Check Engine Light Keeps Coming on and going off

That was silly. The shotgun approach is more expensive than just paying mechanic to fix it.

Just for grins, is the air filter clean? Vaccum leaks on the manifold between MAF and cylinder head? Does the light ever come on when you haven't added any crap to the gas? Perhaps the additives confuse the ECU.

AZ, Thanks for the response. The air filter was cleaned when I cleaned the MAF. Haven't looked for vacuum leaks but that is always a challenge unless obvious hose fell off or broke. Leak in the intake is hard to find but will look this weekend. Any advice on testing would be wonderful.

Yes, the light does come on without additives, in the past and as recently as this morning (came back on this morning driving to work. Filled up yesterday from almost empty tank (added 23 gal to 25 gal tank) so very little additive in the mix).

This has happened off and on over the past several years but is not more constant. The owner's manual says check gas cap so back a couple of years ago, tightening cap seamed to help as light would go out after 2nd startup and warm up or so. Bought new cap. Maybe time for another but never checked code. Anyone know what code a lose gas cap will set?

EVAP (loose gas cap) code would be in the P0440 range, but why are you chasing a loose gas cap when you have two rich mixture codes? (or vacuum leaks or dirty MAF for that matter)?

Look for causes that would allow (or mimmic) excess fuel into the engine, things like a leaking fuel pressure regulator, leaking injectors, false high O2 voltage.

The usual way is to get some carb cleaner (gumout for example) and spray around all the joints; if air is being sucked in, then the carb cleaner will get sucked in instead and the idle speed will change.

It is still possible that the replacement gas cap is bad. :-) Is there a charcoal canister? Loose vacuum lines to/from it?

Actually, this is a dead end; manifold leaks would turn the mixture lean. That's what I get for posting so late at night. aarcuda69062 has more of a clue than I; bottom line is: don't use the shotgun approach. Replacing the MAF already wasted more money than a mechanic would charge to diagnose!

You 'sure' sound like you are just describing a bad gas cap....

Mike

86/00 CJ7 Laredo, 33x9.5 BFG Muds, 'glass nose to tail in '00 88 Cherokee 235 BFG AT's Canadian Off Road Trips Photos: Non members can still view! Jan/06 (More Off Road album links at bottom of the view page)

aarcuda69062 wrote in news: snipped-for-privacy@newsclstr02.news.prodigy.com:

I thought all the mandated OBD-II P0xxx codes were supposed to mean the same thing for all manufacturers, but it doesn't seem like that actually happens.

Honda's code for a loose/missing gas cap is P0457.

Thanks for all of the inputs. Couple of clarifications;

- The CEL (check engine light) has occasionally come on and off for several years. I would tighten gas cap or fill tank and light would go out but never had code checked. Light was out when I had time to get checked. Did shotgun and invested in a new cap without verifying codes. Also, not sure if those actions really impacted the light.

- A few months ago, light was on more than off and got checked. I don't remember code numbers but had 2, "right bank rich" and l"eft bank rich" and that points to MAF from what I could tel from googling. Cleaned MAF. Light Light stayed out for a month or so.

- Cleaned air filter (K&N) and light came on and stayed on. Recleaned MAF and light out for month or so.

- Added injector cleaner and light came on and stayed on.

- Filled with gas, light out after 2nd startup and warmup.

- Next day (tank has maybe 50 miles on it) light came back on (that was this morning). Hope it will stay on as I drive home and get checked. May invest in code reader....

This is why I added the injector cleaner. Thought may resolve some of these potential problems (leaky injector, maybe regulator). How can I test regulator? Also, since both banks are rich, would O2 sensor be likely?

A link to good diagnostic would be great. I assume same codes will be set (rich both banks) so how do I systematically work through the possible causes? What steps would the mechanic do if I took it in?

The messages here have already been helpful and thanks for anything further. Bobby

The 1st step is to chuck the K&N in the trash. Get a decent paper filter back in there.

Every time the MIL is lit, the code is stored. The light can be off and you can still pull the code that caused it. Get the codes, write them down, proceed from there.

Recheck the codes and report back. If it's too rich, then it's not a vacuum leak (that causes a too lean). My guess is an O2 sensor. If you haven't replaced it yet, then you are well overdue for one.

innews: snipped-for-privacy@newsclstr02.news.prodigy.com:

Honda uses a specific number in the evap range to indicate a gross leak in the tank, which almost always (just happens to be) a loose gas cap. They have other codes which point to small leaks, leakage on the canister side, etc. I find Honda evap diagnosis to be pretty silver- bullet friendly, actually.

They are still playing by the SAE rules. Manufacturers can arrive at an end by many different means. The code ranges are specific but the individual numbers within a range can vary widely by definition based on the means that particular manufacturer used to test that particular monitor.

Toyota MDT in MO

The oil from K$N's will mess up a MAF sensor.

It happens. EVAP codes start at P0440 with the P0440 being a sort of catch all code and go all the way up to P0459. One manufacturer may choose to use a code, another may not.

Can't remember ever seeing a P0457 on a Chevy truck...

The trouble code is written to a non volatile memory and can be extracted even though the CEL has gone off. Even if the code is cleared, it can still be extracted via "history codes" with an OEM grade scan tool.

Just so you understand, a coated/contaminated MAF sensor typically -under reports- the airflow into the engine, if less than actual airflow is measured and reported to the PCM, the PCM will inject less fuel than what is actually required, this results in a lean condition as measured by the O2 sensors. I would never expect cleaning a MAF sensor to cure a rich code.

Again, the light needn't be on to extract the codes.

Some injector cleaners actually work, most are just snake oil.

Nothing beats actually seeing gasoline dripping out of it. This may show up as the inside of the intake manifold being washed clean in the area where the pressure regulator is positioned. You may see it as fuel pressure falling off after the engine is shut down but that is not a 100% fool proof test. Watching fuel trims on a scan tool will verify that the engine is being over fueled but that alone isn't a 'smoking gun' that points to a specific component. The fuel pressure regulator is inside the upper intake plenum on your engine, the upper plenum is not too difficult to remove.

Entirely possible but not likely that both sensors have failed in the exact same way at the same time. If it were one side rich and the other side lean, this would be pretty much a slam dunk (after a few simple readings).

I'd verify O2 sensor operation with my scan tool (can I drive them lean on decel?), check fuel trims, check MAP sensor calibration, 5 gas it, test drive it and do a WOT run and record MAF readings up to 5000 RPM and plug those numbers into a VE calculator to rule out a skewed MAF, then I'd bore scope the intake manifold and look for gasoline wash by the FP regulator, check that the EVAP system isn't the problem, pick up the phone and get parts prices.

Thanks for all of the information. I got codes pulled at Autozone and the only ones they saw were P0172 and P0175 (rich bank 1 and rich bank 2). I checked fuel pressure not long ago and key on, engine off was 63 psi. Spec according to Autozone is 60-62 so I think pump and regulator are ok. Comments appreciated. Idling the pressure was 53 psi I believe but will check both again tomorrow to see if they changed. Probably resort to taking to shop or finding good scan tool but cost is probably about the same. It sounds like that may be the best way to test MAF and O2 sensors.

What I read a dirty O2 sensor reads Rich? If that is the case, then dirty O2 isn't it either. Also, appears unlike both would read rich at the same time (this is the 2nd time I've seen these 2 codes). Along the same line, a leaky injector would hit one side only on this engine, wouldn't it? Again, I will test pressure but expect it to read very close to spec.

Checked fuel pressue and running is 55psi and key-on-motor-off is 63psi. Again, open to any advice from here.

Light went out today..