E46 Fanatics Forum banner
1 - 20 of 77 Posts

39 Posts
Discussion Starter · #1 ·
As many other before I'm fighting for a year now against the infamous rough running when engine is cold. Engine always start from first crank and run fine for 5 to 10secs and then start shaking more and more if no throttle applied until throwing codes (P1349 or P1347 most of the time, sometimes P1345 or P1353) it then run very rough as DME shuts injector on one cylinder.

If I drive straight away, the rough idle and shakiness disappears as soon as the temp gauge reach vertical and no codes are thrown.

My car is a 2003 330i with M54 engine, manual tranny. I bought the car as is at 146 000km/97K miles, previous owner strangely forgot to mention the issue. Based on the previous maintenance bills, it seems that the problem appeared around 89 000km/55K miles (they started to change plugs & coils multiple times)

When cold, it is pretty much impossible to keep steady RPM using the throttle, engine goes up and down on a 500rpm range.

I'm also concern about throttle response when at temperature : it feels like there is a turbolag, it is worse at low RPM. From what I read it is normal on a E46 but I have never drove another one to compare. (applying the throttle adaptation reset procedure doesn't make any difference)

Voltage is fine : from 14.0 to 14.2V at idle

Here is what has been done or replaced so far but with very few improvement except for fuel consumption (around 10.5L/100 (22MPG) on city) and may be a little bit better engine response.

- Compression checked and look good
- New Spark plugs
- ICV cleaning (was full of soot)
- New Throttle body cleaning + Gasket
- New Lower intake boot
- CCV replacement + all associated hoses
- New dipstick tube O-ring
- 1 bottle of injector cleaner in the tank
- DISA valve aluminium kit from i6automotive + O-ring
- Replaced pretty much all small vaccum lines as they were toasted (F-connector to fuel filter, from SAP valve all the way to the intake manifold, rubber plug on the back of intake manifold, vaccum checked at the exhaust flap solenoid in the trunk and is fine)
- New Fuel pump
- New Fuel filter
- New Air filter
- New Vanos seals
- New VCG
- Cleaned MAF
- New intake air T° O-ring
- 6 New air distributor O-ring
- New exhaust CPS + O-ring
- New thermostat

Codes are not coils dependent as they don't follow when swapping coils.
It doesn't seems to be O2 sensor related as they are not supposed to be used by the DME until engine is up to temp.

This issue forced me into tons of very interesting reading so hat tip to all forum members for their great post and DIY.

I have to admit that I really enjoyed spending time understanding and working on my car but at some point it would be even better if I could finally solve my issue.

So here I am asking for your kind help since I'm pretty much running out of idea

Here are the actions that I plan to try next :

- Smoke test (I read that sucking jet pump could be leaking air in some cases)
- Injectors microwave cleaning + new top and bottom O-rings
- On a last resort : flash with last DME update using winkfp and a USB/K+DCAN adapter

Does anyone has any other suggestion ?

I have a bluetooth OBD2 adpater with torque App so I can provide some data if required.

Thanks for your time.

Premium Member
28,872 Posts
Suggest you read the 1st & 3rd links below in my signature.

Smoke test is the best thing in this case from what I can tell. Please note you really need to smoke test both the intake air path, both sides of the throttle butterfly AND the crankcase.

Do not forget to also check the power brake booster and the vacuum line under the drivers floor pan that connects to the fuel filter.

Cracked valve cover gaskets and/or leaking valve cover gaskets are a big problem with cold start as well.

Also suggest you read these thread on how to use Torque to graph your O2 sensors:



Also Freeze Frame and Fuel Trim data at warm idle and steady highway cruise would be helpful to figure out what is going on.

Not likely the coils, plugs or even DME flash.

Given that the problem seems to be the #3 & #4 cylinder, I would pull the DISA again and inspect this very carefully and make sure the large O-ring was installed properly. The original silicon in the DISA groove needs to be scraped out and the O-ring fits in the groove. The DISA should also have some resistance when removing or installing the last 1/4 inch.

Also make sure the DISA vacuum diaphragm is not blown, this happens often.

Depending on your MAF, 3 wire or 5/6 wire, there may be an IAT sensor between intake runner #3 ***4, if there is no sensor, there is a dummy plug. Either way there is an O-ring on this sensor or dummy plug, part number the last line in my signature.

Good luck.

39 Posts
Discussion Starter · #4 · (Edited)
Thanks guys, I can tell you that I know these threads very well, almost all the actions I performed earlier have been based on information found inside them.

Here are additional information following your reply :

DISA valve
The O-ring is fine as I scrapped the old silicon gasket myself with a tiny screwdriver for an hour before putting the new one and the valve now sits firmly to the manifold (need to wrench the screws to finish positioning).
The vaccum diaphragm appears to be good as well, I tested it using the 'finger on hole' method.
It used to make some buzzing noise before I changed the flap bell to aluminium as the old plastic one was starting to give up, noise is gone now.

