Log graph of MAF output via Durametric

OK, I haven't played around with the Durametric software too much - but I think I need some help. Background: My 996 has had a stumbling issue for almost a year now. Used to happen rarely, typically the first stop light after prolonged cruising on the freeway. It got bad enough that I couldn't start from a dead stop without the engine cutting in and out. Makes for a bad drive. This weekend I cleaned my MAF with some MAF cleaner and it's significantly better. But not completely fixed - still stumbled once on my drive to work this morning. I figure I'm in for a new MAF, but I'd like some definitive proof first - especially because I already own a Durametric. Ideally I'd like to graph the MAF output to a file with my computer and the next time it stumbles, see some wacko reading come off of it. Is this doable?

Don't know what year is you car but if it's not e-gas, you may want to clean the idle stabilizer. If it's e-gas, you may want to do a throttle body recalibration. Yes, you can log the MAF reading in Durametric and store it in csv format and do post processing with excel afterwards.

How is a throttle body recalibration done? Thanks

Turn key to last position before cranking with foot off the gas. Wait for 30 seconds then turn key to off. You should hear the throttle body servo rotating with some high pitch electrical buzzing noise during the 30s.

Ahsai, It is a 2002 - so egas. I'll try recalibrating the throttle body first as I have disconnected the battery a few times over the past year.

I think it will be beneficial to log the FRA and RKAT values as well. Also check if the evap regeneration valve above the alternator is leaking since you said it happens after along highway drive, where the canister will have very little fuel left in it so any leak will become a vacuum leak. So maybe you (your assistant at the passenger seat) will see that from the RKAT.

Quotes source post: 12038833

You sure it is not 60 seconds?

Quotes source post: 12039674

30 seconds worked for me.

Slight update here. Did the throttle body recalibration last week and promptly left the car at the airport for a week. However it still gave me some hiccups on the drive there and on my commute to work today - so next I plan on logging the data via Durametric. Need to get the cable all setup so I can plug it in and go on my next drive.

Quotes source post: 12038900

I agree with Ahsai, a stuck open evap valve could fit the symptoms described. I assume there are no fault codes.

Quotes source post: 12064415

Correct - no codes logged. Umm - anybody got a DIY on how to replace the evap valve? Is this what I should be looking for?

Quotes source post: 12064656

You say the CEL is off and there are no active or pending codes? You say you left the car parked at the airport a week? Do you do this often? You know where I'm going with this: stale gas. 'course, the symptom has been there a year. Unless you drive the car so seldom as to not add fresh gas for a year... Speaking of year, hard to imagine a MAF acting up for a year without throwing any error codes, or more likely just failing outright. In fact hard to imagine anything that would cause rough running for a year not to throw any error codes. I'm thinking plugs, coils, MAF, AOS, evap valve, anything with the vapor recovery system in fact. Does the engine run ok at higher RPMs? If so, I'm thinking an air leak. Had an air leak that triggered a CEL every once in a while. The error codes pointed to the MAF in that they were mixture related. I do not recall the error codes now. Finally convinced the MAF was at fault -- even though I never could spot anything amiss looking at the air flow reading in real time -- I replaced the original MAF with a new MAF but the error codes came back. Turned out to be a leaking oil tube filler cap. This leak was only significant at low engine RPMs, at idle. The engine seldom idled long enough to trip a CEL. I'm not sure you're going to see much in the way of a smoking gun looking at the MAF output. Observe the short term fuel trims. These want to vary a bit on either side of 0 (zero) but not by much. If you see swings of 4, 5 or more percent when the engine is warmed up and the rough running appears that's a sign the MAF could be at fault. Readings close to or over 10% (+/-) will almost always cause the CEL to come one with a mixture related error code. Had a failing AOS, well, what in hindsight proved to be a failing AOS, that would cause a bit of a stumble when I stopped at a light and the engine RPMs dropped to idle. This was after leaving the freeway after a long drive, leaving the freeway to fuel up or for the night. The CEL would come up about the same time the engine stumbled. Not a big stumble just a enough though to make my heart skip half a beat. Based on the behavior and codes I put this down to a bad MAF. But at the end of my 2K mile road trip over the span of a several days the next morning's code start had the engine smoking like crazy. AOS, obviously. Hard to believe an AOS in your car's case could be causing this rough idle and yet have not failed outright. Given a MAF is so expensive I can't recommend you replace the MAF even if you see the short term fuel trims going wacko. There can be other explanations for this behavior. You know to there is the poor man's MAF test and that is to disconnect the MAF at the wiring harness and clear the codes -- to reset all the adaptations to their defaults -- then drive the car as you were before and see if the behavior appears. If it doesn't that's about as good a smoking gun as you'll find without taking the car to a shop with all the Porsche diagnostic tools and training/experience to use them to their fullest.

