Rosie's return to glory

[TorqueHorse]

Keeping my cool

Well since I'm still waiting on those part to ship over to me, I decided to improvise a bit in the meantime. I found a generic GM coolant pipe fitting at the auto parts store for $5 and thought what the heck.


First I cut off the end that I needed and then sanded down the outside diameter so it would fit into the existing pipe.


Then I whacked it in there nice and tight


Then I cut off the end I didn't need with a dremel cut off wheel


Then I tapered what was left with a dremel sanding bit and then a pass with a metal straight edge to knock down any burrs and hopefully get as flat of a surface as possible for the plastic to bed into the aluminum housing.


Then put the Orings on that came with the piece (it is still a little bit longer than stock so I started with the double oring). Plenty of bed for the orings now and drips can't sneak past.


All filled up again and after the engine ran a bit. Look ma, no leaks.

[TorqueHorse]

Losing my cool

Well, I put everything back together, with no apparent leaks from the outside - check! Cranked it over by hand for quite a few times to hopefully get some oil moving and listen for any last hints of misalignment - check! It had been almost 1 year exactly sine I last turned her over; everything sounded good - check! This is going to be great, so I proceeded to turn her over, and - pflat! No joy Coolant is leaking in there somewhere still, and she stalls immediately unless I'm feathering the gas pedal

Here is a video of like the third try. Besides the stalling, and white smoke, you can see some bubbles sneaking past the gasket from the exhaust manifold to the turbo, and then drops of coolant coming out at the precat/downpipe connection. Whatever leak it is still seems just combustion to coolant, with the oil looking clean and emulsion free.

I'll try and run a leakdown test this weekend, but really at this point I don't see anything other than just pulling her again and getting magnafluxed for cracks, just to be sure. Probably just skim the deck and head too, and might even rebore depending on what the crack checking finds out (yes, I know, I'm building a time machine right now to go back and tell Matt from a year ago to do that to begin with ).

[378S2]

Ah man, sorry Matt that's frustrating.

Oil pressure looks low compared to mine at cold idle - not sure how useful or relevant that is?

[TorqueHorse]

Thanks George! FWIW, I looked back of old videos from before the teardown and the oil pressure looks exactly the same as before.
Results from some of the tests this weekend (cold engine for both):

Compression test, ~3sec/cycles
cylinder 1 180 psi
cylinder 2 180 psi
cylinder 3 180 psi
cylinder 4 180 psi


Leakdown test, TDC, Set pressure at ~85psi
cylinder 1 -1.79%
cylinder 2 -0.58%
cylinder 3 -1.97%
cylinder 4 -1.16%

On the leakdown test, I could hear/feel air coming out of the dipstick tube as expected, and also on cylinder 1 there was a bit coming out of the PVC valve hole in the crank case. But all of this says healthy engine, except for the fact that we actually see coolant coming out of the exhaust when running . So my next thought is maybe its a crack up in the head that wouldn't show during leakdown with the valves closed? Or a cylinder wall/jacket crack that is below TDC piston position, or only opens when it gets hot and expands.

Trying to do the coolant pressure test next, but I keep finding other weakspots in the hoses that I need to tighten. Once thats done, I'll borescope the cylinders while the coolant is under pressure and maybe I'll get lucky and see something.

[esprit888]

Not being there its tough to see where the leak is in the front of the head to block area but if you are 100% sure its coming out the exhaust (as in green) and not a build up of condensation from being idle for so long then that would be head off under normal conditions. One other 'possible' though remote area could be the turbo cartridge being cracked as that is the area of the leak at the front of the block ( i first wondered if one of the coolant lines was leaking or missing the copper washer but if its coming out the exhaust....) if its cracked in the right (wrong) place coolant could possibly be getting into the exhaust, though I've never heard of it on the M100 but have seen it on an Audi. As you say a good coolant pressure test should be able to reveal the location, and if you can do it hot (though that's not the 'safe' way) that would be even better.

[TorqueHorse]

Thanks for the advice! I hadn't even considered the turbo cartridge, though I did double check the connections to the turbo. Yep, still working on a pinhole leak in that crossover aluminum pipe right at the front connection of the prongeron. All my cleaning took out the corrosion everywhere that had been keeping the leaks 'shut'

[esprit888]

I also noticed your compression numbers were high (above new factory spec of 169) did you do the test with the throttle wide open or closed? And has the head been skimmed? ( minimum head height 5.114”)

[TorqueHorse]

Hmm, well I just checked from a year ago and I got the same 180psi before I did the tear down/rebuild. I didn't have throttle WO for either that old test or the new one. I didn't have the head skimmed and I doubt previous owner did (but who knows). I'm also using a rental tool from AutoZone, so I don't know what the accuracy is on those (but I did use a different one each time...)

