Causes of lean conditions

[ike84]

So Blue and I had a fun little sidebar over the past week regarding lean conditions (starting as a discussion specific to single bank leans but then expanding to lean conditions in general) and the diagnostic process. I figured it might be worth while to share with the group. I'm just gonna cut and paste the thread as it went along, putting the poster's name at the top of each message. Please don't take any of this as gospel, I just figured it might help out some folks. Please feel free to add to the discussion if you can!

Sent from my SM-G970U1 using Tapatalk

[ike84]

Ike

Hey Blue,

I have been fortunate enough to not have had this issue happen to me. What I know about them is limited to what I've read on the forums, but there are a ton of posts about Vac leaks!

p1128 and p1130 are the codes most often associated with vac leaks, 28 being bank 1 and 30 being bank 2. its misleading because these codes read out in the durametric as if there is an o2 sensor issue, but what (i think) is actually happening is that the sensors are either reaching their lean limit.

the bank 1 issues seem to be more common. this is because the portion of the vac accessories connected to that bank (via the plenum or resonance tube boots) are numerous and tend to develop problems. this includes the vac cannister, the sai controls, the resonance flapper acutator, all the hoses and check valves that run between then, and (I believe) an additional line that runs to the transmission on the tiptronics. the worst part about these is that they are all virtually inaccessible without removing the plenum +- resonance tube and sometimes the runners as well.

the only vac that runs directly off bank 2 to my knowledge are the brake booster line and the fuel pressure regulator valve. the brake booster line is very robust and i can only think of 1 thread where someone had an issue with it. ive never heard anyone having an issue with the fpr valve or its connections.

the other vac-dependent systems are the AOS and the EVAP system, both of which come into the plenum just after the TB. The AOS system is what it is, no need to rehash that. The EVAP system can cause problems but it is almost exclusively related to the stuff under the right front fender (mainly the check valve) and are pretty uncommon. The most common "EVAP" issue isnt actually EVAP related at all, and that is the fuel tank vent valve goes bad and causes issues with fueling and occasionally causes weird stuff on hot startups.

in a lot of instances, 1128 and 1130 come up together. sometimes its because of a massive vac leak. some posts that I can recall traced that issue back to the MAF and derbis in the airbox or honeycomb before the MAF, although I couldnt quite figure out why that triggered those codes (unless it was moreso that the DME didnt find the O2 sensor output to jive with what the MAF was reading so it faulted the O2 sensors instead of the MAF)

anyway, thats most of what ive learned in trying to cobble together the info over the past few years. if theres something specific about the vac systems that you want to know more about, I'm happy to give more info if I know it. it drives me nuts that porsche does such a ****************ty job of making this information public like audi and vw do.

I think this would be a really good thread to post to get the community's experience. Unfortunately though I think that a lot of the older guys who ran through these issues are no longer around very much.


Sent from my SM-G970U1 using Tapatalk

[ike84]

Blue -

It is the same here I have never had any issue with my Boxster so there's that.
I have probably read all the same posts here on the forum relating to vacuum leaks and other performance problems as you have.
On other cars like early classic's (Jaguar, Austin Healey, MG)
and early American cars that are running Carburetors.
I have seen multiple instances of single bank vacuum leaks I have also seen single or multiple cylinder vacuum leaks on cars running multiple carbs.
Almost every time that it has been a single bank vacuum leak the leak has been at the gasket between the manifold and head. Or in an intake runner.
On Multiple Carb. engines like a Jaguar that is running three carbs I have seen vacuum leaks on two adjacent cylinders.
Each Carb. feeds two cylinders. Those leaks have always been at the Carb. to intake manifold gasket.
So I would say everytime I have seen this type of vacuum leak it has always been very very close to the intake valve.
Or isolated in some way such as a leak in an individual intake runner or base of a Carb. on a multi Carb. setup.

One thing I think I see here on the forum is someone gets a code that indicates a lean condition.
They think Unmetered air Vacuum leak!
They also have misfires.
Vacuum leaks cause misfires right!!!
So they think vacuum leak.
But misfires can also cause the O2 sensors to report a lean condition.
Which looks like a vacuum leak.
Positive fuel trims confirms it.
They cannot find the cause of the misfires so they think single bank vacuum leak.
Now you have a situation where you have identical symptoms for two different types of problems.
Fertile grounds for a misdiagnosis.
So I think there are times when a lean condition is reported on a single bank. That is thought to be a vacuum leak when the lean condition is caused by the misfires.