Valve Cover Gasket
Mine is brand new, I replaced the old one that was in a very good shape as well (it had less than 10K miles). I put RTV sealant on the half moons and junction as advised so for now I would say I discard this hypothesis. Is there any way to discard it for good ?

MAF/Intake T° sensor O-ring
My MAF is a 3 wires and I did replace the small O-ring on the T° sensor located between intake runner 3 & 4.

Power brake booster
I tried to listen to any hissing noise from inside the car that stops when applying the brakes => negative

Fuel filter vaccum line
flexible hoses on both ends are new, I didn't check the rigid tube for cracks as it was not mentioned as prone to failure but I can surely do that.

Smoke test
Is there any comprehensive DIY on how to smoke test effectively ?
I thought about building a homemade smoke tester and use it on the brake booster vaccum hose and on the big nipple with a plug on the back of the manifold. Don't know how to smoke test the crankcase, from the dipstick tube may be ?

This morning it took a look at my STFT at warm idle : -3,1% for both banks but I think these current values are somehow not representative at the moment as I used to have very high values (up to 15+%) before changing the small SAP vaccum lines that run on the back of the maniforld few weeks ago. LTFT are around +7% so I think I may have to first unplug the battery for few minutes in order to reset adaptation values. (there may be an easier way to do that?)

About freeze frame I may have an issue with torque as it does not register any correct values (RPM at 0, STFT at 50% .....). I will have to fix that as well

I will graph my O2 sensors for 3mn from cold start as described in the posts you gave and will provide it. Since the capture has to be done with no throttle input, my car may throw an error code and cut off fuel on one cylinder. will see.

Premium Member
28,872 Posts
A few additional things.

You mentioned the valve cover is new, but this does not mean it did not get cracked when installing it. Not saying this is your problem, but you do need to keep an open mind when trying to resolve these type of problems. Sometimes you need to double and triple check things you think are fine.

But overall you appear to be doing a very through job trying to correct this problem.

If all your misfires are with Cylinder #3 & #4 I am thinking possible vacuum leak in this area that has not been identified. The brake booster connection is centered in the intake directly behind the throttle body as I recall. If the brake booster is leaking, this could be a source of your issue. Simple, quick test would be to disconnect the brake booster and plug the hose or pinch the hose off for a few cold start tests. But this may not the the final test as the other end of the hose connects to a strange grommet on the back side of the intake near the throttle body, very hard to see/access and I think even the grommet is very hard to find a replacement for.

Air distribution manifold can get oil fouled due the CCV design. Might consider replacing the air distribution manifold?

Smoke test, might be worth making a tester. You could try the Scotty method as seen below, but a machine make it easier. One note on smoke testing for only a cold start problem is there may be a minor crack or gasket/hose leak that can only be detected when the engine is stone cold. And even then a smoke test is not always 100% because of a minor crack or leak may not show up easily.

Intake gasket where the manifold meets the cylinder head. I have had these on other cars that I was finally able to detect while spraying brake cleaner on the intake/cylinder head joint while the engine was warming up and the engine speed would noticeably change.

Smoke testing the crankcase can be tricky. Dipstick tube may work, but if it is submerged in the oil, it may take some pressure to overcome the oil. You could try blowing in the dipstick tube and see how easily you can get air moving, otherwise a spare oil fill cap with a connector is the best option. If you use the dipstick tube leave the oil fill cap off until you have smoke coming out the oil fill then cap off the crankcase and watch for leaks.

Possible cracked intake, the issue with all the plastics, if there is a hairline crack it may leak vacuum on initial start, but the plastics are thin and will quickly expand and seal up a bit, this on top of the fact that after about 60-180 seconds on cold start the enrichment is no longer really needed and small vacuum leaks or other problems are less like to be a contributor after the cold start enrichment phase.

You could change ignition coils around just to see what happens, but your problem sounds like a lean cold start or unusual engine mechanical problem.

Cracked or burned exhaust valves could be suspect, but would not go down that path yet.

Injectors could be sent out for cleaning and testing, this takes a few days depending on where you are located.

If you have any cold start data, send it to me even with codes occurring. Also provide any Freeze Frame data you have.

Also try to get cold start data with you working the throttle trying to keep the engine from misfiring.

A few questions:

How many seconds into the cold start does the misfiring occur?

How many seconds will this problem happen for?

If you give the engine throttle and keep the misfiring from occurring, how seconds before you can allow the engine to idle without misfiring?

Does this happen if the car sits for a single overnight or does the car site for days/weeks between starts?

If you have any cold start data, send it to me even with codes occurring. Also provide any Freeze Frame data you have.

Also try to get cold start data with you working the throttle trying to keep the engine from misfiring.

Anyway, hope this gives you some more ideas to work with.

