Running lean machine (3.2L) needs help…
 

I am stumped (!) trying to understand why a 3.2L (stock in every way) would run lean only @ relatively low RPMs (1500-2200) and @ idle… (??)

Checked fuel pressure; Checked and replaced cylinder head temperature sensor; eliminate all vacuum leaks; checked ignition wires, rotor and distributor cap etc; checked all injectors pulse width and cables; checked air flow meter; checked and adjusted to spec both cams timing (stock DME and exhaust system)

Any suggestions what might be the culprit (I know that “lean is mean”)

Thanks!

Leak?
 

How about a vacuum leak or a leaky gasket?

It seems no vacuum leak in the usual places (including intake manifold)

Please keep those letters coming…

How about checking the O2 sensor? Or replacing it. Do you know the history?

You could have an AFM problem, too (air flow meter). Do some searching and you'll find a wiper issue that would read the wrong amount of air coming into the engine. Use search term, AFM, barn door, wiper and I bet you'll find it.


Doug

How did you determine it was running lean?

Joe

These are great suggestions… the AFM was replaced with a working test unit (made no discernable difference) –still wants to be lean (and mean)

O2 sensor may need a second look but was responsive (good voltage fluctuations)

Might be worth trying to run open-loop and see if false air is entering from somewhere (crankcase?)

>How did you determine it was running lean?

Joe:

Exhaust analyzer CO level below 0.5% (hesitating while coasting…)

-no hand waving with these Pelicans!

It's a long shot, but did you check the eight position rotary switch in the DME? Position #3 gives a 3.9% lean setting. Also, the Altitude sensor should only be closed loop above 1,000 feet. If your below that altitude and it's closed, it will be telling the DME to lean out.

Check out the details on both at Steve Wong's (911Chips) web site:

http://www.911chips.com/fuelsys.html

Good Luck!

Have you (long shot), done a top end recently?

I recently did one, and when I tried to set my mixture after a couple hundred miles, it was ridiculously lean. I had to turn the screw almost all the way in to get .65 % CO. Before the rings have broken in, the engine is just not pulling the vacuum it will later.

After about 2500 miles, I checked it again and it was at 3.4 % CO! Should have just left the damned thing alone!

ianc

But isn't that what it should be after the cat with the 02 sensor connected?

My 86 3.2 barely registers any CO at idle and still runs great.


Joe

That would be normal. I set mine at .8% upstream of the cat with 02 disconnected. When I plug the 02 back in, it heads down to .4% or so. Downstream of the cat will be much less, so it depends on where and how you're measuring this exactly.

If it's hesitating or bucking while coasting, there may be an issue with the idle switch which shuts off fuel flow with closed throttle above about 1100 RPM.

ianc

Doug: Did not explore that avenue (yet) since the DME was not touched but will need to consider the altitude compensator (to cover all bases).

Joe:
The measurements were made open-loop and the number I recalled was the upper limit of the very low readings (I know it was lean when the engine RPM was <2000 range). At higher (and increasing) RPMs this problem disappeared altogether.

ianc:
Normally, the air that gets to the auxiliary air-valve (idle bypass) has been measured by the air-flow meter (AFM) which should be in the equation (data map) of the air/fuel ratio settings (for the given temperature)

OR

are you referring the throttle valve (switch), mounted directly on the throttle shaft –if so, worth exploring to be sure since it plays a role in adding some compensation to the pulse-time.

Do you know if it normally signals “closed throttle” during coasting (coasting cut-off by other name) …?

Yep, that's the one.

