ep3_lol

Hey all. Recently, my 2.7 986 has had problems idling just after very cold starts (below freezing temperatures after I've left the car outside all night). Idling is fine in my driveway, but once I drive up it and stop at the stoplight 100 feet from my house, the idle gets very rough and will sometimes backfire multiple times through the intake. Then, at the next light less than an eighth of a mile away, it will idle fine. The car has 66k miles and I replaced the air filter with an STP unit about 1k miles ago. I thought the MAF sensor might just need a good cleaning, so I cleaned it per the Pelican Parts site's instructions. No change. I've read that an oiled air filter can mess the MAF sensor up, so I am a bit suspicious of the STP unit (I bought a bad oil filter cartridge from the brand but figured an air filter would be hard to mess up). Could this be causing a problem, or at 66k miles and 10 years old could the MAF sensor need to be completely replaced? Could it be some other problem? I have no issues at other RPM ranges or once the car warms up.

Macster

Backfires through the intake is a sign the engine is running lean.

This can arise from an intake air leak, a faulty air intake sensor (a function of the MAF), a faulty coolant temp sensor.

Intake air leaks: cracked oil filler tube; leaking oil filler tube cap; AOS. Or in putting on the aftermarket air filter you created a leak after the MAF. Double check your work.

For the leaking cap with the engine idling and the trunk open listen for an air leak past the cap. Wiggle/move the cap about and see if there is any play or you can create a leak. Or just buy a new cap.

A cracked filler tube is harder to diagnose.

A bad AOS can sometimes be confirmed by excessive vacuum at the oil filler tube cap in that it can be unscrewed by not removed against the pressure difference created by the failing AOS while the engine is idling. But this is not foolproof. Some times the AOS though going bad doesn't manifest this symptom.

Removing the throttle body and inspecting it for signs of oil and signs of oil at the hose connector directly after the TB is another sign of a faiing AOS.

A bad MAF (a bad air intake temp sensor) can sort of be diagnosed by disconnecting the MAF at the wiring harness and clearing the error codes using an OBD code reader. This resets the DME's adaptations to their defaults. Then drive the car normally and see if the symptoms remain. Afterwards reconnect the MAF, clear the codes again, and see if the symptoms return.

With the code reader you can also view the intake air temp and coolant temp values at cold engine start and afterwards and note any disagreement between these values and ambient temperature either at cold start or shortly after when the symptoms appear.

Another possibility though less likely -- but not unknown -- is the aftermarket air filter mod has resulted in an upset to the air flow to/past the MAF and as a result the MAF is not able to correctly measure air mass.

ep3_lol

Air goes through the filter before it gets to the MAF sensor doesn't it? I definitely didn't alter any parts of the intake aside from the filter and its housing when I changed it. Also, what is AOS?

Macster

Yes when everything is working correctly all engine air goes through the filter and through/past the MAF.

An intake air leak though describes a condition in which this is not the case. Some air is pulled into the intake system after the MAF.

This air is unmetered air and results in a lean mixture. The MAF injects fuel based only on the air it measures.

The engine will run rough misfire, back fire through the intake until the DME can adapt/compensate. It does this rather well. If you want a demo -- in a healthy engine anyhow -- at idle remove the oil filler tube cap. This creates an intake air leak. The engine will react but recover pretty quickly as the DME adapts. Then when you put the cap back on the engine reacts and then is fine as the DME adapts back.

Now another critical aspect to the proper fueling of the engine is the air flow past/through the MAF wants to be linear. This results in more precise measurement of the amount of air the engine is using and more precise fueling.

In some cases where the stock air box is replaced with an aftermarket air box or possibly even just a different filter this can upset the air flow and cause the air flow to become non-linear, maybe not all the time but during certain rpm ranges as the intake air flow develops waves, and the MAF does not make an accurate measurement of the amount of air and the engine is not fueled properly.

Another aspect of this is sometimes an aftermarket air intake system is deemed a cold air intake system but actually pulls in heated air from the engine compartment which can cause the MAF fits.

Thankfully even with a number of aftermarket air intake filters/systems only a few bring with them non-laminar air flow.

For the moment I think it safe to assume that in the case of your car there's more likely something else going on and one something else has to be an air intake leak.

It doesn't have to be this though.

It could be temp sensor problem either the MAF is not measuring the intake air temp correctly or the coolant temp sensor is not reporting the coolant (engine) temp correctly.

If either of these is higher than actual the DME can lean the fueling and the engine will be too lean and misfire/backfire through the intake is a sign of this.

Now a more sinister example of an air intake leak arises from a failing AOS, air oil separator. This is a fancy crankcase vent valve assembly that arising from air flow through it removes oil vapor from this air before it is routed the intake manifold and then on into the engine where this air is burned.

With a properly functioning AOS the amount of air that gets routed to the engine is insignificant. But when the AOS starts to go bad it delivers a bigger load of unmetered air to the engine and the engine goes lean and misfires/backfires through the intake.

You realize of course that even if there is an air leak or the AOS is going bad, or whatever, it will not fix itself. A proper diagnosis has to be made and a suitable repair done. Someone's going to have to get his hands dirty.

ep3_lol

That is making sense, thanks for the post. Of course, the problem with diagnosis is that I need very specific conditions out of my control to recreate the scenario (don't know of any freezers big enough to fit the car in ) that triggers the symptom. Unless I can find a definite intake leak and be sure I got rid of it I'll have to wait for that, which typically only happens when I'm trying to get to work. In the meantime, does the AOS or MAF sensor have a typical time or mileage replacement interval?

