troubleshooting help - at my wit's end (Mitsubishi 6G72)

So since it got cold at the beginning of the winter, I've been chasing a hard-start problem with my Dodge Raider (Mitsubishi 6G72 engine). The problem is DIRECTLY related to temperature outside - if the car is sitting overnight in 20 degrees, it takes like 10 attempts to start. If it's been driven and stop at a store, it starts right up when I come out. 40 degrees maybe 5 attempts, and so on. Once running, it runs, drives, and idles perfectly, as if there's no problem at all. If it's a warm day it starts in 1 or 2 cranks (and probably will start right up once it gets really warm out again)

Initially I thought it was a fuel issue (losing pressure when sitting) that just happened to coincide with winter, since it seemed as if it wasn't getting fuel immediately. But after replacing fuel pump, filter, injectors, and FPR.....no change at all. It gets good fuel pressure when cranking, no fuel smells, etc. 

Then chasing ignition, so replaced the coil, plugs, wires and cleaned out all engine-related electrical plugs, etc......no change at all. It gets good spark immediately when cranking. My starter is on a separate button and battery is almost-new, so cranks with good power. Ignition switch is OEM but seems to turn everything on fine. The only thing I didn't change is the igniter module, but again, it's getting fine spark when cranking and runs fine so don't think that could affect things. 

Then figured it must be the system that tells the ECU the coolant temp so it can adjust fuel trim (since this is a common issue with this engine), so first I swapped my spare ECU - no change. then swapped the ECU coolant sensor - no change. Then checked the wiring that goes directly between the two (good continuity on both the sender wire and the ground). 

So I think I've pretty much run out of things I can think of. I've had a throttle position sensor warning light for years (replaced it twice, no change, light will not go off) and that's never affected anything at all in terms of starting or driving (including last winter), and it's been on for years before this issue so I don't think that's related. 

So what the hell else could it be? No evidence of vacuum leaks (checked all hoses). I haven't checked compression yet but it had new rings like 20k miles ago and the engine certainly feels as strong as ever once running. Oil is clean. Coolant is clean. 

So what am I forgetting here?

Can we check what temperature the ECU thinks the coolant is?

Two random thoughts:

- Is the starter on a relay? If so, have you swapped it out for testing purposes?

- I'd try a smoke test to be sure you have no vacuum leaks; checking hoses may not be sufficient.

02Pilot said:

Two random thoughts:

- Is the starter on a relay? If so, have you swapped it out for testing purposes?

- I'd try a smoke test to be sure you have no vacuum leaks; checking hoses may not be sufficient.

Yes the starter relay was a prior problem and it actually has double relays now. It cranks hard that's definitely not the problem. 

I will smoke test at once it warms up but it doesn't seem like vacuum would be an issue since it idles perfectly and there are no other symptoms that usually come along with vac leaks. Also while I've been doing all this every single vacuum hose and other areas where that could leak has been disconnected and reconnected, and almost all the hoses are new within the last couple years. That was actually my initial thought when this first started happening so I checked all of them pretty closely. 

buzzboy said:

Can we check what temperature the ECU thinks the coolant is?

I don't have the slightest idea how I would do that on a 1980s ECU. 

But I would think if coolant temperature to the ECU was the issue it would have changed when I swapped out the ecu and the sensor. It's literally a direct circuit. 

Coolant temp sensor was my thought.  There are usually 2.  1 for the Guage and 1 for the ecu.  Are you sure the correct sensor was replaced?

Sorry for the silly question.

Does this engine have a cold start injector?

wvumtnbkr said:

Coolant temp sensor was my thought.  There are usually 2.  1 for the Guage and 1 for the ecu.  Are you sure the correct sensor was replaced?

Sorry for the silly question.

Yep I'm positive. The one for the gauge is a single wire, and unplugging it makes the gauge go dead so it's an easy one to verify LOL. There's also a third one for the air conditioning system but I don't see how that would affect anything.

MadScientistMatt said:

Does this engine have a cold start injector?

I don't believe so. I rebuilt this engine entirely and don't recall running into anything like that nor can I find any part number for one. 

In reply to irish44j (Forum Supporter) :

Could there be a hairline crack in the intake manifold? It would be temperature-sensitive, hard to locate, but definitely create a vacuum leak when cold.

02Pilot said:

In reply to irish44j (Forum Supporter) :

Could there be a hairline crack in the intake manifold? It would be temperature-sensitive, hard to locate, but definitely create a vacuum leak when cold.

if it's warm enough tomorrow I'll smoke the engine. so hopefully that will reveal (or rule out) that aspect. 

other things that made me think of is maybe the IACV isn't sufficiently closing or opening when cold? IDK. It was thoroughly cleaned (or new, I forget) 

I had an almost identical problem on a MK1 GTI once.  Hard start when cold, no problem when warm.  It ended up being injector seals.  When cold they would shrink and let air in. When warm they would expand and seal up and all was well. 

Is there a check valve anywhere in the fuel system that is supposed to maintain fuel pressure at the injectors when the engine and fuel pump are off?

If so, perhaps it's failing to maintain pressure for some reason when it gets really cold. That would explain why extended cranking is needed when everything is cold, but not once everything has warned up.

