e46 325i Running Crazy Lean when Cold

[bobinyelm]

2003 e46 325i Wagon 185k

I've been maintaining a friend's e46 for a few years (having a nice shop and a 2 post lift I find I have a lot of "friends" but I enjoy the friendship and the work, when it's not too frustrating), but this time I am stumped.

The car is hard to start (starts and immediately dies), but if you can play with it and get it to rev to 2500 immediately after firing it WILL run. If you keep it running, slowly as the engine warms it will run at lower and lower rpm until by 135 deg it will even idle. At morm temp (180-190) it idles like a Swiss watch and performs perfectly at all rpm and power settings.

It throws codes (when cold) for exceeding mixture limits (It's trying to go full righ, but it's still too lean). It also throws a lot of codes for "DME Ign Activation" codes and HAS for a few years, though it never showed any misfires. I put a set of coils in when the originals started failing, but while it runs perfectly, it still always showed the "DME Activation Ign" codes, so I don't think it has anything to do with THIS problem (plus being proprietary codes, they do NOT set a CEL. The misfire codes set when the engine is cold and the mixture is super lean before it warms up.

First I suspected a vacuum leak and applied smoke (I adapted a Disco Smoke Generator for smoke testing) and found a cracked MAF sensor boot split, and a new on helped a BIT, but not much. I couldn't find any more leaks. Then I thought maybe an eng or air temp sensor was lying to the DME, but they all showed w/in 2deg of actual ambient. The MAF shows normal readings (13.4 @ idle warm) plus I cleaned it.. When warm, the upstream O2 sensors look normal (see image I photographed off one of my my scanners), as do the downstream (though they don't alter mixture).

When cold (end even hot) the idle O2 signals are REALLY Lean (as expected) at 0.1v, which is why I suspected an unmetered air leak. I even took it to a Cert BMW Tech (independent shop) w/ 30yr experience figuring HE could find the problem, but he was stumped. He offered it "could" be Vanos seals, maybe fuel pressure (that's one thing I didn't check), but given how well it runs hot and the hot O2 sensor signals (Cycle.3v-.8v), I doubt it, though neither of my scanners will show a PID for STFT or LTFT, but just that the Fuel Trim is full limit).

I gather that until some temp is seen by the sensors the DME is Open Loop so the mixture is governed by some internally stored "nominal value" for mixture until the temp allows Closed Loop. Closed loop seems to be when the engine starts running normally I think (the O2 BiSi and B2S1 below show Closed Loop running for sure


.

The throttle body is clean, the IAC seems to be working properly (rock solid 750rpm or so idle hot), and the DISA is working (I can activate it w/ the scanner, so the diaphragm isn't leaking (plus I just put this on on a year ago for Lucas).

Any and all suggestions welcome-

Some days I am glad my DD is a Honda (though I will never sell my e24 635CSi). I confess being less enamored on Bimmers made after the e24/e28/e34/e36 era due to the complexity of the electronics.

Bob

Below is the DTC List and an Image of the 2 Pre-Cat )2 Signals showing Closed Loop O2 Cycling in case it rings a bell in anyone-

Luke 2.jpgLuke 1.jpg

[dworthy]

You do have a bit of a head scratcher here, so what are the STFT and the LTFT? Did the Indy give you those by chance?

Yes the oxygen sensors are cycling properly, but I don't think they are the problem as everything is happening in open loop operations, not closed as you noted.

The trouble with using generic hand scanners is you will not get all the info you need, unless they are loaded with BMW codes. Have you tried to use INPA or ISTA to look at the DME and the Live data aka PID's?

Which temp sensors did you look at, for there are two. One in the lower hose, and one in the head. Granted the lower hose is mostly for the electric fan, but it too is a data point the DME uses.

I agree that this is a cold start problem, and typically the engine will require more fueling. It would be worth testing fuel pressure at the schrader valve on the fuel rail here, should be 50 PSI +/- 5%.

Which brings me to my next point, did you take a look at the vacuum line on the FRP down on the fuel filter? Do you know how old the fuel filter is? BMW recommends replacement at 100k, but heck I wouldn't go more than 75k on them myself.

[whiskychaser]

Your codes E3 and E4 equate to BMW codes 227 and 228. They mean that the DME can't adjust the fuel any more for what it thinks is the amount of air coming into the engine. You don't mention it but I would expect it to trigger the SES light. So you have to decide if you have an air leak, not enough fuel or one of the sensors is telling lies.

