Warm restart 3.2 takes 5 seconds unless gas is near empty?

[Hunt3R]

1986 Carrera 3.2
If my gas tank is near empty my car starts right up as soon as I turn the key, everytime, no matter how hot or cold the engine is... otherwise if I have gas then whenever the engine is warm it takes approximately 5 seconds of cranking to start the engine and the idle will then sometimes bounce a few times before settling down where it should, like the engine was flooded from the cranking perhaps.

The gas level threshold where the problem goes away is somewhere around when the low gas warning light starts to come on.

The hard warm start begins once the car has sat for 15+ minutes with a warm engine. Start it back up before the 15 minutes and it will start right up the instant I turn the key. Cold starts when the car has sat overnight or longer always start right up no matter how much gas is in the tank.

Youtube video of fuel pressure during hard warm start.

[Hunt3R]

Has anyone with a 3.2 come across ANY kind of starting problem related to the level of gas in the tank?...

I need clues from anywhere at this point! I have just tried so many things and nothing ever fixes it. I even replaced the fuel pump recently thinking that had to be it, and at first the problem was completely gone. I was running the car for weeks with very low levels of gas in it, only putting in 2-5 gallons at a time, then as soon as I fill up the tank the first time the hard warm start comes back again. Drive the tank down and once it gets near empty the problem goes away again and everytime I start the car it starts the instant I turn the key. I can just keep the level of gas low and the problem doesn't start happening until I fill up the tank again.

[burtm5]

when it won't start take off the gas cap. then see if will start then we can discuss. best Henry

[Hunt3R]

I've actually tried testing that a bunch of times but it never affects the starting problem. I've tried leaving the cap off for hours when I park it too.

I've also verified that the hose is not clogged running from the fuel expansion chamber thru the fender to the fuel evaporation tank in the front fender. There is no hose running from there back to the carbon canister as that hose was clogged somewhere so the fuel evap tank in the fender just vents to open air there, yet I've never smelled gas in my garage.

[cgarr]

What's the pressure do when you shut it down? I should hold for several minutes. Is that gauge on the fuel rail? Should hold min of 15# for 20 min or you got a leaky injector or bad check valve in the pump. I would guess injector if it doesn't hold maybe flooding the engine making it hard to start.

[rick-l]

Originally Posted by cgarr

Should hold min of 15# for 20 min or you got a leaky injector or bad check valve in the pump.

WAG

or the regulator bleeds off with a full tank and not an empty one

[Hunt3R]

Yeah the gauge is attached to the test port on the left fuel rail, I never remove it. FP goes up around 34lbs when I shut off the engine and holds for around an hour before it will start slowly dropping and get near 0 within a couple hours. I can't find any pattern to how it starts related to FP at that time so it doesn't appear to affect the issue.

The new fuel pump came with a new check valve too. I replaced the previous pump because I had the gas tank out recently to replace the belly pan due to a rust spot near the suspension arm and the guys doing the work said they saw some crud in-line in front of the fuel pump. I had hoped they would find the cause of the problem when doing this work but that was all they found. And it still had the same warm hard start when they put it all back together so I thought well maybe that crud in the line had been clogging\damaging the pump. But no, the new fuel pump and new check valve seem to work the same.

I had the whole fuel tank drained too during all this, twice, the 2nd time it looked clean. Also changed the fuel filter afterwards, it looked clean although when I changed the fuel filter a couple years ago when I first bought the car that one was pretty dirty.

[cgarr]

Well that's holding good pressure and the correct amount. Are you sure its a fuel problem? Poor spark? did you change the plugs? Just throwing some things out there?

[rick-l]

The gauge reads zero when he starts cranking it. Is it supposed to hold pressure?

[cgarr]

The relay under the seat runs the pump maybe a swap to check it. You can manually run the pump thru this plug too. The pressure jumps right up when he started it, maybe zero from hooking up the gauge?

The book only calls for three major things: System Pressure: Fuel Delivery Volume: Residual Pressure: Did you check volume yet?

[rick-l]

Originally Posted by Hunt3R

Yeah the gauge is attached to the test port on the left fuel rail, I never remove it.