Quotes source post: 12064745

Yeah - I do leave it un driven for periods of time. I've had stale gas issues as well - but I've cycled 3 or 4 tanks in the last 2 months, driven the car 'often' for my standards. The trouble started when I was driving it alot over a year ago. I really really doubt stale gas. I agree - no codes - I'm not thinking it's the MAF anymore, but I'd love proof. I will admit, this problem started roughly around the time I changed my plugs a long time ago. Just thought of that. Wouldn't that eventually throw a code?? Engine runs great up to 6k. I don't take it up there much, but never a problem there. Only stumbles accelerating from a stop or <15mph in 1/2 gear. Me too - I was hoping you guys could figure that out if I record some RKAT values. Heck, I don't even know what RKAT is. Got it. Car does smoke sometimes on startup, but I've attributed that either to stale gas in the past, or the 'cold' weather we've had here sometimes. Thanks for the long reply Macster.

RKAT is the mixture adaptation near "idle". FRAU is adaptation for the middle range and FRAO is for the upper range. Post your results and we probably figure it out. Does it idle smooth after some period of time or always rough?

Quotes source post: 12064878

Based on what you posted I'm inclined to blame the plugs, or coils. If the behavior appeared concurrent with plug change the plugs or coils have to be suspect. What plugs did you use? I ask because in at least one case a Cayman owner replaced the plugs in his car's engine, and used plugs that were according to one cross reference chart the "same" as the Porsche plugs, but which proved to not be the same to the point he had to remove the new plugs and install the real McCoy plugs. I have to comment further on this plug business. The Porsche plug has an extended tip. The benefit of this type of plug, of this extended tip, is at low engine speed the tip is exposed to the heat of the combustion and thus runs hot which helps to keep the plug clean. But at higher engine speed the extended tip is exposed to the incoming charge and thus is kept cool. Even with the right plugs installed there's always the possibility the plugs were not installed properly, or the coil/plug connection is not good. There is also the possibility a plug is just bad/marginally bad. This coil/plug connection is extremely important. Years ago in my 1996 Mustang GT the engine out of nowhere and out in the middle of nowhere developed a miss at freeway cruising RPMs. I say miss but there was no check engine light. But the behavior was real enough. Long story short I traced the problem to a bad coil to plug wire connection. It appeared that when the plugs were changed the tech failed to secure this one coil wire to its plug tip. For a while this was ok but over time the plug tip eroded until the miss developed. But at this time the tip was eroded enough that the coil wire would not snap onto the plug tip like it was supposed to. My fix was to use pliers to squeeze the coil wire connector down so it would "snap" onto the plug tip. As a back up as soon as I could I bought a new set of coil wires but never used them. So you have to be sure each coil is connected securely to its respective plug. And you have to be sure each coil is likewise securely connected to its respective wiring harness connector. Techs tell me that often a car comes in with a miss and it is the coil connection at the harness that is at fault. Thus my advice would be before you looked too far afield for the cause of the engine behavior you should seriously consider revisiting the plugs/coils to ensure everything is secure. And if you strayed off the reservation regarding spark plugs to seriously consider replacing the existing plugs with genuine Porsche plugs. You don't have to buy these at the dealer parts department, you can buy the same plugs from another business I guess. The important point is to buy the right plugs. Oh and then install them correctly and of course this includes making sure each coil and plug connection is sound and each coil and wing harness connection is sound. To kind of jump around and to address a possible intake leak, do not overlook the oil filler tube cap. With the engine warmed up and idling see if by just trying to move the cap you can cause the engine to react. You might be able to hear the air leaking past the cap. A co-worker of mine heard this which is how I learned the cap was leaking.