The DME determines the rich mixture on cold start by reference to the coolant temperature sensor. So that would be a good place to start. You mention you have a MAF reading of 13.4. Assume that is kg/hr because if it is g/sec, it is about four times what it should be. As dworthy says, checking the STFTs and LTFTs will give you further clues.

[MarcoZandrini]

To add to the great advice already provided, you could take the car to a reputable indie shop that specializes in bimmers. Find a shop at www.bimrs.org. It sure sounds like you have a vacuum leak. The only way to find all of the leaks is to use a professional smoke machine.

[bobinyelm]

You do have a bit of a head scratcher here, so what are the STFT and the LTFT? Did the Indy give you those by chance?
None of the scanners used (One X431 Launch PROS, One Autel DS708 MaxiDAS, and an Autel MK906) show STFT and LTFT as I suspect they are PIDS not outputted by the DME. On MOST vehicle makes, sort term and long term fuel trim ARE displayed by these same scanners, so I think it's just how the car chooses to offer info

Yes the oxygen sensors are cycling properly, but I don't think they are the problem as everything is happening in open loop operations, not closed as you noted.
Yes, agree 100%

The trouble with using generic hand scanners is you will not get all the info you need, unless they are loaded with BMW codes. Have you tried to use INPA or ISTA to look at the DME and the Live data aka PID's?
These scanners are all pro grade scanners with bi-diectional capabilities, though not made specifically for BMWs. I and my Indie shop owner (a friend of 25 years) went though all offered PIDS as presented in Live Stream Data. As I mentioned, he's a factory trained BMW tech that has kept his BMW certification current by attending frequent training classes and he uses his Autel 906 daily on all manner of imports and specializes in BMWS, MB, Porsche, and I've been with him doing Aston Martins and Ferraris. That 906 can recode ecus and modules, so I think it is pretty sophisticated and cost him around $3k in US Version.

Which temp sensors did you look at, for there are two. One in the lower hose, and one in the head. Granted the lower hose is mostly for the electric fan, but it too is a data point the DME uses.
Looked at 1) Eng Coolant Temp, 2)Eng Oil Temp, 3)Eng Air Inlet Temp, and 4) Rad Outlet Temp, and all agreed w/ the proper ambient +/- a degree or two before engine start cold (39deg when I checked them). That was my first thought, actually since the DME uses coolant temp at least to set open loop fuel flow initially).

I agree that this is a cold start problem, and typically the engine will require more fueling. It would be worth testing fuel pressure at the schrader valve on the fuel rail here, should be 50 PSI +/- 5%.
Yes, and I MAY be able to remove the fuel rail cover on the e46 w/o pulling all the wiring over the back and the cowl/wiper blades like on the e90 (which intrudes even more). I HAVE a F/P test gauge and will do that when I retrieve the car from my friend's shop where it's now sitting. But if it DID have low fuel pressure, I'm not sure it would allow the vehicle to run perfectly when hot either, especially at high throttle settings, and produce upstream O2 voltage cycling as in the image I took (that was on the Launch X431 PROS). Certainly worth the time I guess to solve a mystery.

Which brings me to my next point, did you take a look at the vacuum line on the FRP down on the fuel filter? Do you know how old the fuel filter is? BMW recommends replacement at 100k, but heck I wouldn't go more than 75k on them myself.
He gave me his service records going back to when HE bought the car and started having me tend to it at about 130k if I remember right (180k now). He bought it from a husband/wife M.D. couple and they always dealer serviced it, but that doesn't mean everything. But if it were a fuel filter, you'd think it would restrict enough that at high rpm/high power would be where it would affect more since the required fuel flow at full-throttle 3000rpm is many times the cold idle flow.

I didn't snap an image of the LiveData, but at cold idle (as soon as it would idle that is with 120-130 deg or so coolant temp the injector time was around 4.5 ms. The MAF value was around 4-5 gr/sec on my tool which seemed reasonable too. His tool read in pound/min but he said it was good.

On the vacuum press line to the FPR, no, but we DID do a good smoke test and none came out, nor did spraying brake cleaner dow and around produce any reaction, so we "assumed" it was OK ("assume" is a dangerous concept I know)

Thanks. Definitely the next is to check fuel pressure.