Regarding the turbo cartridge leak. It looks like I could jury-rig a hose to bypass the turbo coolant feed/return lines. Does the turbo need that coolant in the first 30 seconds of start-up idle? The leak shows that quickly, so I'd be able to rule that out if so. I'm just thinking I can look in a cylinder to see evidence of coolant leaking from head/cylinder wall, but I don't know how I could see inside the turbo.

[378S2]

If the leak was from the turbo, and getting in to the cylinders, then it'd have to go via the intake valves wouldn't it? So if you pull the intake hose off and you find coolant in there you will have confirmed that it is a turbo fault. Right!?

[GeoffSmith]

What about disconnecting the air intake pipe from the wing to the throttle body and running it?

[TorqueHorse]

Oh yeah, ok I get it now. Nice, I'll try that, thanks folks! FWIW, I just got confirmation this morning that a new ECU is on the way so I'm hoping this one will work with my elanscan interface and can get some better pinpointing from sensor feedback as well.

[esprit888]

:agree: What about disconnecting the air intake pipe from the wing to the throttle body and running it?

Geoff, That won’t work if it’s a leak from the turbo cartridge to the exducer side and out the exhaust via the cast turbo housing as Coolant would still be circulating through the turbo. It would however eliminate the inducer/compressor wheel side and if coolant shows up at the pipe exiting the inner fender than bobs your uncle, but there is no boost at idle so pushing it through the intercooler and up the hill is a big ask. My money though is still on the head.

[esprit888]

Hmm, well I just checked from a year ago and I got the same 180psi before I did the tear down/rebuild. I didn't have throttle WO for either that old test or the new one. I didn't have the head skimmed and I doubt previous owner did (but who knows). I'm also using a rental tool from AutoZone, so I don't know what the accuracy is on those (but I did use a different one each time...)

Regarding the turbo cartridge leak. It looks like I could jury-rig a hose to bypass the turbo coolant feed/return lines. Does the turbo need that coolant in the first 30 seconds of start-up idle? The leak shows that quickly, so I'd be able to rule that out if so. I'm just thinking I can look in a cylinder to see evidence of coolant leaking from head/cylinder wall, but I don't know how I could see inside the turbo.

compression tests should always be done with the throttle flap WIDE OPEN. Counterintuitively this will lower the readings and they will likely be in spec. As I’m a lazy bugger and don’t want to read back over the whole thread were any engine internals replaced ? (Rods/pistons) The Isuzu cast iron block used on the M100 is a strong bugger and I haven’t seen one fail in away that would give you your symptoms.

[TorqueHorse]

Update, that there really isn't an update.
I was hoping that changing the coils, HT leads, and ECU might change the situation a bit, but it still stalls if I stop pumping the accelerator. It hasn't overheated again, but I still can't find if/where coolant is leaking. I can do a pressure test and it will hold 15, but not 18 psi, and I don't see where it is going (And I followed your tips and checked in hoses before and after running the engine and saw nothing). I've borescoped the cylinders while the coolant system is pressurized and didn't see anything obvious. There was fluid on top of the piston, but it smelled like gas. Trying to test the fuel pressure, but having a hard time getting the gauge on the feed line. I thought the return line would work, but get no reading there. Of course I'm using rental gauges for both of these, so I also have to wonder if the gauges/fittings even work right. I thought the ECU would help at least get better feedback on what is happening when I try and run it, but I'm having trouble logging there too (separate thread).

@esprit888, yes I did change out a couple of pistons (due to scratches and insufficient diameter), but the car was kind of having these intermittent "is it overheating, or is it not" symptoms before the tear down.

So next up is to test the fuel pressure, and then work through the injectors I guess (even though I got them cleaned and checked before the tear down)...

But the good news is that I did get rid of my HVL tapping. Little wins I guess.

[TorqueHorse]

A little bit of progress

Finally got the car to idle on its own tonight, yay Boy, is there still a lot to fix
Made the saddest of jury-rigs at the fuel feed line to actually get the fuel pressure gauge to connect. Yes the gauge was broken too, so that took hours of my time to troubleshoot and fix. Found out my fuel pump is a little weak (only 38psi at ignition ON and 40 when running) but the backflow checkvalve is shot and it loses pressure over time. Replacement on the way, but I wasn't convinced that was the source of my woes. FPR seemed to check out in spec though.
Finally was able to start pulling some data down from the elanscan (it is a loose USB connection from cable to computer) and after days of just looking at a series of about 30 second runs and trying to make sense of all the cool things I could see, I finally found an interesting correlation between the target idle speed and the IAC position. Every time it tried to drop target idle from ~1.2k rpm down to 950, it would jump right back and the IAC would spike right around the same time. So, before tearing the fuel system apart I followed the EMJ trouble shooting guide and disconnected the IAC and finally the engine could hold an idle on its own. I have a replacement one that I hope works, so that will be next.
Now that the engine could run on its own I was able to run it up to temp and have the thermo open. That's when I finally got a reading of combustion gasses from the block tester sniffing from the rad tank. And in addition to that, there is oil leaking somewhere around the oil cooler, and one of the pump bearings is going out btwn pwr steering, oil pump and aircon. Lets hope it is the aircon since that didn't work already anyway. So she squeals like a pig, smokes like a chimney, and bubbles like a bottle of soap, but now I finally have some solid leads to chase down this weekend in the light and with a somewhat stable patient.