Ok, here it is. We're looking at ~80 seconds of data here - the failure happened multiple times, you can fixed - might see the RPM oscillations. I think it happens right when the MAF readings jump way above 400. That was the car 'hiccuping' or stumbing during acceleration. This is at the end of a 10 minute drive - car was running freeway, this happened in a parking lot. I'm also going to attach the raw data in excel format. The few things I notice. FRA and RKAT are near constant for the 'failure' period. FRA between bank 1 and 2 is pretty closer. However RKAT is way off between the two banks. Does this matter?

Two more things I noticed. No misfires were recorded during nearly an hour of driving. Second - My MAF & Hot Film MAF values drop to 0 6 times, I'm pretty sure at the exact point of failure (hard to tell - time based). That can't be right, can it??

I can't tell too much from your graphs and spreadsheet. RKAT bank 1 is ok even though it is different from bank 2. RKAT and FRA don't change instantly. MAF mass flow and voltage values look ok. Since there are no fault codes the problem is not occurring (or occurring long enough) during routine system checks to be flagged. I would take a look at; 1) cam deviation for both banks - should be close to zero 2) actual cam advance verses spec angle - actual should be close to spec 3) disconnect the electrical plug on the evap purge valve and see if the valve is leaking. 4) plot rough running for all cylinders at idle and when it stumbles. 5) try disconnecting the battery to reset all the adaptation values and then read RKAT while the engine idles for a while. Don't forget to turn the key on and then off without starting the engine to set the throttle. Plot the minimum variables too.

I think George already covered a lot. The only thing I want to add is like you observed, the MAF signal drops to 0 when other signals around those dropouts are smooth. Now I highly suspect your MAF as you also said the hiccup occurred during acceleration, where the MAF input has great effects. You can see that the MAF signal drops to 0 are real (as opposed to diagnostic I/O errors) because you can see bank 1 FRA reacted to it by dropping from 1.04 to 1.01 and bank 2 FRA raising to 1.02 from 1. Since the MAF signal drops, bank 1 removed fuel but then bank 2 added back fuel to compensate for the wrong adjustment (in reality the air sucked in didn't change during the dropout and didn't dip to 0 in 100's of milliseconds for sure). You can also see the more abrupt throttle position changes around those dropouts as well, could be from your gas pedal input reacting to those. Since MAF is expensive, you may want to get a professional opinion before getting a new one. Another way is to remove the MAF and drive the car around and see if you still have the same problem. Or if you have a friend who can lend you his/her MAF for testing...

Quotes source post: 12072420

I need to think about this more. In the meantime I found it equally interesting that the MAF jumps high before it goes to zero. see below.

I am not convinced you have a bad mass flow sensor and would try logging the following in 2 tests; First increase the sample rate to maximum. 1st data set; RPM Mass air flow (the one without tank ventilation) in kg/hr Injection time oxygen sensing bank 1 oxygen sensing bank 2 Actual throttle value (not throttle position) Actual cam angle bank 1 actual cam angle bank 2 In this run I would be looking to see 1)what the injection time does when the MAF goes high and then low, 2) if the oxygen sensing goes rich in both banks during acceleration. This is instantaneous unlike adaptation values. 3) see if both cams angles are about the same. 2nd data set misfire counter for cylinder 1,2,3,4,5,6 In this run I would look to see if there is any particular cylinder missing or entire bank or all when the engine stumbles. If I recall correctly "misfire counter" will record a misfire and not necessarily throw a CEL. Let's see what you find!

