Hard warm starts...

[Cherokee]

Hello everyone I am new to the dsm forums but I hope I have found the right place for my question. I have a hard start problem that I have been trying to figure out for months. I have a 1996 Mitsubishi Eclipse RS with a 420A.

Heres the symptoms. First thing in the morning it starts up perfectly, no hesitation what so ever. After it warms up for a few driving for a bit or for a long time, if I shut it down for more than a minute it wont start back up right away, I kinda have to fight with it. If I crank it for 10 - 20 seconds it starts up, or if I prime the fuel pump a bunch of times by turning my ignition on and off a bunch of times. After it sits for a few hours it is fine again. I have replaced e the OEM fuel pump in fear that was my problem. Nope same problem still. I know my car isn't over heating I have replaced the radiator cap ans the thermostat (they were both a bit weak but not completely failing), and I have got it up to full operating temp and tested all around with a temp gun. I also replaced radiator after it popped a hole. The only thing I can find at all wrong with it is I do have some low intake vacuum, around 12 inches at idle. I have hooked up a smoke machine to a vacuum line until smoke was coming out my cold air intake, with no luck of finding a leak. I am an automotive student at Western Technical Collage in Wisconsin so I am decently knowledgeable when it comes to cars. Any help would be greatly appreciated.

 

[2GFWDbeast]

this may sound funny but this happened in my 97 gst and i replaced my gas cap problem solved. Did a little reading about it from other car forums besides dsm's. Basically what is happening is vapor lock. The gas cap being bad wasn't allowing vapors to escape as there are built up. Its a cheap fix. It worked for me.

 

[Cherokee]

Ok i will go get a gas cap on Monday. Thanks for your help. Hopefully it works.

 

[mrdavidbradley]

I have the same problem hopefully this will fix mine too.....

 

[Cherokee]

ok so a new gas cap is not the answer... still the same problem, and only getting worse. it it taking longer and longer to start my car. Any other ideas?

 

[TunaTalon]

Have you tested the ECT sensor (engine coolant temperature)? If it's bad or the coolant level is below the sensor you can get that problem.

Without a reading from the sensor the ECU thinks the engine is full cold and will make the mixture extra rich.

 

[Cherokee]

I flushed my coolant about a month ago. Unrelated problem was the old stock radiator popped a hole. And if it was trying to run rich why would it be so hard to start? Wouldn't the continuous cranking just flood the engine rather than finally starting? I'm not ruling out that it is bad but i just don't understand why running rich would cause major delayed starts. Also if it was my ECT sensor why would turning on and off my ignition about a dozen times make it start first try? i know i ask a lot of questions but i like to fully understand the issue.

 

[luv2rallye]

Sounds like the classic hot start vapor lock problem cars have. On DSM turbo models it is usually a faulty fuel pressure solenoid (FPS). Since you have the NT however, you don't have a FPS so I'm not sure what it is, perhaps the aspirator solenoid valve. PM Doug99RS who has a 99 NT for good ideas on this.

 

[Cherokee]

what is that? i have never herd of anything like that before.

 

[Doug99RS]

Hey Chase, thanks for the PM as I'm not on here a lot lately. Feel free to PM anytime and I'll go to the proper thread and help if I can.

The gas cap, in theory, shouldn't affect your ability to start or not. The fuel pump picks up liquid fuel from the bottom of the tank and then pressurizes it to send it to the rail and then injectors. On a 96 and newer, that's the end of the line for the fuel. On the 95 model NT they still ran a return line so fuel that once entered the engine bay could then go back to the tank. More on that in a bit. The gas cap is vented to allow ambient air in but should NOT allow vapors out as this would increase vehicle emissions (Vehicle emissions are from the entire car, not just content of the exhaust stream). If a cap fails, it can allow gas vapors out as well as allow air in but neither of these situations would cause a no-start related to fuel.

I've run in to one problem on a minivan where I was testing for evaporative emissions and failed to remove a piece of test equipment and the result was that the fuel gauge started bouncing around. The bottom of the fuel tank actually pulled upward a little away from the tank straps and when you removed the fuel cap, the tank went back to proper shape and the gauge reported accurate fuel levels. The point being... a cap that leaks vapor out causes a check engine light but the car still runs fine. A cap that doesn't allow air in would cause other symptoms but not a no start, from my experience and training/understanding of the system.

