Help solve my automotive mystery

[deraltekluge]

At this point I've concluded that it has something to do with putting fuel in the tank, and I have 4 theories:

1) My car doesn't like gas stations, and is trying to condition me not to go to them.
2) My car really likes gas stations, and wants to stay once it gets there.
3) Water in the tank.
4) Clogged fuel filter.

1 and 2 don't seem scientific enough for my engineering mind. 3 and 4 have some holes as well since it runs without problem with just a little pressure on the gas pedal.

Have you tried stopping at places other than gas stations? Try it, stopping for the same length of time it takes to get gas, and see what happens.

[Tom Stormcrowe]

The indication they have water in their tanks at the bottom is the fact that I got a tank load of water comntaminated gasoline from them recently and it was right after their delivery. (The stir effect).

[gitarzan]

Have you taken it to a car parts chain store? Most will pop on an analyzer for free. It can eliminiate a lot of issues, all kinds of sensors...

[jccaclimber]

Unfortunately I managed to throw something in my back way out of whack at the climbing gym tonight (had to have someone come get me so I could get home as I can't drive as a result). As such, it may be a bit longer than I had planned before I can lean in to the hood and probe some things with a DMM. The good news is that I can breathe without significant pain and cramping now, so all should be well in a day or two.

If I remember correctly (it's been a while), you can access the fuel pump from under the back seat. Pull the seat bottom out, and there should be a removable plate. You can then take the fuel pump out, look at the pickup screen, and see if there's any crap in the bottom of the tank.
Most Japanese cars are this way, and since the Probe is a Mazda.....

That would be really great. I was concerned that I'd have to remove the hard lines or fab an access panel like the poor souls that do 3rd gen Chevy F-bodies.

If this were a throttle-body injected engine, I'd say it could be gunk in the fuel occasionally clogging the injector. But I believe this is multi-port injected; one clogged injector would not completely kill the engine, and it seems unlikely you'd have multiple injectors clogged just enough to kill idle but not enough to affect off-idle performance.
Some kind of clogged pickup screen in the tank seems possible, but I'd think that would also affect performance when off-idle - which you don't report. It would be nice if you could measure fuel pressure when the event occurs.
A sticky IAC valve, or intermittent MAF sensor connection fit the intermittent failure, but don't explain the correlation to refueling. You have only two data points. With 15-20 I'd be more confident - how about running a longer experiment?
One issue I know you have is too many engineers and not enough technicians.

I actually have 3 data points. The two times that I filled the tank, and one time I pulled in up to a pump (3/4 tank at the time) only to realize that I forgot my wallet. On the other hand, it looks like I'll get 2 data points per week until I solve the issue.

Have you tried stopping at places other than gas stations? Try it, stopping for the same length of time it takes to get gas, and see what happens.

Yes, I've tried waiting a minute then restarting the car every time I get home or to work (car is completely warmed up by then as well) and have no issues. I have also stopped a couple other places and had no problems. Finally, yesterday I went so far as to open the gas cap and then wait the estimated fill up time before restarting the car. No problems what so ever, unless I happen to be parked at a gas station (which I'm sure was a fluke the third time).

Have you taken it to a car parts chain store? Most will pop on an analyzer for free. It can eliminate a lot of issues, all kinds of sensors...

It is an OBD1 car, OBD2 didn't start until 1996 for these. None of the parts stores or local shops I called have an OBD1 scanner. The good news about OBD1 is that unlike OBD2 you can get the codes out through the dashboard lights by putting a couple paper clips in the right places. No engine codes.

[jccaclimber]

The indication they have water in their tanks at the bottom is the fact that I got a tank load of water comntaminated gasoline from them recently and it was right after their delivery. (The stir effect).

Seems reasonable enough to me. Thanks for the heads up.

[deraltekluge]

No, that's very unlikely. The likely thing is that your car simply doesn't like to do hot starts. Something's wrong with the various sensors and controls that set the mixture and whatever. It may be a matter of adjustment, or it may require changing some components.

[Siu Blue Wind]

dirty fuel injectors?

fuel injectors.

multiple injectors clogged

[jccaclimber]

No, that's very unlikely. The likely thing is that your car simply doesn't like to do hot starts. Something's wrong with the various sensors and controls that set the mixture and whatever. It may be a matter of adjustment, or it may require changing some components.

I've done several successful hot starts since this problem occurred.

[jccaclimber]

You're really pushing that aren't you.
If I drain the tank and don't find anything odd, I'll go about pulling the four injectors for you. How can I test them to see if they are bad?

[PlatyPius]

You're really pushing that aren't you.
If I drain the tank and don't find anything odd, I'll go about pulling the four injectors for you. How can I test them to see if they are bad?

Remove the entire fuel rail and injectors. Turn the key on and off several times to pressurize the system. Look for drips.

The other option is to put a fuel pressure gauge on the fitting on the fuel rail, pressurize the system, turn the key off, and then wait. Watch the gauge to see if pressure drops.

[jccaclimber]

Remove the entire fuel rail and injectors. Turn the key on and off several times to pressurize the system. Look for drips.

The other option is to put a fuel pressure gauge on the fitting on the fuel rail, pressurize the system, turn the key off, and then wait. Watch the gauge to see if pressure drops.

I like the idea of seeing if they are dripping after taking them out, that makes sense. Also something I'll do when my back works again. I decided that it would be a good idea to go to work today, and that was a terrible idea.

[jccaclimber]

Progress!
With the help of a friend, and lots of ibuprofein (vitamin i), we managed to drain and drop the gas tank. We couldn't see in it all that well, but did discover that food coloring is completely soluble in water, and completely insoluble in gasoline. We threw in two drops of food coloring and swished it around a bunch. There are a few drops of water in the tank, but I would guess less than a teaspoon in total. The fuel pump pickup sock was nearly completely clean, and the pump starts every time (couple dozen starts to drain the tank as fuel pump test pin in the diagnostic port was oddly missing).

With that as a dead end I went ahead and removed the IACV (Idle Air Control Valve). It was coated in carbon, and nearly completely seized. I cleaned it out by wiping my finger through it, and using a toothbrush on the spring. It seems to be teflon coated, so carb cleaner would be a bad idea. That said, the screw in the end of it sets preload on the iac valve, it isn't really an idle screw if you will, and using it to change the idle will 1)Require a significant change and 2)Completely mess up the function of the IAC. However, with the tank re-installed and the IAC re-installed, the car seems to function fine, including when I filled the tank at the gas station. I'll post again in a couple days if the issue stays resolved.