Quotes source post: 12073039

George, good point and observation on the rise in MAF reading preceding the 0's. In the 1st data set, is it better to log the MAF voltage instead as I think that's the raw value? I assume voltage is the direct reading so if we see 0v, that would be pretty convincing that the MAF signal is breaking up/grounded. Although MAF in kg/hr should be the direct translation from voltage but I'm not sure if there's any substituted values used when the input signal drops out. Also it may be beneficial to log the gas pedal position reading to make sure that it's clean. It can also help use separate the cause from the effect some.

I just want to say - Thanks George, Ahsai and Macster for all the help. I did a short driveway experiment last night - no failure point, just running values: Camshaft position 1 deviation: 3.52 Camshaft position 2 deviation: 2.06 Those values never move, I read +/- 6 is spec, so I guess I'm ok there. I watched Spec Angle vs Actual and they stay very very close, trending right next to each other in the live graph. I took the truck to work today, so I'll have to do these experiments tomorrow. I'm going to do a third one as well - drive with the MAF disconnected. I want to blame the MAF - it's $$$, but sooooo easy to replace.

Experiment 2-1 and 2-2 completed. The first graph is both MAF (HFM) and Mass Air Flow [as selected in Durametric]. I got the failure at the end of my drive home - the MAF (HFM) goes to 0 on 7 different instances. The second graph is RPM. Nothing exciting there. The third graph is labeled. Few things I've noticed: 1) Although MAF (HFM) goes high and then dives to 0, the Mass Air Flow doesn't. 2) Injection time seems to stay high for a period before the failure happens. I recorded oxygen sensing, cam angle, throttle and pedal values, but I don't know what to make of them. Next experiment (same dataset): So I have misfires. Not included in this data is about 60 minutes of drive time where I got other random misfires (maybe 4-6 total). Great So - I'm not sure I'm any further with this new data. I know the MAF (HFM) goes to 0 - which it shouldn't. I know injection time stays high before the failure (makes sense, I'm accelerating from a stop usually).

Those 0 MAF signal lasted only a split of a second each time so I wonder how much (if any) they contribute to the hiccups. Now you said you saw random misfires. I take that as misfires in different cylinders? The excel shows only cyl#5 had misfires. How old are the coils and are the connections all tight?

interesting data. i can't figure out anything conclusive but saw that when second misfire occurred was same time as when MAF went high and then to zero. see below. can you let it idle for a while and read RKAT?

Quotes source post: 12076940

It may be the zero MAF reading concurrent with misfire is just the DME shedding load and responding with no data to the Durametric request.

Mystery solved! It was a bad 5v supply wire to the MAF. It was intermittent but later on the wire broke completely. https://rennlist.com/forums/996-foru...0102-help.html

Yes! Very happy that this intermittent issue has been fixed - it was causing me to drive the car less and less. It's almost like having a brand new car in the driveway.

Quotes source post: 13220432

I'm very happy for you too Yours was a very interesting and atypical case. I can imagine...you were essentially driving with a crazy MAF The DME had to substitue predefined values whenever the MAF was losing the 5v reference.

[QUOTE=You know to there is the poor man's MAF test and that is to disconnect the MAF at the wiring harness and clear the codes -- to reset all the adaptations to their defaults -- then drive the car as you were before and see if the behavior appears. If it doesn't that's about as good a smoking gun as you'll find without taking the car to a shop with all the Porsche diagnostic tools and training/experience to use them to their fullest.[/QUOTE] Macster, this is a from this thread. Could you be more detailed about this procedure, such as when you say "disconnect the MAF at the wire harness and clear codes", is this with engine on or off? Off I would assume. Then when you say, "then drive the car as you were before and see if the behavior appears.", is this with the MAF harness reattached? Reattached I would assume. And most importantly, "then drive the car as you were before and see if the behavior appears. If it doesn't that's about as good a smoking gun as you'll find." So you're saying if it doesn't drive the same after this means MAF is most likely Bad or Good? And if it does drive the same this means MAF is Bad or Good? I'm sorry but they way you wrote this just confuses me. Sorry, if this is naive, but just want to do this right since it seems you know this is a good test for MAF And final questions, why it is that you have to disconnect the wire harness and then clear the codes? Is this different from just clearing codes with the wire harness attached?