/steel, much better than say air or diesel fuel. They go around in a complete circle, through the easiest path, much like electricity. Magnetism and Electricity have a lot of influence on each other, without getting into the weeds. When there is an "air gap", which is the case when the solenoid is retracted, the field behaves differently and has less magnetic force than when there is no air gap. Thus, the way this thing seems to work, is it's electrically energized by the coil, causing the armature shaft to pull in and enable the pump cycle, while at the same time closing the air gap. That in turn affects the magnetic field and hence the current going through the coil, which is sensed by the PMD on the same wire that is driving the coil. Or something like that. Scope traces forthcoming for comparison.)

Anyway, This screw is BROKEN! So the plate isn't closing the magnetic circuit.

Pic's below:

635861


(Above) Sort of an exploded view of what you have to pull apart. The two pins on the left are the terminals that connect the wire harness to the solenoid itself. I sort of jacked up one of the insulators slightly trying to get them out.


635862


(Above) Hex cap screw in red is suppose to go through the hole in the metal place and screws into the end of the armature shaft. The screw is broken. It is lock-tite'd in there, so we'll see if we can get 'er out.

As the armature shaft goes in and out, the plate moves with it and opens and closes magnetic circuits that include the plate and whatever the coil is wound on (not shown).

I'll get it put back together and report back.

Hex cap screw: $0.10 at ACE (probably more like $0.02 normally)
New Injection Pump:$3500 quoted.

We'll see...

P.S. I just managed to get the broken screw out of the end of the shaft. It is LEFT HANDED. Yay.. dunno if ACE has left handed metric cap screws or whatever it is.... OD .108", 2.75mm. I heated it up with a torch to burn out the locktite. I might reheat and quench to re-establish the hardness before reassembling, but it's probably not necessary.
 
#74 ·
I am quite sure it's not left handed. Just regular threads.. The shaft shows lots of wear so it might not work, at least polish it. The travel is only .005" so if it's not perfect it will throw a code.Just don't turn the solenoid in the head or that will be the end of it. (It's set on the bench with the head disassembled.) So folks this is what causes a runaway . > Broke or loose armature screw.
 
#75 ·
I am quite sure it's not left handed. Just regular threads..

The shaft shows lots of wear so it might not work, at least polish it.

So folks this is what causes a runaway . > Broke or loose armature screw.
Oh, it's definitely left handed. It appears to be a 4-40 left, ~3/8" long. Can't find a 4-40 left anywhere though, 6-32 left, yes. I'm guessing it's left handed because the armature shaft wants to precess for some reason. It is also a metric 2.75xsomething...

Yes, polish.

Yes, broken or lose armature screw.

The point of most of this thread for me is (since we're not running yet...) to determine why it causes the problem (runaway or no start). It appears to be the case that the pump technically could operate, but somehow the feedback to the PMC is getting screwed up.

And yes, in a way, it's a big waste of time like you said, but I didn't want to start tearing things apart and also maybe it's now documented... :) Plus, I like the the challenge.

Anyway, as soon as I find the screw or another armature shaft (with a right handed thread), I'll put it back together and cross my fingers. Stay tuned.
 
#76 ·
>> shaft wants to precess for some reason <<<< If you meant to say the plate is recessed the armature is not any good....We discarded shafts that had that much wear on the shaft and never tried to reuse one that the screw broke..... It just took way to much time to redo the whole pump, when it could have easily been replaced the first go around. They are quite precision and would not pass calibration specs once ran. I am curious to as why they runaway and just don't die when broke. Typically the truck will sit there idling perfectly and you just barely touch the throttle and then away it goes wide open. Your best bet now is find a good used one from one of the members here... A new one is a little expensive and the money would be better spent on a new pump.
 
#77 ·
Some additional updates from my research:

Apparently Stanadyne had over time discovered issues with the fuel solenoid armature design, though screw failure doesn't seem to be one of them. They came out with a new design in early 2000's is seems, and that one was designed to have less friction.

From http://www.stanadyne.com/dealerportal/ssi/english/Service Bulletin/550.pdf

635897

And from http://www.stanadyne.com/dealerportal/ssi/english/Service Bulletin/561.pdf

635898


My pump has the armature assembly on the left above. The one on the right seems to be available for purchase through some dealers for about $80, but I'm not sure if it will replace the old one. Anyone know?

Interestingly, there is this "anti-rotation pin", which I think speaks to the left handed thread on the armature assembly.

I just called the regional diesel service shop I use and they said they can probably get most any part from Stanadyne. Stay tuned.