Sent from my SM-G970U1 using Tapatalk

[ike84]

Ike -

You are definitely right about that. Sorry that I digressed from your original question earlier. I have very little experience with older style carbeurated motors but from a basic principle standpoint, to have a single bank lean condition you are either going to have to have unmetered air in that banks intake or lack of fuel isolated to that bank. In this respect the M96 is somewhat unique. These motors just don't have the problems of blown intake seals or head gaskets (thank god). But they do have a screwy ass overly complicated vac layout (i said to STL recently that biggest beef with Porsche is that while I see beauty in simplicity, Porsche sees beauty in complexity lol)!

When the 1128 codes have come up in the past as isolated codes I think the majority of cases were tracked back to the vac-assisted components I was describing earlier. I think that is purely due to design - those components are normally routed back to the plenum boot, which is direct line of flow to the intake valves, so bank 1's pull is going to be largely diverted through those components. Things like hairline cracks in the vac reservoir or leaky check valves are nightmares - being inaccessible while the motor is in running condition they would be a nightmare to diagnose and fix.

I do think though that the overwhelming number of lean conditions is going to be due to unmetered air and not a lack of fuel. This is for 3 reasons - 1, the fuel pumps are more than up to the demand of our cars (they are the same for a 2.7 and 3.2, and although it is a different model number the t fuel delivery volume is the same for the 996 3.4) so even if they are running on the low side of pressure there is still more than enough oomph. 2nd, the dme can adjust a significant amount without throwing codes (I dont know the exact limits but running a 996 tune on a 2.7 motor my LTFTs are at -17% with no codes) so you would need a massive loss of fuel to cause this imbalance. third, the rails are in series, so you would never have low pressure on bank 1 and good pressure on bank 2. also, as a side note, every time I have read about a faulty injector (mine included) they have always been stuck open, not shut. an injector that won't open may be the exception here and create enough of an imbalance as to create a single bank lean but I haven't heard of this happening before.


so, with all this being said, here is the daignostic alogrithm i would likely recommend for this situation.

1 - vac test.

if vac test is abnormal, find the leak.

if vac test is normal, move on.

2 - fuel pressure test

if fuel pressure is normal, move on.

if fuel pressure is low, test the pump output and act accordingly.

3 - sensor replacements

this one is a bit tricky. theres no way that I know of to test/verify that the MAF and O2 sensors are working properly (except for voltage readouts which gets complicated).

but, pre and post cat sensors can be swapped. if its a bad sensor then the codes should change. if its only a single bank, sensors can be swapped from one bank to the other. Maf can be cleaned.

4 - if all else fails, investigate injectors

what do you think?

Sent from my SM-G970U1 using Tapatalk

[ike84]

Blue -

Along with unmetered air and low fuel as causes for single bank lean conditions. You have to add misfires. They can cause a lean signal because the oxygen is not being used up in the combustion process. So the O2 sensor sees excess oxygen in the exhaust. which is in a misfire case, unburned fuel. But the O2 sensor only sees the oxygen. So the ECU sees a lean condition. The other thing is valve train issues. anything from valve timing to valves not sealing at the seat. Or weak springs, or bad lifters. On an engine with four cams like ours. I believe the cams have a few degrees of adjustability by loosening sprocket on the front of the cam turning the cam and retightening the hold down bolts.
So you could have just one cam on one bank out of time and cause a lean condition.
So four ways that I see a lean condition could happen on a single bank the M96.
For what ever reason ECU's don't always report both banks as having a lean condition even when there is a common vacuum leak.
Sometimes they report just one bank. like the P1128 and no P1130.
I have seen it numerous times on other makes.
Now why they report both banks one time and only a single bank at others. Is beyond me.

I agree the most common cause of those two codes is kinda split between MAF issues and common vacuum leaks. Then exhaust leaks.
Yep the fuel pumps are pretty robust and as long as good clean fuel is used and maintenance is done. There are usually no problems with restricted fuel filters. So fuel delivery issues are down the list when I am trying to diagnose a lean condition. That is true on just about any make of car for me.
Good point on the LTFT I think the limit is something around 20%-25%. All your points on the fuel system are spot on.