Curious if you (or anyone) knows what the idle injector pulse duration SHOULD be for this engine at idle? Trouble is the vehicle won't operate at idle speed AT ALL until warm. In fact, below around 2500rpm cold it simply will not run at all.

It is as if the ignition is just plain SHUT OFF below that rpm-no misfire, no stumble. Just OFF like fuel or ignition is switched off. As the coolant temp rises, it will run at lower throttle settings/rpms until at say 135 coolant temp it will run roughly at 1000rpm, and by maybe 150deg it will idle at 750rpm foot off the throttle.

Bob

[bobinyelm]

To add to the great advice already provided, you could take the car to a reputable indie shop that specializes in bimmers. Find a shop at www.bimrs.org. It sure sounds like you have a vacuum leak. The only way to find all of the leaks is to use a professional smoke machine.

I AGREE it is classic vacuum-leak symptoms, which is why I and the Indie shop owner (my old friend) DID smoke tests (though my machine is a "home-made" his is a $900 pro model from MAC Tools off the truck), AND we liberally sprayed brake cleaner everywhere to locate a leak. No leak heard, either.
My smoke FOUND the cracked upper intake boot (one that goes to the MAF at the air cleaner), but it was a very minor leak and only made a slight running difference.

[bobinyelm]

Your codes E3 and E4 equate to BMW codes 227 and 228. They mean that the DME can't adjust the fuel any more for what it thinks is the amount of air coming into the engine. You don't mention it but I would expect it to trigger the SES light. So you have to decide if you have an air leak, not enough fuel or one of the sensors is telling lies.
EXACTLY, It was telling us that the DME was at MAXIMUM correction and even then it could NOT supply enough fuel to make the engine run cold. First suspicion was lying sensors, which is why I used live data to verify all temp sensors, and MAF values.

And YES the SES light IS on. If I cancel/erase it w/ the engine ALREADY HOT I can drive for long distances w/ no SES light coming back on. It's the cold start sequence until warmup that turns the light on.

The DME determines the rich mixture on cold start by reference to the coolant temperature sensor. So that would be a good place to start. You mention you have a MAF reading of 13.4. Assume that is kg/hr because if it is g/sec, it is about four times what it should be. As dworthy says, checking the STFTs and LTFTs will give you further clues.
Yes, and the coolant temp sensor agreed +/- a couple of degrees w/ ambient and also the inlet air temp, oil temp, and the radiator coolant output temp before engine start (Ign Pos II).

If memory serves I think hot I got 4.5-4.7g/sec on my scanner. Impossible to get a COLD MAF reading because at less than about 2500rpm the engine cold won't run at ALL. It's like the ignition or fuel is just flat CUT OFF and it dies.

It takes throttle pedal manipulation after start (anywhere from 1/4th to 2/3rds throttle position) to even keep it running until it warms up to maintain 2500rpm.

Thanks,
Bob

[whiskychaser]

Maybe check the air flow figures? 4.5 - 4.7 g/sec at idle looks high. Have 2.2 and 3.0 M54s - they both read 3.6 g/sec at idle when hot (90C+). You might also want to plot a graph of the air flow against the rpm starting at idle and going up in stages to say, 2000 rpm. The graph should be a straight line. If it is not, I would suspect the MAF. I doubt you will get accurate fuel flow figures using the schrader valve on the rail because it actually restricts flow. But you should be able to check the pressure easily enough. The fuel trims would be useful but if the adaptations have been cleared you will need to do a run to reset them.

[jclausen]

We had an e46 that had a leaking brake booster and the smoke came out under the dash

[dworthy]

We had an e46 that had a leaking brake booster and the smoke came out under the dash

That was another thought I had, as you can also induce a lean condition by pumping the brake pedal too.

[bobinyelm]

We had an e46 that had a leaking brake booster and the smoke came out under the dash

Good thought

I can unplug the booster hose and plug it as a check I guess.

Bob

[bobinyelm]

OK, and sorry for he delay after my last post, but the owner had to take his car back regardless of the poor running.

