Carrera 3.2 Rough Idle — Removing oil cap increases idle, solves issue?!

[Myami]

Hey Pelicans,

So a new issue has arisen and the symptoms are oh-so interesting.

My '84 Carrera has developed a rough idle. And when I say rough, I mean you can feel it in the car. I first noticed it after an oil change and brake fluid flush. When idling, the car sits around 800/850 and the needle wiggles a little bit. It doesn't hunt, but it's definitely moving. And there's a strange "ticking" sound from underneath the engine. I suspect this is the cat internals (or something else) rattling, but that remains unidentified.

Today, I tried the old "remove the oil cap" trick to test for vacuum leaks. Oddly enough, the idle increased to 950-ish and the engine smoothed out within seconds. The strange ticking went away as well. Impressed and confused, I shut the car down. When I fired it back up an hour later, the idle remained at 950-ish and the engine's newfound good behavior continued.

Any thoughts on what's going on here? I think this tells me that I'm running rich (hence the lean condition solution) and that I don't have a vacuum leak (or I do and the previous owner altered the mixture to put a band-aid on the problem). But beyond that, I'm puzzled. Any thoughts or tips from the experts as to the cause/solution for the problem would be much appreciated.

Finally, on an unrelated--or maybe, related--note: My hand brake light (the one on the left/driver side) has been illuminated since the flush. I have not disconnected the battery yet. Figured I would share, in case this is a actually a clue.

Thanks in advance!

[porschedavid]

I had a similar thing on my 1984 3.2 but it was also hunting a bit.i changed all vacuum rubbers hoses and runs normal now

[cmcfaul]

Assume you did not over fill the oil when changing. Is there air blowing out of the oil tank when the cap is removed?

Chris
87 Carrera

[Myami]

Originally Posted by cmcfaul

Assume you did not over fill the oil when changing. Is there air blowing out of the oil tank when the cap is removed?

Chris
87 Carrera

I was worried the shop did this as my oil level gauge shows that the car is rather ... full. Would this cause the idle issue I've described? And, if so, what's the easiest way to remove some oil??

[Lyle O]

You would know it was overfilled when you go up hill, around a sharp corner, etc.: you will start blowing blue smoke out your exhaust like a giant fog machine. It could cause idle issues if pronounced or prolonged. Best way to remove the oil is to use a hand pump and draw from the oil tank. You could also use the bottom drain plug of the oil tank, but it gets a bit messy...

[jlex]

The easiest way is to take it back to the shop and have those guys take out some of the oil. Make sure they know it's supposed to be checked with the engine warmed up and while idling. The level should ideally be halfway between the marks.
If you aren't changing your own oil, then you may not know how it comes out in a torrent when the plug is removed from the oil tank... I wouldn't go there. Plug removal from the engine itself may not provide as great a river of oil but could still get messy.
Let the oil change "experts" handle it.

[darrin]

Originally Posted by Myami

I was worried the shop did this as my oil level gauge shows that the car is rather ... full. Would this cause the idle issue I've described? And, if so, what's the easiest way to remove some oil??

I wouldn't trust the gauge for this -- mine generally stays pegged at the top when standing still after warmed up when my oil level is normal.

Instead, I'd check the oil at the dipstick with the engine running and warmed up. If OVER the top line, you could have an issue, if below the top line, likely not the source of your issue (though still slightly overfilled).

[Walt Fricke]

Darrin - the oil level gauge system is reasonably accurate for its purpose - to warn you before the oil pressure gauge or idiot light does that the oil pressure is low. However, if you pull the connector to the gauge off of the gauge, or the sender, the needle will peg.

The sender is a wire coil, with a sliding contact moved by a float on an arm. Eventually the sliding can wear through one of the surfaces of the coil. At low oil levels you may get a normal reading, but at higher ones the gauge may peg.

Which is another reason the usual advice is to check the dip stick frequently, and trust it more than the gauge.

[darrin]

Originally Posted by Walt Fricke

Darrin - the oil level gauge system is reasonably accurate for its purpose - to warn you before the oil pressure gauge or idiot light does that the oil pressure is low. However, if you pull the connector to the gauge off of the gauge, or the sender, the needle will peg.

The sender is a wire coil, with a sliding contact moved by a float on an arm. Eventually the sliding can wear through one of the surfaces of the coil. At low oil levels you may get a normal reading, but at higher ones the gauge may peg.

Which is another reason the usual advice is to check the dip stick frequently, and trust it more than the gauge.

Walt: COMPLETELY AGREE -- my comment was in response to the OP's observation that since his gauge needle was high (pegged?), his oil was overfilled -- sounds like you'd agree from the above that the dipstick (and not gauge) is the proper way to confirm an overfilled situation.