As far as diagnostics of lean codes:
First thing I like to do If possible is:
Cold start and just let it idle and warm up.
This way I can listen to it,observe RPM,and sounds weird but feel it.
Then a road test.
If it is a suspected vacuum leak then:
let the engine idle so it steady's out if it will.
Then slowly bring the RPM up to about 2500-3500.
Here is my reason for doing that:
If it is a common vacuum leak:
Lets say a 1/4" hole in the intake.
and at idle let say your closed throttle is equal to the 1/4'' vacuum leak.
you have a 50-50 or a 1 to 1 ratio of metered air to umetered air.
As you open the throttle and bring the RPM up that ratio changes.
If you have a 3" throttle body at WOT the ratio is now 12 to 1 in favor of metered air.
So when I bring the RPM up slowly to 2500-3500 the engine smooths out and runs better and better because the Total amount of air entering the intake is being metered more and more accurately as the throttle opens more and more.
If the engine will not smooth out. It is probably not a common vacuum leak.
Sometimes I do the above with a scanner that can show fuel trims.
With a common vacuum leak or a single bank vacuum leak fuel trims will be positive.
I run the RPM up to 3000RPM and hold it as steady as possible for 60-90 seconds.
If it is a vacuum leak fuel trims will start to come down to more normal readings. (All other things being in correct condition).
I hope you can make sense of the above.
Those two tests are very simple and together they are very effective in telling you if the issue is a vacuum leak.
If those two tests are done correctly and you don't get the described results. It's 99% not a common vacuum leak.
If I determine it is a common vacuum leak find it fix it.
If I still have a lean condition:
Then the Vacuum gauge test to see the overall condition of the engine. Valve train,rings, everything I can get from two or three tests with the vacuum gauge.
If I have not found proof of a lean condition at this point.
Then I will look at sensors to see if they are faulty.
Then the fuel system.


Sent from my SM-G970U1 using Tapatalk

[ike84]

Ike -

Ahhh, misfires, I knew I was forgetting something!

I think of misfires very differently in these engines. If you break down the causes of misfires (not preignition or detonation, that is a huge pet peeve when people use those 3 terms interchangably, which you do not, and for that I thank you ) you are either going to have one of two situations - the person tinkered with something and didn't put it back together OR a catastrophic event has happened and your engine is now likely toast.

Here's what I mean by that.

Bonehead changes spark plugs, doesn't reconnect coil harnesses properly, and now there are misfires. Or removed injectors to "clean them" and damages them in the process or didn't hook them back up properly. I had a mentor who used to say "whenever a problem arises, always looks to where the hands of man have been". I think that is true for a lot of issues that get brought up on the forum and is one of the hardest things for us to help with for obvious reasons.

If it's not that, it's a catastrophic event. The timing chains on these motors don't stretch, so we don't gradually lose timing the way high mileage engines of the days of ol would. However, one systematic issue I'm aware of with timing is the chain tensioners. When those pads wear down, the dme is able to monitor the deviation and throws a code if it gets out of spec (which I think is 9deg). When the pads completely go, the chains jump the rails. Now the whole valve train is free floating and with this being an interference engine, you're now in the market for a new one.

The one exception I can think of was the guy recently who had a bad lifter spring after filling up with ****************ty gas. That was incredible luck that his engine survived unscathed there and he is the only person that I can rember seeing post something along those lines. Otherwise, I can't think of anyone else who has had valve train issues over the past few years.

The other notable exception here is what Stl is experiencing, which looks for all the world to me like a short in his wiring harness. Damn rodents. Again though, probably 1 in a million luck.

My point of all this is that I think of misfires in its own category, usually heralded by other leading events, active symptoms, and (usually) additional codes.

For your testing scheme, I think you are spot on. There ain't nothin wrong with feeling and smelling your engine. It's amazing how much information you get from your senses if you let them work!

I think your testing for a vac leak is brilliant. Reason and method must always accompany the madness. I would only point out though that the car must be in closed loop operations for it to be valid. A lot of us understand that principle, but there are some that definitely wouldn't. Not their fault, it's a complicated system, but its another factor that makes it difficult to help sometimes.

Sent from my SM-G970U1 using Tapatalk

[ike84]

Blue -

Spot on Spot on.
Yes misfire, backfire , detonation, knock, pre ignition,
All misused and abused terms.

One thing I forgot in my Diagnostic process of a lean condition is>>>exhaust leaks before the O2 sensor and at the bung weld. Amazing what a pinhole leak you can't even see does to the O2 sensor when it is at the bung weld.

Yes the engine must be fully warmed up for the Vac leak testing I described. That's why the idle, look, listen, feel, smell period followed by the test drive.
Another thing I like to do if time allows is:
have good strong flashlight, and a cold drink, (insert favorite alcohol here). pop the hood and just look, then look some more. don't touch anything, then look some more.
Look and don't tough anything until the drink is gone.
Amazing how many times a person has brought there car me and I found the problem that way.

You mentioned difficulty in testing MAF Sensors.