I did a full DME Live Data Readout when the engine was relatively warm (It won't run with enough stability when cold because the throttle needs to be continuously manipulated a LOT to keep the engine running-

PID: Value:

Eng Speed 702 RPM
Idle Speed Integrator 3.2%
Coolant 165 deg
Coolant Rad Out 66 deg
Oil Temp 135 deg
Intake Air 65 deg
Ambient Pres 1008 mBar
Air Mass 15.3 kg/h
Air Flow 0.8 v
Inj Duration 2.8 ms
Bat Voltage 13.6 v
Timing 8.6 deg
Vanos Inlet Spec 112.5 deg
Vanos Inlet Act 111.4 deg
Vanos Spec -105 deg
Vanos Ext Spec -105.7 deg
Oxy Sensor Control Bank 1 Not Active
Oxy Sensor Control Bank 2 Not Active
Oxygen Sensor Controller Bank 1 1.0
Oxygen Sesnor Controller Bank 2 1.0
Oxygen Signal Sensor Bank 1 Upstream 0.02
Oxygen Signal Sensor Bank 2 Upstream 0.03
Adaptation Mixture Add. Bank 1 0.00 ms
Adaptation Mixture Add. Bank 2 0.00ms
Adaptation Mixture Bank 1 Multiplicative 0.0%
Adaptation Mixture Bank 2 Multiplicative 0.0%
Air Adaptation w.o Driving Position 0.0 %

I did a F/I Fuel Pressure check and at idle it read 51psi, and 15min after eng shutdown 45psi so the check valve is working.

For the above readings, the engine was not quite up to full normal temp (about 1/3rd scale) vs hot at 190 deg coolant temp

NOTE:
1) I found it strange that it showed no additional fuel being added even though I get 2 codes, one for each bank showing Fuel Trim Limit Exceeded Bank 1 and Bank 2.
2) Is "Adaptation Mixture" the same as Fuel Trim? Strange it shows ZERO There is no readout for Short and Long term Fuel Trim (STFT and LTFT) as I do see that for most vehicles I scan. I can only imaging the BMW does not offer those as PICs.
3) Note that at idle at 150deg or so the Pre-Cat O2 voltages show 0.02v to 0.03V which is VERY VERY lean, but even with TWO smoke checks no leak could be found anywhere at all.
4) Note the Adaptation Mixture Additive for Banks 1&2 were shown as ms rather than %, so I guess it reads the actual extra milliseconds of open time for the injector commanded by the DME? More time = More fuel = Richer

As always, any ideas or feedback welcome!

Again, when partly warm (130deg + coolant temp) the vehicle runs fine, and the Pre CAT O2 signals are like in my first post in this thread

[whiskychaser]

It would be useful to know what your fuel pressure was when you turned the key and before you started the engine.

With regard to your notes:
1. Right now it is not showing fuel trims. But you can see that the DME is adding extra fuel - you have injection duration of 2.8ms. Normal would be about 0.2ms. IIRC, the max is about 0.35ms
2. Adaptation is not the same as fuel trims. If you have cleared the fault codes, the adaptations need to be reset. The engine needs to be hot and you need to perform a drive cycle or you will not set the OBD readyness codes. The DME will ignore the O2 sensors and do its own thing. Once the LTFTs are set, the DME adds that number to the STFT. If the total is above 11, it throws in the towel. It puts the SES light on and registers the codes in the fault memory.
3. At 15.3 kg/hr, you air mass appears to be too high. So you have to decide if you have an air leak or if the MAF is telling you lies. If the STFTs go down as you open the throttle, you may well have an air leak - the proportion of extra air coming in at idle is going to be small compared to the full bore of the throttle body
4. Depends what kit you are using. DIS will show them both as percentages. But more time does equal more fuel

[bobinyelm]

It would be useful to know what your fuel pressure was when you turned the key and before you started the engine.