I also experienced an extreme example of a failed sender -- needle would flop around constantly (as if it drank a quintuple expresso) -- sender rheostat wiring was beyond repair, but replacement sender put all back to normal.

[cmcfaul]

if oil was changed by a non Porsche specific shop it is likely over filled. They typically fill it to spec when cold. When hot, it will be a couple qts overfilled. the oil then gets sucked into the breather / air filter / intake. Just makes everything oil soaked / dirty including the plugs.

Checking the oil level could not be easier. Just use the dip stick, with the engine running, hot and level.

Chris

[Myami]

Well, guys. I think we can count this issue as identified--see attached. The orange connector for the O2 sensor failed. So, now what???

(Note: While pretty handy, I don't have the time I used to have to "play with cars," as the wife puts it. Hence the oil change and "regular maintenance" type stuff being farmed out. I still get grief when the car is down for a day or two, but it's the path of least resistance.)

[steely]

Ugh I was gonna volunteer / suggest O2, but backed off when I saw the Oil related posts.

There is (or was) a gentleman here who used to fab and sell an O2 connector here that was compatible - you might try that path and I'll post too if I find the thread.

[Myami]

Originally Posted by steely

Ugh I was gonna volunteer / suggest O2, but backed off when I saw the Oil related posts.

There is (or was) a gentleman here who used to fab and sell an O2 connector here that was compatible - you might try that path and I'll post too if I find the thread.

Yep, thanks. Yet another O2 failure ...

I PMed Louis85 (the guy making the newly fabricated connector) and am awaiting his response. While a bummer, it’s much better than hunting for a vacuum leak. Leaks and shorts are my least favorite problems.

And, to close the loop on the overfill, thanks everyone for your helpful responses. It sent me off scratching my head and searching through the engine bay for readily visible issues. I appreciated the possibility of an overfill, but the work was done by a solid Porsche shop and I watched them warm up the car up before doing the second fill. Proper procedure, notwithstanding the pegged gauge. :-/

[steely]

Found it.

Check w/ Pelican Louie85, maybe he's still "in biz".

O2 Sensor connector


EDIT: I'm too slow LOL

[Walt Fricke]

On an SC, an O2 sensor failure typically results in more emissions and somewhat worse fuel mileage. The features of the CIS Lambda system which adjust the fuel mixture are also turned off (as I recall the matrix) at idle.

The 3.2 controls idle with that intake air bypass valve. You wouldn't necessarily need an 02 input to control idle speed, since the engine RPM signal isn't affected by the O2 senor.

Could someone with 3.2 ECU experience explain the connection?

I'm sure some operating parameters are affected by a failed O2 sensor (which is what a broken wire looks like to the ECU).

Just curious.

[cabmandone]

I'll be a bit surprised if that fixes the issue. I ran my car for a year without the 02 sensor hooked up before I had had a chip installed that did away with the 02 sensor and allowed me to install a wideband AFR gauge in its place.

[Quicksilver]

First a question: What year is your car? (Or has the DME been upgraded to the 28 pin ROM?)

The idle speed is controlled by the DME's programming. The earlier DME with the 24 pin chip has the idle set at 800 RPM. The later 28 pin DME has the idle set to 880 RPM.
When the idle position switch is engaged the DME uses the Idle Control Valve to adjust the idle speed. There are threads on how to test and clean the ICV if that is the issue. The "barn door" AFM is ignored with the idle position switch triggered.

The most likely culprit is a vacuum leak. The fact that the idle goes up with the oil cap off would lead me to believe that the addition of the air from the oil system is exceeding the quantity of air needed to keep the idle down where it is supposed to be.
- The first thing on the list is the intake manifold gaskets. They are notorious for cracking and the plastic spacers erode too. You can quickly check and see if any of the intake bolts are loose (18ft-lbs). If any are loose you can be pretty certain that the gaskets will have cracked. If you have an IR thermometer you can test how fast each of the cylinders warms up. Start with the engine dead cold up on jackstands. Fire the engine up and leave it at idle. Use the IR temp gun to see how fast each of the exhaust ports heat up. If you have cylinders that heat noticeably slower then you have a vacuum leak (or a really odd injector issue). If the intake gaskets are the issue you will need 12 gaskets and 6 intake spacers.
- At the car's age all the rubber in the engine compartment is suspect. There are some larger vacuum lines on the back side of the engine that can crack and they are hard to see. A cheap USB endoscope used with a smart phone or a laptop may help in locating a possible issue. (And if the fuel lines to the injector rails haven't been replaced with newer lines make sure you do it before a Car-B-Que.)