Not arguing there, just sharing what I know. I wish all no-starts were a gas cap as I would buy stock in Stant.

Back to the fuel return line point:
Return lines were a bit of a problem back in the day but the fix was to do away with it. The term VAPOR LOCK actually deals with liquid gas inside the fuel rail boiling/evaporating and thus the fuel injectors would spray no gas or just vapor. The reason this was happening was because as the fuel enters the engine bay it gets hot. The injectors spray what they need and then the balance of the fuel is returned back to the tank. Cold gas in the tank is warmed by this hot gas and then sent back to the rail to get even hotter. See the cycle? Eventually it would get to the point where the fuel would vaporize and you'd get a no-start on a hot-restart. It was never a problem cold because the fuel would cool back down in to a liquid. Vapor lock was a problem on systems with a return line but should be gone for returnless problems.

The only exception I've run in to there was on a 4.0L Cherokee on the middle two injectors. They came out with a recall to install a "debris shield" on the intake which happened to be right above the exhaust manifold (inline 6 cylinder). The two middle injectors now get heat soaked and the debris shield just retained the heat even more. I swapped the injectors around, wrapped the rail in heat-wrap by thermo-tec? and the debris installed during a national recall campaign must have just fell off.....

I'm babbling, I should get on here more often I guess.

Back to the patient before us.

Are there any codes? If not, does your check engine light actually work? If you don't have any codes then can you drive it to a parts store and have them pull the codes anyways? If you get there and they tell you they can't because it's a mitsubishi then tell them to try anyways.

Engine coolant temp sensor (ECT) is a good suspect if you had to guess but I really hate throwing parts at stuff.

The ECT is going to add more fuel when the engine is cold and pull it out when it's hot. If the sensor is providing the wrong information then obviously mixture won't be right and you'll have a long crank time or a no start all together. You can remove your sensor and check the resistance at ambient temperature and then check the resistance as you heat it up in a pot of water on a stove. I can give resistance readings if you actually pull it out and are in the process of checking it.

If you get too much fuel then the engine will flood causing a no-start but you will also wash down the cylinder walls. The age old saying "I have fuel, I have spark it should start." doesn't take in to account that you have so much fuel that the piston rings can no longer build compression to set off the air/fuel charge.

I would recommend a compression test to help rule this out but it's results ALONE probably won't tell us everything.

One thing you did mention that was WAY out of whack was engine vacuum. You should be seeing 18-19 inches of vacuum on a good engine. A reading of 16-17 could indicate timing off one tooth. Vacuum at 12 inches... I've honestly never seen that on a car at idle and if it ran smooth and only had 12" then I doubt it would accelerate worth a darn.

I had a bad relay for the fuel pump? on the firewall once with my RS (which met an untimely demise a few years ago for those that didn't know). I had a battery relocation kit with a 100 amp circuit breaker that would trip when the relay shorted somehow. It would spin and spin and never catch and eventually trip that circuit breaker. Since you don't have a circuit breaker... I'm wondering if you've still got a bad relay, just no way of knowing what circuitry is tripping out.

When it doesn't start then I'd pull off a plug wire and make sure it's getting spark. I'd also want to know the history of the spark plugs, what brand and what the current gap is set to. I'd also want to know the age and brand of the wires. If they're original or Mopar then they will have a date on them like 04-00 indicating they were manufactured in 2000, the fourth month.

When it's not starting also keep tabs on the tachometer. See if it bumps up and holds to maybe 200 rpm while the starter is spinning it over. RPM on the dash would indicate a functioning crank sensor. If you lose crank signal then you get no spark and thus no start.

Um.... I think that's enough for now It's alot and I'm sorry but no-starts with no codes really require a good bit of checking first to give you the right direction to go in. Please read carefully all the stuff and do NOT omit an answer to a question I've asked, it may be that bit of information that leads us down the right path.
Thanks,
Doug

 

[Cherokee]

Wow, I never even expected an answer like that. I am amazed. Don't even worry about babbling I love to soak up any info I can. I'll do my best at giving enough info.

I have no codes at the moment and haven't for some time. Also, no the check engine light doesn't work. I don't know if it’s just a faulty light bulb or the pcm. Only reason I would suspect it might be the pcm is it is hell trying to get a scan tool to work on it. Only tool I have found able to work is a Snap-On ethos working under global.

