Oil smells like gas

[pacem]

I did an oil change on 93 Caprice with 5.7L (120K miles) and for the first time on any car, I smelled gas in it.

What is it - problem with the PCV system and why would faulty PCV cause that?

It has a failing fuel pump, because the first time I start it during the day, it takes about 5 secs, and then each consequitive start is instant. I heard that a bad diaphragm could cause fuel to leak out back into the tank. But I don't see how that can cause oil to smell.

Or, leaking injector?

 

[Matt_S]

I think you nailed it with that last one: leaking injector. The fuel is bleeding off into the engine after shutdown. It's probably also leaking during operation, causing the engine to run rich and adding to the fuel problem.

There's a way to clean the fuel injectors by hooking the fuel rail up to an aerosol can of cleaner and running the engine off that cleaner. I know 3M makes a system like this, you can check with a dealer or mechanic. That might be in order here. Otherwise, a couple of treatments with a good, strong fuel system cleaner might help or the injector may just be shot and need replacement.

 

[Kestas]

If it is a leaky injector, you could run into a hydrostatic lock problem if a cylinder fills up significantly with fuel. This can bend a connecting rod.

If you suspect a leaky injector, I would recommend having all your injectors rebuilt. If you search the internet, there's a number of places where you can mail your injectors to have them tested, cleaned, and flow balanced for as low as $25 per injector.

 

[pacem]

I think I will replace the injectors altogether, they are wrong for this engine. It's a civie caprice which got a 350 from a "perfomance" caprice (9c1 package) and I kept the old 55# injectors, instead of using the bigger 68#. I wonder if that contributes to the problem.

In any case, I was told that smaller 55# injectors will cause this 9C1 engine run lean which is bad for it.

 

[Gary Allan]

I'm not sure of your injection system ..but given the mileage and the symptoms ..it may be just the computer compensating for several clogged injectors. The PCM, on most systems anyway, monitors O2 readings in open loop. When it detects lean conditions during too many events ..it broadens the pulse width until it doesn't see them any more. This is to compensate for a weakening fuel pump or clogging injectors. The problem is that it's "numb" ..it only reads the average, so the "flow imbalance" between injectors (which others have touched on here) can have you "dumping" in one or two injectors ..while being quite lean in others.

Pulling the plugs should give you an indication of which injectors are the offending ones. You've just got to determine if the offending injectors are actually the bad ones ..or the only good ones that you have left.

If revenues permit, I'd also recommend a flow matched set of replacements.

 

[tpi]

Does the '93 Caprice use throttle body or port fuel injection? Many of the early '90s full size V8 Chevrolets used TBI. Two fuel injectors.

 

[brianl703]

quote:

---

Originally posted by Gary Allan:
The problem is that it's "numb" ..it only reads the average, so the "flow imbalance" between injectors (which others have touched on here) can have you "dumping" in one or two injectors ..while being quite lean in others.

---

The Ford Fuel Injection book by Charles O. Probst (the one that covers 1987-1993 models and is considered the be THE definitive reference) claims that sequential electronic Ford fuel injection systems can adjust the air-fuel mixture on a cylinder-by-cylinder basis. (Obviously this is not possible with batch-fire systems).

It doesn't go into details, but I can think of at least one way to do it is to vary the pulse width of each injector separately, looking for a response from the O2 sensor. For example, if the computer is compensating for a lean condition, and has increased the fuel injector pulsewidth as a result, it can then start cutting back on the pulsewidth of each cylinder (separately) until it starts reading a lean condition again. Then it can make use of this data to apply a correction to just that cylinder's injector pulsewidth.

Does Ford actually do this (or something similar)? Is the book correct? I don't know--the best way to find out would be to get a Ford vehicle, and a set of injectors, one of which is slightly more clogged than the others.

Install them and drive it around for long enough that the computer can learn those corrections.

Then connect a dual-trace scope to the clogged injector's leads and one of the non-clogged injector's leads, and compare the pulsewidths. If the computer is indeed able to compensate for a clogged injector separately of the others, it should show an increased pulsewidth compared to the others.

I don't know if anyone has ever done this experiment. It would be an interesting one.

 

[pacem]

right, there are only 2 injectors in this TBI setup.

The 350 engine has 120K miles on it but everything else on Chevy Caprice has 175,000 miles. With that kind of mileage, I wonder what might be wrong with it.
There is some kind of fuel delivery problem for sure, because it takes 5 seconds to start the first time I start it on any day.
I am thinking fuel pump but that, and injectors would be good to replace.

 

[Gary Allan]

quote:

---

the one that covers 1987-1993 models

---

That was their famous MAF 5.0. Highly sought for mddifying. The very complex system tunes itself continuously. The system can handle fairly radical cams and any displacement up to 7.0 liters.

A worked 351 adapted to this MAF injection system would put many a SBC to shame ..and be totally streetable.

The typical speed-density system isn't nearly as "smart" ..but tends to be far more reliable (less stuff to go wrong)

 

[brianl703]

Well, for the Mustang 5.0 it wasn't MAF until 1989 (1988 California Mustang 5.0s are also MAF).

Now you may know from reading my other posts that I have a 1988 Mustang 5.0-- It is a 49-state model so it had speed density. Now it has MAF.

The conversion between speed density and MAF involved:

1)A MAF computer ($60 from wrecking yard)
2)A MAF sensor ($25 from wrecking yard)
3)A MAF harness plug and computer connector terminals ($30)
4)MAF bracket and intake duct/pipe ($30)

Installed all of the above, swapped two terminals in the computer connector, addded two terminals for the MAF input, disconnected and plugged the vacuum hose to the MAP sensor which became a BARO sensor instead, and enjoyed much improved throttle response.

The only potential reliability problem with MAF is sensor element wire contamination, and that can be easily fixed by cleaning the wire with some non-residue electrical contact cleaner or brake part cleaner. Older Ford MAFs make it easy by using slotted screws, not the tamperproof torx screws with expoxy-filled heads that they use on newer models. Oh what fun it is to get that crap out of the screw so it can be removed. (After I remove just enough expoxy to get the screw out, I burn the remaining expoxy off by laying the screw on a burner on the stove and turning it on).

 

[brianl703]

You can check for leaking fuel injectors with a fuel pressure gauge.

The system should hold pressure for at least 5 minutes after engine shut-off. If it does not, suspect a leak.

 

[pacem]

good idea.