Years ago I had an episode that I call "The Rubber Explosion Incident". I did the intake manifold gaskets and discovered cracks in some vacuum lines behind the engine. During reassembly I had trouble reconnecting the large "plastic" vent line that connects the top of the engine to the oil tank. The hose finally fractured when I forced it. When the replacement showed up I was surprised to find that the hose was actually very soft rubber. (And I later I discovered there is an air restrictor that needs to be taken from the old line to the new line so opening the oil cap doesn't kill the engine.)
So, yeah, all the rubber was toast.
__________________________________
Other possible issues? A couple that come to mind are: An 'iffy' DME that needs the solder connections resoldered. Injectors that need cleaning.

BTW - Updating a DME to the 28 pin chip is a very worthy upgrade. The #1267355358 ROM gets you a smoother idle plus the same programming as the ClubSport chip (without the raised redline). When they put in the 28 pin socket at the same time they can reflow the solder which will increase the life of the DME and possibly solve problems.

[Shorty62]

Does anyone have a vacuum hose diagram for a 1988 3.2 I know i’ve seen them in the forum somewhere but my searches haven’t turned up anything. It’s like I need 2 universal joints and a couple of good mirrors to see what’s going on at the back of the engine.

[Shorty62]

Dr google just came through for me so disregard the above post - thanks

[Myami]

Originally Posted by Quicksilver

First a question: What year is your car? (Or has the DME been upgraded to the 28 pin ROM?)

The idle speed is controlled by the DME's programming. The earlier DME with the 24 pin chip has the idle set at 800 RPM. The later 28 pin DME has the idle set to 880 RPM.
When the idle position switch is engaged the DME uses the Idle Control Valve to adjust the idle speed. There are threads on how to test and clean the ICV if that is the issue. The "barn door" AFM is ignored with the idle position switch triggered.

The most likely culprit is a vacuum leak. The fact that the idle goes up with the oil cap off would lead me to believe that the addition of the air from the oil system is exceeding the quantity of air needed to keep the idle down where it is supposed to be.
- The first thing on the list is the intake manifold gaskets. They are notorious for cracking and the plastic spacers erode too. You can quickly check and see if any of the intake bolts are loose (18ft-lbs). If any are loose you can be pretty certain that the gaskets will have cracked. If you have an IR thermometer you can test how fast each of the cylinders warms up. Start with the engine dead cold up on jackstands. Fire the engine up and leave it at idle. Use the IR temp gun to see how fast each of the exhaust ports heat up. If you have cylinders that heat noticeably slower then you have a vacuum leak (or a really odd injector issue). If the intake gaskets are the issue you will need 12 gaskets and 6 intake spacers.
- At the car's age all the rubber in the engine compartment is suspect. There are some larger vacuum lines on the back side of the engine that can crack and they are hard to see. A cheap USB endoscope used with a smart phone or a laptop may help in locating a possible issue. (And if the fuel lines to the injector rails haven't been replaced with newer lines make sure you do it before a Car-B-Que.)

Years ago I had an episode that I call "The Rubber Explosion Incident". I did the intake manifold gaskets and discovered cracks in some vacuum lines behind the engine. During reassembly I had trouble reconnecting the large "plastic" vent line that connects the top of the engine to the oil tank. The hose finally fractured when I forced it. When the replacement showed up I was surprised to find that the hose was actually very soft rubber. (And I later I discovered there is an air restrictor that needs to be taken from the old line to the new line so opening the oil cap doesn't kill the engine.)
So, yeah, all the rubber was toast.
__________________________________
Other possible issues? A couple that come to mind are: An 'iffy' DME that needs the solder connections resoldered. Injectors that need cleaning.

BTW - Updating a DME to the 28 pin chip is a very worthy upgrade. The #1267355358 ROM gets you a smoother idle plus the same programming as the ClubSport chip (without the raised redline). When they put in the 28 pin socket at the same time they can reflow the solder which will increase the life of the DME and possibly solve problems.

Thanks for the detailed post. For completeness, my car is an 84 and I do not believe the chip has been upgraded to a 28-pin. I have a SW Chip in hand and plan to install it this weekend, so we'll know more in a few days.

As was predicted, repairing the O2 sensor did not solve the problem. While the car was in the shop, I had a smoke test performed and confirmed that there are no vacuum leaks. I also had a valve job performed and the engine runs beautifully when I'm on the throttle. Unfortunately, I still have an idle that can only be described as "stumbling".

So, based on the information in this post/thread and from what I've been able to find while searching, I'm left with the following potential causes: (i) fouled injectors (I've used a few bottles of injector cleaner to no avail); (ii) faulty ICV (I will search to find the testing / cleaning procedure); and/or (iii) faulty distributor (??--not sure if this could cause the stumbling idle, but my PPI revealed some wear inside the distributor).

Will report back when I know more. In the meantime, any thoughts/advice would be greatly appreciated.