Also, I haven't found any DS4 parts diagrams or lists with part numbers. Anyone know if they are to be found publicly? (I hate it when they do it dealer only... I'm a DIY guy)
 
#78 ·
Ah ha, someone else out there has discovered how it works too:

"The PCM sends a digital Fuel Inject control pulse to the FSD, which is (a) a power driver that applies battery current to the FS solenoid coil via the collectors of the paralleled PNP driver transistors, and (b) a comparator circuit that monitors that current for the changes that result when the flux density changes as armature\plunger begins moving, continues forward motion, and stops as the valve seats - the resultant digital Closure-Time pulse is sent to PCM which starts counting clock ticks for injection pulse width timing - PCM has been counting ticks since sending the injection pulse to the FSD, so CT is included in Injection Pulse Width timing - FS is merely a coil, responding to applied power - power on is a 1, coil energized, power off is a 0, coil de-energized "

In short, when the armature screw breaks or for whatever reason the air gap doesn't close by the movement of the armature, it screws up the closure time pulse, and thus the fuel mapping, which likely explains why the pulse width kept diminishing in the 3-4 pulses I was seeing from the PMC before it stopped putting them out entirely.

Still trying to figure out what armature I can find, or the 4-40 left handed cap screw. I found some stainless ones, but they are only 70,000PSI, vs 140,000PSI for a standard (grade 8) socket head. Given that the old screw broke, I'm not so sure one rated for half the strain will do it...

Also, the armature is rather scored at two places along the axis, in a rotational manner. I can polish it if I can find a replacement screw, but I think I'd rather just replace the whole armature assembly.

Does anyone know if the newer armature design is compatible?
 
#79 ·
#80 ·
Best bet is to replace IP and call it a day.....Odds are the new armature will not work... The poppet lift is critical and it's not likely a replacement will still be in spec. They have to be set on specialized equipment. One other thing is the wear inside the IP where the armature scores the inside metal.
Spec is five thousandths.. Lift. Hard to set on the work bench with the proper equipment. Impossible to set on vehicle.
 
#81 ·
Dear friends,

I'm reviving this thread after a long hiatus.

Status update:
1) the truck starring in this thread to date is running. I put in a reman pump and it started right up. Okay. great. (It's got a knocking that gets scary when it warms up, but that's another issue...)
2) I acquired another very similar truck with the same engine and same problem. I.e. not starting. After doing the usual diagnostics I put in a reman pump, thinking it would start, same as previous, but no. So now I'm diving into this one. They are conveniently sitting side by side though, so I can go back and forth and compare.

What I've done is a) all the usual gremlin removal/ground inspection/etc tricks, which turned up nothing. b) gone high tech again and threw it on my scope.

BONUS! Here are the signal traces of the working truck referenced to battery/chassis ground, as promised last year:

Font Science Space Parallel Symmetry

click for hi res

This is off the first truck with the reman, running. Interestingly, trace "C" is supposed to be the ECU/PCM ground which is supposed to be pretty close to battery ground (plus some slight variability due to wire resistance), and it's bouncing around compared to battery ground something fierce. I'm surprised it's running at all. I may short pin C to the battery ground just to see if anything changes, and if so, this truck has a bad ground issue, but not bad enough to cause a failure it seems.

I haven't drawn it up yet, but I'll post a simplified schematic of how this whole system is supposed to work, since I've got all the puzzle pieces sorted out now. Interestingly, "B" is the fuel solenoid command from the PMD, but the PMD also monitors that pin at the same time as it's driving it, to see what happens, which is namely that the plunger of the solenoid itself pushes in and the small steel plate on the end closes an air gap. That in turn affects the amount of current needed to drive the solenoid, and in turn feeds back on "E" when the solenoid is in fact depressed or open. Or something close to that.

What I don't have is the ECU/PCM software, which is where the headaches lay. Read on.

On the second truck, when I go key off, on, wait for glow plugs, start, I capture consistently four pulses on "A", the ECU/PCM fuel on command pulses. Three normal looking ones, and a short one, and then it goes dead until I key off again. This is exactly what I was seeing in the first truck initially and also after I replaced the broken solenoid plunger. Clearly, the ECU/PCM is not happy about something, but it's not revealing what. In other words, it tries three times, whimpers on the 4th, and terminates it's attempts.

I found an old Snap On scanner from around 2001 and got it communicating with the OBD1 on this truck, and it throws a 49 and 62 as well as real time parameter data coming from the ECU/PCM

DTC 49 - Service Throttle Soon Lamp Circuit Fault (I read that could be due to a faulty throttle position sensor, but it shouldn't be a show stopper according to a service manual that discussed it. Additionally, the scanner shows the throttle position moving as it should when I press it).
DTC 62 - Turbo Boost Sensor Circuit Low (it's disconnected, should matter)

So what the heck is going on??!!
I might be having problems with the feedback ("E") on this reman.... shouldn't be though.... in an ideal world....
Additionally, the grounds seem a little floaty. I did manage to tie something together for a little bit which caused the ECU/PCM to send control pulses continuously (vs 4 times), but then it stopped.

Will a GM Tech1 scanner do more/tell me more than the Snap On? It does show various parameters related to the IP.

Additionally, I have no idea if I've got this new pump aligned within some acceptable range. I put it in the same as the old one, but there aren't any other alignment marks. Or did I miss them? What's the proper alignment when you don't know what it's supposed to be? Right in the middle of the range maybe?

Thanks!
 
#83 ·
As mentioned ensure the pump is level.What I usually do is center up the top mounting stud on the timing cover within the top elongated slot on the injection pump.
 
  • Like
Reactions: OkDually
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top