Yes, when throttle is closed, switch is engaged and this tells the DME to activate the idle stabiliser valve. If throttle = closed, and RPM > ~1100 (can vary), then fuel flow is cut off until RPM drops again. If there's some flakiness with this switch, then that may account for the jerkiness during coasting.

ianc

<I>
<b>
If it's hesitating or bucking while coasting, there may be an issue with the idle switch which shuts off fuel flow with closed throttle above about 1100 RPM. </b> </I>

I had the same problem after pullin an engine. Throttle linkage was bent at the tranny pivot. Threw the switch helter skelter at the throttle (adjusting screw for the 4-8 mm gap is a PITA)

GO SWITCH GO SWITCH ... RAH RAH RAH

good call Ianc!

Good idea for the problem. I have the same stuff when I'm coasting along at about 2000RPM. It just wants to hunt for a rpm and misses and gurggles. I'll look into this also.

I had a lot riding (idling) on this much overlooked microswitch potential problem.

Regrettably, it does not contribute to the persistent lean mixture at idle and low RPMs. A related problem area could be the harness going to the DME. A good way to eliminate the culprit is to short (or open circiut) the harness form the switch socket cable) to the DME and see what happens. In my particular case the idle RPM went well over 1200 and returned to ~880 when the switch (jumper was put back into the circuit) which indicates that part of the A/F mixture works fine.

I still think that false air (bypassing the AFM) can be a causative factor… any more suggestions from the Pelicans…?

After the cat is not where you should be taking your readings since the job of the cat is to clean up the exhaust and make you think it's lean when it's fine.

I would concentrate on the ignition and forget the mixture.

Joe

As Ian said, fuel flow is cut off to the injectors on deceleration which activates the idle microswitch, and does not come back on until 1100 rpm. If you are still registering CO or O2 sensor voltage above 0.1 volts while decelerating above 1100 rpms, your idle microswitch may not be activating.

Another common cause of lean low rpm issues is a worn air flow meter track. Most Carreras and 964s over 70k miles have worn AFM tracks, especially in the idle and low rpm areas where the engine operates most of the time. You can see this when the black plastic cover is removed, and you see almost white wear tracks on the black resistive surface where the wiper contacts have worn through. The worn spots decrease the signal output voltage to the DME which will make it run lean and hesitate. Reposition the wiper arm up or down the shaft, and carefully bend the copper wiper arm to track on a new surface while just applying enough pressure to maintain consistent contact. Then adjust idle base CO as per spec.

To check for air leaks, use engine starting fluid (ether) and spray around all suspected points while the engine idles. Disconnect your O2 sensor and idle control valve during this test, and if you idle shoots up, you have a leak. Common causes are sucked in intake manifold gaskets between the manifold and heads.

Joe,

Just so there is no confusion, the lean condition determination is NOT based on values at the exhaust pipe (past the Cat) but at the point of the O2 Sensor, which is in front of the Cat (!)

The reason I am harping about the -false air- is that blocking (grossly) the air filter box intake does very little to change the idle (… does this hint at anything familiar such as large vacuum leak...?)

That's very helpful, Steve!

To be sure, I will chase it down and report back…

I had the good fortune to locate a new Air Flow Meter from an ’94 964. The new AFM improved the low RPM response (off-throttle coasting) but my idle is still rough. I found and fixed bunches of vacuum leaks (false air)… (I decided to replace the intake gaskets -on order form the Host)

Many of you may know that it is quite possible to refurbish the AFM (redirect the rheostat sweep-arm to a new track). I recall seeing a write-up about this on the 911 List. Below is how it looks (on my now back-up unit) -note how deep are the old grooves (pitted ~ idle range -not seen) in the ceramic substrate. I repositioned the sweep to a lower (virgin) track, which makes this unit good for another 100K miles.


Tested the output voltage and was pleasantly surprised to see how (after all these years) linear (flap angle 0-110 deg) it actually is (!)


http://forums.pelicanparts.com/uploa...1160602889.jpg

Yep, doing this helped mine a lot too.

ianc

Below is the before picture of the sweep arc (flap angle), showing considerable pitting (first ~ 10deg) and can conceivably throw the ECU into oscillations hunting (look-up table) for the correct air-fuel mixture…


http://forums.pelicanparts.com/uploa...1160677045.jpg

More than one cause...
 