Since the pump runs when he hits start I would say it is not the relay.

[Hunt3R]

In the video the FP was near 0psi because it had been sitting for a couple hours.

I have tried changing the fuel pump relay with my backup, and then back again but swapping them doesn't seem to affect the issue.

I have also tried jumpering the fuel pump relay socket to run the pump when I was expecting the hard start. I ran it for about 40 seconds and fuel pressure was holding at about 37psi this whole time. Then I plugged the relay back in and started the car and it still had to crank for 5 seconds.

I have not attempted the fuel delivery volume test, but it's a test for the fuel pump specifically eh? Which I did just replace with a new one...

I change the plugs once a year with Bosch copper, they are coming up due again this summer but changing them before didn't seem to affect the issue.

I believe that the issue is fuel related though because it's affected by the level of gas in the tank. Also, one time when I expected the hard start issue I took out a spark plug and kept it in the plug wire and grounded it to the engine ground and had someone start the engine while I watched the plug. It cranked for 5 seconds before starting and that entire time the plug had a bright blue spark that pulsed. I repeated the test twice later with the same results. I do believe that the pulsing spark I saw is expected, correct?

Thanks for all the ideas everyone.

[Hunt3R]

I should add that as I drive the gas tank down to where the problem goes away it's not an instant transition from cranking for 5 sec's to starting right up. As the gas starts to get low, on a warm restart it will choke a little right when I turn the key and then maybe crank for only a couple seconds before starting. Eventually as the gas gets near empty this turns into an instant startup as soon as I turn the key with no choking.

I'm testing how much gas I have to fill back up with (from empty) in order to trigger the hard warm start issue. I put in 7.0 gals the other day and so far every warm restart has either started right up instantly or choked slightly and then started, as I just described.

[javadog]

Have somebody start the car for you while you watch the fuel pressure. See if there is a difference in how quickly the pressure rises from zero, with both a full tank and near-empty one.

JR

[rick-l]

Good idea. See if the fuel lines are full of air with a full tank and not siphoned out with an empty one.

[Hunt3R]

Originally Posted by javadog

Have somebody start the car for you while you watch the fuel pressure. See if there is a difference in how quickly the pressure rises from zero, with both a full tank and near-empty one.

JR

Good idea, I took a video of the fuel pressure when starting the car with a near empty tank (gas light on) and it started right up:

1986 911 3.2 Fuel Pressure when starting normally - YouTube

Just like in my other video where it cranks for 5 seconds, the fuel pressure in both cases immediately jumps up to about 32psi as soon as it starts cranking. The difference is in this video with an empty gas tank it starts up right away while in the other video it just keeps cranking and the fuel pressure then slowly rises a little more before it finally starts up. Unfortunately, I don't believe that gives any kind of clue to the problem.

I also decided to remove the Steve Wong chip and put the stock chip back in the DME. The idle was then too low and needed to be raised a good bit so first I re-tuned the base A/F ratio to 14.2-14.4 and then I reset the base idle to 800 and I've been driving it for a week now and it did not affect the hard warm start issue at all, it still happens exactly the same way with the stock chip in the car, and it still goes away as I drive down to an empty tank. Was worth a try though just to rule it out.

[rick-l]

Just grasping at straws here but if you look at the videos a couple of times... with the empty tank the pressure jumps before the first compression stroke and the full tank video it builds more gradually and takes at least one full compresion stroke before it raches full pressure.

Now no idea what sould cause that.

[cgarr]

Here's one for you: a full tank is screwing with the fuel pump check valve and letting fuel drain back into the tank. An empty tank lets the check valve operate as normal and keeps fuel from draining back so it starts right up.

[rick-l]

Never mind. They don't look that different.

[rick-l]

Originally Posted by cgarr

Here's one for you: a full tank is screwing with the fuel pump check valve and letting fuel drain back into the tank. An empty tank lets the check valve operate as normal and keeps fuel from draining back so it starts right up.

I was speculating that a leak was letting air in when the return line was immersed in gas but that makes no sense either.

Does the charcoal canister let air through?