I don't think it would be the ECT sensor because my gauge is still working fine compared to my aftermarket gauge I installed that actually reads out in degrees Fahrenheit. Yet at the same time I don't rule it out because I haven't tested the resistance.

I have done compression tests before with normal readings all the way across, about 200 psi give or take 5 psi. Also I did this with a very warm engine. I also have done power balance tests on all cylinders with about a 200 rpm drop with each.

My vacuum problem is very mysterious. I installed a vacuum/boost gauge because of this and it was half the price of the actual tester that you need the hood open for. At my idle of about 800 rpm I have about 13 - 14 inches, when I am breaking to come to a stop my engine sits perfectly at 1000 rpm until I come to a full stop. During this time I have about 17 - 18 inches. Once I reach a full stop the rpms drop back down to around 800 rpm with 13 - 14 inches of vacuum. Idle is smooth as ever even at that vacuum. The only way I can force my manifold to create 19-20 is letting go of the throttle at 2500+ rpms and during the deceleration of the rpms it bumps up to around 24 inches and gradually drops to my "normal" range. The vacuum problem doesn't affect my acceleration at all. I still smoke any civic or v6 mustang that pulls up next to me and I keep up right on the v8's back bumper.

As far as the ignition components they are all brand new within the last 6 months. The battery, coil pack, plug wires, and spark plugs. The battery is an Everstart, the coil pack and plug wires are both Mopar, and the spark plugs are Bosch platinum 2. The spark plugs were pre-gapped so I didn't feel the need to check.

I don't have the car tonight due to the girlfriend needing it to get to work so I will check for the crank sensor in the morning. But now that I think about it, I don't remember ever seeing the tachometer indicating any rpm during cranking. But it could be it just slipping my mind.

I hope I have relayed enough info for you to graph out some sort of idea. If it’s just the fuel pump relay I will replace it either way tomorrow just to see if it makes a difference. I usually don't just throw parts into a car without knowing if it’s going to work but it’s just a relay.... no big deal.

p.s. As I said before don't worry about feeding me too much info. I am a young mechanic wanting to learn everything I can. It’s not a bad thing, trust me.

 

[Doug99RS]

The all OBDII cars have two different networks used to access information from the DLC. When you are using the Ethos you're going under a generic network that doesn't require information such as year, make, model, VIN or other information. It's more or less just a plug-in and the scan tool determines the communications protocol.

When you go under year/make/model/vin then you're using the manufacturer specific network pins in the DLC and it may require adapters, personality keys or a two-connector plug to access ALL the necessary pins. On my RS, there was a second connector behind the 16pin DLC that I had to use with the DRB and lie to the scanner and claim it to be a talon. Then I could access make/model specific information like airbag, trans etc modules.

I would initially suspect that you may be missing one of necessary pieces of hardware or even software in the Ethos to access it. I believe Snap-On is running 10.4 now which may or may not include additional vehicle applications to help you. Chances are though even with 9.4 or older it's not going to be the scan tool, rather a key or missing connector adapter.

Getting that check engine light to work should be hard. The bulb should always have power to it and the PCM grounds the circuit to cause the light to come on. With a multi-meter at the correct 40 pin connector for the PCM you should be able to ground the circuit and force the light to come on.

Get rid of the Bosch platinum and put in NGK or the OEM Champion. Bosch go in euro cars and that's it. No questions. NGK is an awesome plug that I'll put in any import and even some of the chrysler's switched over to them from the factory or were recommended plugs per tsb's addressing rough idles and other issues over the years in a number of their vehicle lines.

The stock ignition system will work better with plain-old copper plus or the NGK v-power plugs and you shouldn't run any platinum or multi-ground strap style plug. They simply don't work in a chrysler product unless they came in there from the factory (Chrysler minivans had platinums).

I would suspect a faulty gauge being your issue. If it idles and runs as good as you say then check for a pinched vacuum line, pluck the head of the gauge to see if the reading changes and verify that with the key off the gauge at least reads 0. I've run through a lot of vacuum gauges over the years and have yet to find one that can last more than a year or two. I think about the only one I trust right now (subject to change tomorrow) is the vacuum gauge on my hand-held mighty vac vacuum pump. It seems to report accurately where as some of my other gauges didn't flutter when I had a timing issue or displayed the wrong vacuum reading.