I finally got around to order a set of intake manifold gaskets. When elevated the left side manifold (no need to remove the injectors or the fuel line supply) I was delighted (imagine that) to discover that indeed one of the gaskets was not seated correctly. Obviously, there must have been and air leak here for some time now (PO spent a bundle in high-profile p-shops chasing the rough idle and poor low RPM response). Another ‘tell’ was that the cylinder temperature associated with that intake leak tended to run cooler…

The use of engine starting fluid (ether is my choice) idea is a good one (Thanks Steve W!) but it tends to give lots of ‘false positives’ if you are not careful to avoid the volatile vapors reaching the air filter box. This was the lower (cylinder surface) gasket which is harder to reach with starter fluid but it worked well in my hands and gave me reason to “believe”

http://forums.pelicanparts.com/uploa...1161725385.jpg

The take to the garage message is that persistence, aided by smart Pelicans, is at times a better solution than just taking it to the p-Pros.

After all that said…it better idle smooth

Cheers,
Jascha

Interesting: I would assume that due to a leaner mixture, that cyl would tend to run hotter rather than cooler.

Good find!

ianc

Yep, that's my problem too. My intake gasket was bad on # 4 and I replaced it. Now the car runs like a champ. Thanks to all for the help.

I agree, all things being equal, the combustion should have been hotter in that cylinder…

Maybe so, but my #4 cylinder was not firing at all. It had compression but no fire. If I would take off the spark plug wire, the engine would'nt change tune. So it was cooler.

A lean running cylinder will be hotter than a rich cylinder but the key word is running.

Just a slight difference is A/F mixture and you have little or no combustion, just a bunch of cool air and even cooler fuel passing through an overly lean cylinder. Gasoline has a very limited flammability range.

"The flammability range of gasoline is between 1.4 and 7.6%. If the ratio of gasoline to air is less than 1.4%, then the mixture is to thin to burn. The mixture cannot burn when it contains more than 7.6% gasoline because it is too rich to burn."

Cliff,

That is quite telling of what I found –an intake chamber (in question) that was still “wet”

(I ran a cold engine just long enough to put the p-car on ramps and inspect the CV boots)

When installing new intake gaskets, it is imperative that the intake manifold gasket surfaces are perfectly plane with one another or the gaskets will get sucked in again in less than 10k miles! This can happen if the intake nuts were retightened repeatedly in an attempt to cure the leaks. Take the manifolds and check the gasket surface against an absolutely flat surface, such as a granite slab, or a polished piece of marble or porcelain tile. If not flat, you can take it to a machinist to have them plane them, or you can attempt to sand them flat with sandpaper taped to a granite countertop. The phenolic spacers should also not be warped, and it is also important that they still have the concentric grooves molded on both sides to grip the gasket. If they are barely there, or scraped off when trying to remove the old gasket, replace them with new ones. I use a dark brown sticky Permatex aviation gasket sealer on all gasket sides for additional insurance. Be sure to torque the manifold nuts with a torque wrench to the exact specification of 16-18 ft-lbs (IIRC) not any more or less. Come back after driving it a few thousand miles, check and retorque again if necessary, because the gaskets will have shrunk and settled in.

Good advice from Steve. I should probably think about doing that one of these days... ;)

ianc

Steve,

I don't recall seeing such "groves" (in some circles that’s a bad sign) on any of the spacers that I removed in the past. Can you point to it (or describe it again) in the attached image...?

http://forums.pelicanparts.com/uploa...1161744061.jpg

Hard to see in your picture, but you can kind of see the concentric rings on the surface of the picture, going around the doughnut hole. They are probably milled in after molding. When you're holding them in your hand, it's very obvious. The grooves are very similar in size and spacing as the surface a strip of standard desktop staples. You can run your fingernail across the grooves making a zipping noise.