[TorqueHorse]

Runs better without O2 sensor?

Uhh, ok feeling like I'm chasing my tail a bit here. I've been going through the list of things to check and the best I've gotten it to idle is when I disconnect the O2 sensor. It keeps surging otherwise, but even when disconnected, it still wanders a bit. Video of today's test

along with some elanscan's from today, as well as the past couple days doing other tests like having the IAC disconnected and then going cylinder by cylinder disconnecting injector and spark.

Rosie_Sept2021_elanscans.zip

(180.4 KiB) Downloaded 800 times

To recap:
I've replaced the ECU, coils, leads, and plugs;
I've replaced the IAC valve; the EGR valve isn't stuck; the EECS is closed; I've changed most vacuum hoses and the others check out; I disconnected the exhaust from pre-cat and didn't affect performance;
The fuel rail pressure and return are in spec and the injectors are all getting a signal, but I have not checked the individual injectors yet for leaks;
Some new sensors, and others seem to check out (O2 sensor is brand new BOSCH 15701);
I set the base timing as best I could, but it wouldn't hold a constant idle (it is better after adjusting the CAS up a bit, but still not perfect); Throttle bypass is set as low as it will go;

FWIW The smoking, squealing, and bubbling have subsided for the most part.

I know there is coolant leaking in the engine somewhere, but I'm just trying to eliminate as many other causes as I can to be sure that is the only problem before tear down. Anyways, I'm new to reading elanscans and don't really know what a good one should look like, so if you see anything else, or have tips on what I should try next, I'm all ears. Otherwise I will be starting to take it all apart again. Thanks all!

[TorqueHorse]

Its been a long weight

@HJ2, I finally have the engine out again and I was able to weigh everything for you. About 425 lbs. (max).


Transmission ~80 lbs


Exhaust manifold & Turbo ~40 lbs


Hoses, brackets, and pipes, etc. ~45 lbs


Alternator, Starter, Fuel rail, jackshaft ~40 lbs


Engine Block, Head, Clutch/flywheel, Intake ~220 lbs

[TorqueHorse]

Process of elimination

I have the turbo off, so I was able to pressure check the coolant lines going through the turbo cartridge and it held 20 psi for 30 min. Even though it wasn't hot, I'm going to rule this one out for now and move on.




I was also able to track down where my oil cooler was leaking from. I thought the matrix had cracked, or maybe one of the washers didn't seal, but it turns out it was just the hoses weeping at the hose to metal connection. I had planned to do the hose swap out as part of the first oil change post rebuild, but this kind of forces my hand to get it done. (Out of frame, I was juggling my air compressor in one side and plugging the other with my thumb and somehow taking the photo too).

[FunElan]

I've followed most of your build, I really hope you don't have to tear it down again.

[TorqueHorse]

Thanks @FunElan, but with combustion gasses showing up in the coolant, it had to come apart.

This should be as blue as the stuff in the bottle.

Pulled the head apart and how quickly did it all get so dirty. I don't think I clocked even 2 hours of run time.


I couldn't see where it was coming in exactly on the block, unlike RonR's great recent post


But I could really smell it up by the water pump side. You can see the corrosion getting into the part for the head bolt.


So I took both to the machine shop and got them skimmed down a few thousandth's of an inch. It ended up the block was more out of true than the head.


There was still a tiny bit of pitting on the head and the block at that area, so I ended up putting a little bit of high temp JB weld in there and flatten down with a razor blade as it dried.


Waiting a bit to start on the rebuild until I have all the other little things done, so that I minimize the time from assembly to running with oil. I think I had oil starvation on my brand new bearing the last time I put it together since it was >3 months assembly to run-in before



AND I think I may have found the reason for my shite performance (fingers crossed). It ends up that some PO had put the wrong injectors in, and they only had a flow of 230, vs the ~320 that we are supposed to have. Won't know until I get it all back together again...again.

If only I had checked the specs the at the beginning of all this