In your situation, check the readings with the Ethos and see what it says. Also, verify that your barometric pressure reading on the scanner matches your elevation. Barometric pressure readings are taken the split second the computer is powered up to crank the car. By the time the engine begins to rotate, the reading is already stored to the computer. Improper baro readings will affect air/fuel charge as well. Unfortunately, there is no comparison like engine vacuum or compression which should be 18ish and 180+ with no more than a 20 psi variance. Baro pressure is different all over the country and you'll have to find out what your reading should be for your specific region.

The engine always shows higher vacuum when you're decell'ing. On a good engine and a known good gauge you'd probably see an initial spike of maybe 20-21 inches, possibly higher. With the throttle plate shut the engine is starving for air and if you use the trans to decel as well then you're increasing vacuum above your average automatic trans decel so it's going to always show higher.

Percentage wise I'd say that 13-14 inches at idle probably would yield a 17-18 inch on decel reading which kind of leads me back to questioning the gauge and not the engine.

You can swap the relays around without having to buy anything. On the automatic model there's three on the firewall next to the brake reservoir and on the manual there's only two. I've got an "engine bay tutorial" floating around on here if you want to do a quick search using those words and my screen-name, I think it's archived by now.

When you do get the car back, you may want to just peak at the back side of the block, above the oil filter. That's your crank sensor and you'll want to make sure there's no oil in the connector and the wiring is not melted or coated with oil. That's not a diagnosis or anything, just a quick, free check to help see if it's wiring related to the crank sensor.

Also, with the Ethos you can graph the ECT readings, two or three pids at a time I think with the Ethos, 4 with the solus pro and then on up with the others.

I'd put in... ECT, TPS % and see if there's a pid for Cam/Crank sync and make sure you've got it the whole time. Sometimes the TPS will report about 5-10% open depending on the scan tool which isn't a massive deal. If I had a TPS code then I'd break out a meter and check the signal wire going back to the PCM and verify the voltage agrees with scan tool readings and then verify readings agree with spec at "closed position". Sometimes there's no getting rid of that 5-10%. And the 420A doesn't need base idle adjustments.

If the check engine light works when you ground it then the computer is faulty. Replacing it may fix all of your worries and if one circuit inside is faulty, chances are that others are on their way out as well and you may be chasing your tail with sensors and wiring when all of your problems are contained in the PCM, they're just intermittently showing up in other ways.

 

[Cherokee]

Well after I got the car back this morning i went down to O'rileys Auto Parts and purchased a new fuel pump relay.... IT WORKED! My car starts up every time now. There was some really old dielectric grease all over the place and looked like it was rusty in some spots. I cleaned the whole connector and replaced the relay and now my car starts every time now. The only thing that still bugs me is my CPS isn't reading any rpm during cranking but the second it fires it shows normal rpms. I don't know if it’s something I should worry about or just to leave alone. Also I purchased new NGK VPower plugs. I’m going to install them this afternoon. Just wondering what the reason for the NGK plugs rather than Platinum 2 plugs.

P.S. Thank you so much Doug, you are truly a dsm wiseman. I bow in your glory. Thanks again oh wise one.

 

[Doug99RS]

One, they're over-priced for what your car really needs and two most vehicles develop issues down the road much quicker than they would have if the correct plug were used.

Anyone that's been diagnosing cars for a while or is familiar with a particular brand of car for a while will tell you that when it comes to a spark plug, OEM or a select few other plugs work well for the life you should expect out of the plug.

I wouldn't worry about the lack of RPM while cranking. So long as the engine cranks don't even try to chase it. The instrument cluster is a module and it receives information from yet another module for a number of things that cause needle movement or bulb illumination. So while cranking for your application, it may not even get TOLD to move yet. It's more or less a cheap/quick check for certain makes/models.

I'm glad you've got the car running but just don't get your hopes up yet. Give it another couple of days driving and then report back. If it would have acted up ten times between now and then but doesn't, then it's fixed and you can pay me my diagnostic fee of 1,000,000,000,000 pesos. I'll give you my paypal addy once you report back.

 

[Cherokee]

. Will do. Thanks again.

 

[Cherokee]

It has been working for a couple days now. No problems... Thanks a lot again, you have no idea.

 

[Doug99RS]

Thanks for the update and glad it's fixed!