So a very important concept to try to use in diagnostics:
Is how things correlate to each other. Things that work in tandem and things that work in opposites or ratios.
The example I gave you of the ratio of metered air to unmetered changing as you open the throttle is an example of opposites or ratios of opposites. So you use that to diagnose Vac leaks.

So if you look at the MAF sensor and the TPS (throttle position sensor) there is a tandem or a parallel correlation between the two.
That correlation is voltage as it relates to Air flow.

The more air past the MAF the higher the MAF signal voltage. The more you step on the throttle the more air you get and the higher the TPS voltage goes.
There is a direct correlation between the two.
So the voltage signals will parallel each other on a graph.

So here is how I test a MAF sensor.
With engine off and at ambient temp hook up a scanner that has graphing capabilities (The Durametric)
look at Intake Air temp with engine off. Is it within 3-5 of ambient temp.
If yes next step.
Intake air temp is measured by the MAF.

Now get the engine full warmed up.
Find TPS voltage. put it on the graph.
Now find MAF sensor voltage. (not air flow not resistance).
MAF voltage.

Now when you step on the throttle you will see both voltage signals rise on the graph parallel to each other almost like a set of railroad tracks.
Let off the throttle the signals will drop again parallel to each other.
If there is a problem with the MAF you will most times see one of three things. The signal is biased high, The signal is biased low, or there is a stutter it doesn't run parrallel to the TPS signal.
Two other things to be aware of: there is a little lag time in signal response from the MAF when you go WOT... Air flow and RPM's increase slower than your foot moves.
The same thing happens when you come off throttle quickly to closed throttle decelerate.

So hook up your Durametric and play with it. once you get used to what you see you will understand how the test works.

When I do it I get everything set up and go for a 50 mile drive (loop). With the laptop on the passenger seat so I can watch it.
This Lets me see the two signals under all driving conditions.
If there is any kind of anything not right with the MAF I will see it at the same that I feel the car not respond correctly.

It is not an absolutely fool proof test. But it is pretty damn good

Sent from my SM-G970U1 using Tapatalk

[ike84]

Ike -

how does a pinhole at the bung weld influence the 02 sensor reading? that is something ive not encountered before. since the exhaust is positive pressure, i would think that the exhaust would not intermix with atmospheric pressure and instead you would just have a bit more exhaust gas pressure moving past them senosr.

for the maf sensor testing, i have a few questions. you said that the voltages should parallel each other in change. but what is the differenece between them? MAF>TPS or TPS>MAF? Also, the the absolute value difference between the two of importance? you mentioned biasing so I'm assuming that means that the gap is either too large or too small. I've never seen OEM descriptions or specs of these voltage readouts so I guess it may just be from experience? lastly, does this hold true on other vehicle makes as well that use this type of intake setup (the parallel between maf and TPS sensor voltage)?

btw, this is a good discussion! youre a smart guy blue and I appreciate the back and forth and your experience/insight with all this.

Sent from my SM-G970U1 using Tapatalk

[ike84]

Blue -

How does a Pin Hole at the bung weld influence O2 sensor readings?
First think about "your own" experiences with exhaust leaks. Large ones!!
They make a sound (a pulsing sound) so that positive pressure you refer to is not constant. It pulses. (A light should have just turned on in your head)
So "Any" type of exhaust leak ahead of or at the O2 sensor allows oxygen into the exhaust stream and the O2 sensor sees it.
Exhaust leaks especially pinhole leaks at bung welds and all welds in the exhaust system are grossly overlooked when attempting to diagnose lean conditions. They are probably the most over looked cause of a lean condition.
Not just by guys like you and I but dealers,shop, everyone.
Also pinhole leaks in welds are damn hard to find.

Common lean conditions as caused by Intake Vacuum leaks are actually simple to diagnose. The methods I described for you seldom fail if done correctly.
But when people have lean conditions caused by exhaust leaks (especially small ones) or failing or biased sensors (other then the MAF) they get stuck. The are convinced they have a intake vacuum leak.

Another thing I think happens is a mental block.
Most people think of lean conditions in terms Fuel/Air ratio.
When they think of exhaust they think in terms of that fuel/Air ratio after combustion. Few people consider the possibility of ....Exhaust/Air..... ratios changing "after" combustion.

Hope this sheds some light on pinhole exhaust leaks and there relationship to The O2 sensor

Sent from my SM-G970U1 using Tapatalk

[ike84]

Ike -

thats really intersting, i truly never would have thought that air would be able to get through a pin hole like that. now that im thinking more about it though, i suppose thats how the exhaust sucker setups for negative pressure on a crankcase works! thats brilliant!

Sent from my SM-G970U1 using Tapatalk