With regard to your notes:
1. Right now it is not showing fuel trims. But you can see that the DME is adding extra fuel - you have injection duration of 2.8ms. Normal would be about 0.2ms. IIRC, the max is about 0.35ms
THAT is a biggie as I did not know what a normal idle injector pulse time is. 1/5th of a millisecond or 200 microseconds is damned short.
2. Adaptation is not the same as fuel trims. If you have cleared the fault codes, the adaptations need to be reset. The engine needs to be hot and you need to perform a drive cycle or you will not set the OBD readyness codes. The DME will ignore the O2 sensors and do its own thing. Once the LTFTs are set, the DME adds that number to the STFT. If the total is above 11, it throws in the towel. It puts the SES light on and registers the codes in the fault memory.
I essentially did that a few times (cleared codes, cleared adaptations, and driven the vehicle that would at least equate to a drive cycle). That tells me I think the main problem remains? The data does not display "ordinary" STFT and LTFT like I am used to seeing/reading so I'm not sure whether it's that the DME doesn't output it, or whether this scanner( Autel DS708 MaxiDas) with this vehicle won't display it. I could try a simpler scanner (Autel 509 I have). I have a Launch X431 PROS Mini but it shows the same PIDs as the DS708.
3. At 15.3 kg/hr, you air mass appears to be too high. So you have to decide if you have an air leak or if the MAF is telling you lies. If the STFTs go down as you open the throttle, you may well have an air leak - the proportion of extra air coming in at idle is going to be small compared to the full bore of the throttle body
I understand. At idle the leak is large by proportion of metered ar, but at more open throttle that same leak as almost inconsequential. What would you expect to see for a MAF reading at idle? Something like 13?
But a leak of unmetered air would likely LOWER and not RAISE the MAF reading (since not all the intake air would be going past the MAF)? So a HIGH MAF is more likely a BAD MAF?
As to "deciding" if I have an air leak or a bad MAF, given I did a smoke test as did a friend with an indie shop the easier (but not the cheaper) thing would be try another MAF, but I hate to spend my friend's $ frivolously, but on the other hand neither I nor my shop friend could find a vacuum leak, so maybe the MAF is the logical step? I don't think my shop friend has a customer's e46 in that we could substitute the MAF to see.
4. Depends what kit you are using. DIS will show them both as percentages. But more time does equal more fuel

Tonight was the first time I saw a MAF Sensor DTC displayed, but I have always seen the ignition codes on it, though I don't know why (even when it had no problems or SES light).

I'm afraid my diagnostic skills are sub par here... I appreciate the hand-holding.

[whiskychaser]

I may not have explained note 1 very well. Above 0.2ms STFT the DME is adding fuel. Below and it is reducing it. Note sure of the actual baseline on the 2.5 but I believe it is 1.8ms. Totally agree it can be confusing when trims are expressed in different units. Yes, I would expect your airflow to be around 13.2 kg/hr.

Probably the easiest way to check if you have a vacuum leak is to remove the oil filler cap. If the rpm does not go down a tad, it is likely you have got one. Don't repeat this often as it will end up throwing a code - because it is getting too much air.

Totally agree on not spending money unless you are convinced you are on the right track. Once it is hot, try graphing the rpm against the air flow in stages as I mentioned above. It was the only thing which convinced me that my MAF was shot. The graph should be a straight line - mine was like a dog's hind leg. When I compared the readings of the old MAF to the new one, the difference was anywhere between 16% and 40%.

[bobinyelm]

To check the brake booster diaphragm which I read can be a source of a vacuum leak I pinched the hose off but it made no difference, and to check a good pinch I had a hard pedal and zero boost.

OK, did what you suggested earlier-

Got the engine fully hot (around 190 coolant by scanner).
Reset Faults
Reset Adaptations
Drove car about 20mi until all readiness monitors reset.
Readings suggest still very lean, though only Bank 1), but I switched to an inexpensive Autel MS-509 and it displayed Fuel Trim numbers the others wouldn't.

Here are the latest numbers in BOLD. The MAF readings are in pounds per minute from the Ms-109 at the bottom.

2003 BMW e46 325i
(Readings below at 1/3 temp gauge reading (not fully hot)
Inside a shop at 60deg Ambient Temp)
Engine running too lean to idle until temp gauge is off the mark. Smoke test showed MAF boot leakage-replaced boot and no more smoke leaks, but still way too lean at idle.
Pre-Cat O2s show normal cycling w/ eng hot (0.3-0.8v) both banks.
DISA Valve not leaking vacuum and is 1yr old and operates normally w/ test actuation.
Pinched Brake Booster hose
Used bore scope to check throttle body for deposits-looked clean.
OBD II Scan shows the following DTCs
1) Code 8 DME Signal MAF Signal
2&3) Code CA DME Fuel Trim Bank 1 & 2 at Control Limit
4-9) Code 1 A-G DME Activation Circuit Cyls 1-6
10) Code D2DME Error. Ign Current Meas at More than 2 Cylinders


Test Readings:


Fuel Pressure Running 51psi = 3.5bar
Eng Speed 702 RPM
Idle Speed Integrator 3.2%
Coolant 165 deg
Coolant Rad Out 66 deg
Oil Temp 135 deg
Intake Air 65 deg
Ambient Pres 1008 mBar
Air Mass 15.3 kg/h
Air Flow 0.8 v
Inj Duration 2.8 ms
Bat Voltage 13.6 v
Timing 8.6 deg
Vanos Inlet Spec 112.5 deg
Vanos Inlet Act 111.4 deg
Vanos Spec -105 deg
Vanos Ext Spec -105.7 deg
Oxy Sensor Control Bank 1 Not Active
Oxy Sensor Control Bank 2 Not Active
Oxygen Sensor Controller Bank 1 1.0
Oxygen Senor Controller Bank 2 1.0
Oxygen Signal Sensor Bank 1 Upstream 0.02
Oxygen Signal Sensor Bank 1 Upstream 0.03
Adaptation Mixture Add. Bank 1 0.00 ms
Adaptation Mixture Add. Bank 2 0.00ms
Adaptation Mixture Bank 1 Multiplicative 0.0%
Adaptation Mixture Bank 2 Multiplicative 0.0%
Air Adaptation w.o Driving Position 0.0 %
The Following were recorded from Autel MS-509 Scantool
(These readings were unavailable on Autel DS-708 or Launch PROS Mini)
This scanner showed ONLY Bank 1 and Bank 2 Lean (no other codes)
STFT Bank 1 15.6%
STFT Bank 2 21.1%
LTFT Bank 1 0.0
LTFT Bank 2 0.0
MAF (in #/min) 0.425- 0.454 lb/min = (12.4 kg/hr)

I didn't take a F/P reading before start, but after sitting overnight there was some pressure left, and the pres came up to 50psi instantly upon start at least (I was inside the car and didn't observe the gauge until after start). Most BMWs I've had with bad check valves run smoothly seconds after starting even with no pressure until keyed. This one takes several minutes until the temp gauge is will off the peg. Twenty minutes after shutdown the system pressure fell to 45psi and then quite steady.

I noted today that occasionally Bank 2 upswtream O2 seems to go dead, but it doesn't affect the running, and it's back in a minute or so.
(See Below)

Normal Hot Running Upstream O2 sigals resenble:


Luke 2.jpg

[bobinyelm]

Tonight was the first time I saw a MAF Sensor DTC displayed, but I have always seen the ignition codes on it, though I don't know why (even when it had no problems or SES light).<br><br>I'm afraid my diagnostic skills are sub par here... I appreciate the hand-holding.

[bobinyelm]

Totally agree on not spending money unless you are convinced you are on the right track. Once it is hot, try graphing the rpm against the air flow in stages as I mentioned above. It was the only thing which convinced me that my MAF was shot. The graph should be a straight line - mine was like a dog's hind leg. When I compared the readings of the old MAF to the new one, the difference was anywhere between 16% and 40%.

So you are suggesting that I take readings of the MAF Flow vs RPM and plot it. Sounds like a good plan.

I guess you are meaning to do it unloaded in NEUTRAL with no load on the engine, yes?

It would seem that under load, the engine at higher manifold pressure (lower vacuum) the MAF would be greater at any RPM than if the engine were not in gear?

I will try that.

Frankly my friend is likely beginning to to doubt my ability to diagnose the problem (I cannot blame him) so he may take it away soon to find a shop that says they can fix it. Maybe I am missing something obvious I am not seeing.

There are several unknowns at work. One is the "DME Activation" codes on all cylinders, though they have been there a long time, yet they never seem to bother the vehicle operation. ANOTHER is the intermittent near zero Bank 2 O2 readings that again doesn't seem to affect the drivability.

Obviously, however, the BIG problem is cold start and running until the eng temp is well off the cold peg.

Absent a MAF problem, it ACTS like a pretty significant air leak, but again if that leak is SO BAD that it prevents the engine from running AT ALL below 2200-2500, wouldn't be big enough to cause poor idle or drivability when the engine has warmed up? That's a rhetorical question, of course.

I've just never seen a vacuum leak SO BAD it presented these symptoms cold, yet was unnoticeable (except electronically) when warm and presents a butter smooth idle, flawless acceleration and normal drivability.