Macster

Neither the MAF or the AOS have scheduled replacement intervals.

I replaced the MAF in my 02 Boxster at I forget what mileage now when the problem was due to a leaking oil tube filler cap. This replacement MAF lasted maybe 100K miles then failed. I replaced it with the original MAF and it is in the car yet.

The first AOS (not MAF!) lasted 80K miles. The 2nd lasted around 100K miles. the 3rd is going on 90K miles.

But other owners have faced shorter replacement intervals for these items.

While the MAF is easy to replace it can cost hundreds of dollars for a OE one and while some use "OEM" to save money and have good luck some do not. I choose to use whatever the factory sends to the dealer.

The AOS costs maybe $100 or so but is expensive to replace.

Directioneng

Hi Macster and EP3,

Starting a vehicle that has been out all night in freezing weather puts work on the ECU as it calculates the fuel requirements. As the vehicle warms up the ECU transitions into normal temperature operation. Is it possible that the problem is simply a transition glitch while the processor sorts the fuelling out?

Macster

No. Assmuming everything is working as it should, the ECU has the intake air temp, the coolant temp, and knows the amount of air being used by the engine. It has fuel maps that allow it to very quickly calculate/look up the proper fueling (how long to leave the injector open) and when to trigger the spark.

The ECU does this all the time. At cold start, idle, hot idle and up to redline and everywhere in between.

If the ECU can keep up with the engine running at redline it can certainly keep up when the engine is running at idle rpms, even if the engine is cold.

ep3_lol

The thing that's been strange is that it idles fine when it is down a hill behind my house, surrounded more or less by structures, but it only starts to have the problem when it gets up on the street in the open about 7-8 feet higher in elevation. Windows will not fog up or get frosty where it's parked, so there must be some atmospheric differences, but I'm not sure how this would make a difference.

Anyway, it hasn't been cold enough for it to happen again. I'm going to check for a leak this weekend.

Macster

Hard to imagine the car's temp would be any different a few feet higher or lower in just a small difference in distance. I mean there are micro climates and there are micro climates but one isn't going to have it arctic on the north side of his street and Sahara on the south side.

My WAG would be the slight difference in elevation is not a factor but possibly the car's levelness or lack thereof. If the rough idle occurs when the car is not level but does not occur when the car is level or doesn't occur when the car is tilted one direction vs. another I'm thinking some kind of fuel pick/supply problem.

ep3_lol

It's level both places.

Macster

It is a puzzle then.

All I have left, and it ain't much, is could a temp sensor fail in such a way that a few degrees below some threshold has it providing a false/wrong reading while a few degrees warmer has the sensor otherwise behaving ok?

That is going to be a tough one to verify

gfl

I've read all this (quite interesting, really) and can;t say i have an answer. But i do have a few comments:

1. MAFs typically last a LONG time unless they are fouled. They are sometimes fouled by an over-oiled replacement air filter, whcih you just added

2. Whenever you change something, and a seemingly unrelated things goes awry, i always suspect the thing that changed. I would consider cleaning the MAF and putting the known good Air Filter back in. Have you seen objective flows tests on air filters? The best was a GM paper element in the test i saw. K&N was in the middle. QED.

3. AOSs have no replacement interval, but do fail. Typically, however, they have telltake signs of failure, such as blowing smoke on startup or smoke bombs under some track conditions. Mine began to go at 40k

4. When you did the A/F, did you undo any of the wiring or tubes on the intake? When i replaced my AOS (nasty job, BTW) there was one connector i didn't snap completely together. The car ran, btu did odd things, especially on decel and return to idle. After an hours search and 4 seconds repair, all was well.

Oh, i'm impressed by anyone who has 80 + 100 + 90k on his car!

One IMS too?

Grant

ep3_lol

So I mentioned that I used an STP brand filter. Are you saying that those are indeed over-oiled? The box did not specify whether it was oiled or not.

This makes sense, but the cold temperature triggered this problem. It was the first time my car had been that cold in about a year, in which time I've changed a few other things as well, including a "snorkel delete".

Every start-up or sporadically?

Thanks for the response.

Macster

Missed the info about the aftermarket filter. I hate these things. They may have been ok years ago -- I used them on my motorcycle and modified car engines -- but no way I'd use one on for instance either one of my Porsches.

And now I'm prooperly warmed up, I hate cold air intakes, too. One might as well install bad plugs, cracked coils, and fill the tank with crappy gas.

As for the snorkle delete that again brings up the specter of non-laminar air flow past/through the MAF.

If you have eliminated all other reasonably possible causes for the engine's behavior and are now faced with expensive diagnostics time or expensive throwing parts at the symptom instead you might just revert back to the stock air intake system.

In this case you probably should clean the MAF just in case it got oiled from the air filter oil then using an OBD2 code reader clear the codes even if there are none to reset the adaptations to their defaults. These are probably wrong due to -- if my suspicions prove correct -- due to the MAF getting confused from the less than ideal air flow past/through it and the filter oil.