Sonic said:

I had an almost identical problem on a MK1 GTI once.  Hard start when cold, no problem when warm.  It ended up being injector seals.  When cold they would shrink and let air in. When warm they would expand and seal up and all was well. 

Yeah that was one of my original thoughts, some kind of injector leak or similar  but I just replaced the injectors and the seals so that's not it.

Is there an Intake Air Temperature sensor?

earlybroncoguy1 said:

Is there a check valve anywhere in the fuel system that is supposed to maintain fuel pressure at the injectors when the engine and fuel pump are off?

If so, perhaps it's failing to maintain pressure for some reason when it gets really cold. That would explain why extended cranking is needed when everything is cold, but not once everything has warned up.

no, I've found nothing to indicate one exists (and I replaced all the fuel lines a few years ago and didn't see one). loss of fuel pressure was my initial diagnosis (hence the new pump, injectors, and FPS) since it seemed to fit the symptoms. That's the most frustrating part - I was SURE that was the issue lol. 

To check that theory I actually put a momentary switch on the fuel line so I can manually prime the system up to pressure before cranking (assuming it was losing pressure). It makes ZERO difference even if I run the pump for 10 seconds to build pressure before cranking. none. So at this point I'm almost certain it's not a fuel pressure issue. 

1988RedT2 said:

Is there an Intake Air Temperature sensor?

Not a standalone one, and I haven't found any clarification about whether there is one integral in the AFM. The workshop manual wiring diagram shows this of the AFM. Obviously there's a resistor there, but not clear if it's measuring temperature as well as airflow. I assumed the coolant temperature sensor played that role but maybe I'll have to look more closely at the AFM. 

But would that fit with the "it's 20 degrees out but the engine is still warm" thing? The engine being warm wouldn't affect the AFM being "warm" as it's measuring the outside air temp. IDK. 

Yes, the resistor in the AFM has got to be measuring temperature. Flapper type AFMs need this measurement, and unlike hot wire types they can't read it indirectly.

So can we agree it is a lean mixture on cold start up?  After a few tries the next morning stop and sniff the tailpipe.  Not a quantitative test but it could verify spark vs fuel.  Since the fuel pressure was ruled out that leaves just not enough coming out of the injectors. If you can disable the ignition do that and the next morning crank it for 10 seconds and sniff the pipe again.  It should smell rich.  If not you need to learn the EFI system and see where the signal it needs is and look for it.  If you can find it and manually trick it by grounding or adding a resistor somewhere problem found.

So this morning first thing I did a smoke test and no evidence of any vac leaks anywhere. The whole intake system appears to be airtight.

Then pulled out the MAF (pretty sure it's original, as it has Mitsubishi logos on it). Nothing of particular note - the resistor is right up top and appears undamaged, solders are still good, etc. No evidence of any damage, etc. I sprayed it out with some MAF cleaner but didn't see anything indicating it was dirty or grimy or anything. The plug contacts looked ok but I cleaned them up a bit better. Was hoping I had a spare one from the parts Raider I used to have, but couldn't find it. You can get them for under $100 online (thanks global Pajero market and the fact that the same MAF is used on a bunch of Mitsu trucks). Even if that's not the issue, I may order one just to test and to have a trail spare since the one in there is probably 35 years old.....

After all that, I went to start it up (it was about 40 degrees at the time, not teens like last week) and.....it just started up after about 2 second of cranking, no problem (after sitting for about 18 hours overnight). So hmm. IDK if cleaning the MAF did something or if it's just the warmer temperatures. It's supposed to be 60 on Wednesday so may let the truck sit undriven until then and see how quick it starts cold in the afternoon when I get home from work. 

When it starts, does it act like it was flooded?

If so, that all aligns with a dirty throttle plate.

Pete. (l33t FS) said:

When it starts, does it act like it was flooded?

If so, that all aligns with a dirty throttle plate.

No, once it starts it high idles to warm up as usual, and then drives normally.

Throttle plate and entire TB was cleaned when I had the IM off (and wasn't all that dirty in the first place)

Had a similar issue once on a digifant vw with a bosch flapper maf. It would get a little stuck when cold. Good cleaning fixed it. 

Ill note though that this isn't a flapper valve MAF so far as I can tell. I know my Porsche 924S has a flapper-type flowmeter, but this one looks to have a hot wire/hot film-style setup; I didn't see anyting resembling a moving flapper. IDK...

In reply to irish44j (Forum Supporter) :

It's neither flapper or hot wire. That honeycomb is a vortex generator, it uses it to count the incoming airflow. I don't recall them being problematic. There is an air temp sensor and I believe a barometer in there too. If you think the sensor is bad, unplug it when it won't start. It should bypass the counts and start off a limp home table. 
 

I'm leaning heavily towards coolant temp sensor related. I know you checked the wiring and sensor, but it sounds just like what used to be a very common starting issue with turbo DSM's when the wires got brittle. 
 

Is the fuel pump running when it won't start? I know fuel pump relays were an issue sometimes. I know most Mitsubishis had test connector under the hood on the fire wall where you could run the fuel pump with a jumper off the battery.