I wonder about some mechanical problem (gap between two components) that leaks when cold, yet when warm causes the gap to close up, blocking the air that was leaking in cold. Probably sounds too bizarre to consider.

(BTW, that mangled post above is unfixable. When I try to edit it, it freezes my computer up every time and won't load well enough to allow me to edit and delete all that text).

[whiskychaser]

If you can't get rid of that junk, maybe message the mods and ask them if they would not mind doing it?

Plot a graph of air v rpm when the engine is hot (90C) and in park or neutral. I would use your diagnostic kit's rpm meter if possible. Going off the tachometer in the car is not going to be accurate. This might sound random but the check the connection of the plug on the MAF is tight and clean. Maybe confirm what the MAF code was?

[bobinyelm]

OK, After l-o-n-g drive NO SES light and no MAF code of course.
The MAF Code was "8 MAF" and checking online resources it verified "8" is an MAF code, but didn't specify more.
The Autel MS509 (simple scanner) would not read that code despite it said "Vehicle Specific Codes Found, Select Brand" which I specified as BMW.
It also didn't show all of the other Injector Activation Codes- Just the Bank 1 + 2 Lean Codes. The more sophisticated Autel/Launch displayed ALL codes.
In any case, the code didn't come back.
I removed/replaced the MAF connector many times along the way and sprayed w/ Contact Cleaner at one point. I also sprayed MAF w/ CRC MAF Cleaner
I noted the STFT MUCH higher at low RPM than higher RPM as one might expect with an unmetered air leak (the unmetered air becomes a lesser % of total airflow as the engine digests greater quantities of air w/ RPM of course).
The Bank 1 & 2 STFT were in the 20% range at low RPM but dropped to about 12% at 2500 RPM (in neutral)
I did the MAF vs. RPM Chart your recommended and the results are (hopefully) below, and not another huge mess like in the last post (I will contact the moderator-Thanks).
The chart was ruler straight to 4000 RPM then bent up a bit (dashed line is straight, solid line is actual). "Xs" are actual data points. I took the data using the tach, and went back and used live-data to check the tach calibration and it was never more than 30 RPM different, which amazed me.
So I think the MAF is no longer suspect.
I used a can of brake cleaner and re-sprayed everything I could reach and This Time I got a bit of RPM fluctuation (almost nothing, but consistent) when spraying down at a 30deg angle and aft just inboard of the oil filter, likely under #1 or #2 intake runner. There's a good bit of brittle "flex" plastic corrugated tubing under the intake manifold there as I remember from an e90 starter replacement I did. It had "squeeze and pull" ends that toast after several years. But even using a piece of rubber tubing to the ear and shoved in there I can hear no air leak.

Here is the chart (please computer Gods, show me a chart and not trash)
e46 MAF Chart.jpg

[bobinyelm]

OK, I give up. I tried to find a way to contact a moderator but didn't see one (to get the following post "cleaned up" of text trash

https://www.bimmerforums.com/forum/s...5#post30597765

Is there a clickable box or link for hat somewhere?

Thanks-

[shogun]

I deleted the non working image link trash. Try again , you can also try to test in the test forum https://www.bimmerforums.com/forum/f...hp?123-Testing

[bobinyelm]

I deleted the non working image link trash. Try again , you can also try to test in the test forum https://www.bimmerforums.com/forum/f...hp?123-Testing

Thanks!

[Muzz258]

Hi

Sorry about coming late to the party, been out road tripping and away from the computer.

Your symptoms are that of a bad air leak, as mentioned by others. I purchased a 6 cylinder Z3 that was delivered to me hot and it ran well. Cold start was the same symptoms as the car in question. Found a large hole in the rubber tee-off to the Idle Control Valve. Replaced the boot and all was well. Your smoke testing should have found this type of large hole easily. I found mine with a visual inspection.

Other areas that cause large air leaks and are harder to see are:
- Broken CCV tubes, specially the ones in/under the inlet manifold. The CCV drain tube to the dip stick tube is a classic.
- On the CCV regulator/cyclone there is a nipple in the center between the top and bottom parts. On this car it needs to be capped. The caps rot and fall off. Use a borescope to check it.
- The vacuum ports on the end of the inlet manifold next to the fire wall. The caps rot and fall off.
- If it is fitted with a secondary air pump, the rubber hoses under the inlet manifold.

[bobinyelm]

Great suggestions-Thank-you!

Bob