39 Posts
Discussion Starter · #7 ·
Yes I took good care of the new VCG when replacing it but I see your point about keeping an open mind

I cleaned the air distribution manifold with carburetor cleaner and pressurized air when replacing the 6 O-rings. I didn't damage it but it seems that time and heat did bend/shrink it a little bit : The nipple over cylinder 2 intake doesn't stay in place (the last millimeter or so) without pushing very hard over it. I workaround that using RTV around the hole, even though not perfect it seems to do the job for now. I thought about replacing it but this little piece of plastic cost around 50 euros.

Ignition coils have already been swapped with no success.

Cracked exhaust valve should show up during compression check right ? the one I did few months ago was really balanced on all cylinders

I plan to clean the injectors myself using a microwave cleaner and a DC generator to open injector while being cleaned. This will require some preparation on my side as well as the homemade smoke tester I want to make. Hopefully I will have time to do that in few weeks and smoke test it using the directions you provided.

As you mentioned, intake manifold gasket or brake booster connection underneath the manifold are good candidates for misfire around cylinder 4. Hopefully the smoke will provide answers as the brake cleaner method I tried failed on both.

Anyway, today I gathered data from an "almost" cold start as I did it coming back from work (car sat for only 6 hours outside)
I let it idle for around 8 minutes and got a pending fault code P1348 but no CEL so I then drove away home. It is mostly city and no steady highway driving this time but I can surely manage to do that next time if required.

Data and graphs are available here : http://tempsend.com/931EF10BB2/F0A9/ck1caribou v1.xlsx
I'm no expert in this but it does look strange as STFT are all over the place and sometime go back to zero for no apparent reason. O2 sensors data didn't showed up from the beginning, I have no idea why. I may need to reduce the number of simultaneous recorded values.

Finally, here are my answers to your questions :

-How many seconds into the cold start does the misfiring occur?
Around 10secs, as soon as RPM goes under 1000RPM

-How many seconds will this problem happen for?
I timed it today and shaking stopped after 3 minutes 30 seconds

-If you give the engine throttle and keep the misfiring from occurring, how seconds before you can allow the engine to idle without misfiring?
It depends, on how cold the engine was. I would say it stops as soon as engine coolant is up to 70°C (160 F). so roughly between 3 to 10mn

Does this happen if the car sits for a single overnight or does the car site for days/weeks between starts?
It happens all the time when cold, doesn't need to sit overnight as proved today. Worse misfiring case I had was in the winter after 3 nights under freezing temps. Car misfired so badly that it died few times on me and I had to clear the codes as they didn't went away after switching ignition off.

Thanks for your support, it is really appreciated

Premium Member
28,872 Posts
Thanks for the reply and the graph link.

LTFT and post-cat O2 sensor data is in question.

Can you confirm the build date of your car? Need month and year.

Strange how the post-cat O2 sensors take so long to report?

Starting to wonder if the O2 sensor wiring is mixed up?

I have been seeing some of these cases lately. Not sure why you are only misfiring on a few cylinders, but again, need to keep an open mind and assume anything can be wrong at this point.

You are located in Paris France?

Does your car have catalytic converters?

What is the history on the O2 sensors?

Do the O2 sensors have any splices or homemade connections?

Need to get a better understanding about that is going on here.

39 Posts
Discussion Starter · #9 ·
My car has been built on February 2003, one of the last ones with 5 speed manual. Production code on realoem is EV51. O2 sensors are the ones with round shape connectors

I am indeed in Paris France and yes there are catalytic converters on the car.
I have no history related to O2 sensors, everything looks to be normal regarding wiring (not sure about what you mean by splices). I think neither pre or post-cat sensors have been replaced so far, at least it is not listed in the maintenance bill I got from the previous owner.

Premium Member
28,872 Posts
Ok, thanks for the answers.

I believe since your car is a pre April build then the O2 sensors are standard not wideband sensors.

Time to assume nothing is correct with the O2 sensors.

I think you need to unplug each O2 sensor one at a time, see what graph drops out as well as wait until you get a CEL and note what code is triggered.

What you need to understand is the O2 sensor connectors on the car are not marked Bank #1 or Bank #2. Also the front and rear connectors are a different gender.

There are 3 scenarios where O2 sensor connections can go wrong.

1. O2 sensor can be connected to the wrong sides so the O2 wiring is swapped side to side.

2. Connectors on car and sensors can be connected to each other due to the different genders on the sensors and engine harness connections. This could be just one side of the car or both.

3. Probably the worst situation is where someone puts the rear sensor in the front and the front sensor in the rear then connects the proper connectors.

Any of these situations can drive the DME and Fuel Trims crazy.

Best to trace and mark the connections on sensors, then disconnect the sensors one at a time and figure out what connection is for what sensor and label them.

I will need to look over the O2 sensor graphs one more time to see if I can figure out what is going on, but something looks out of place and not sure I can put my finger on the issue quite yet.

39 Posts
Discussion Starter · #12 ·
I did my homework today, unplugged every sensors one by one waiting for the codes to pop up
Here are the data and graphs associated with every tests : Link

Test 1
Unplugged Bank2 pre-cat O2 sensor => Got code P0050 and code P1348 as I let the engine idle from cold

Test 2
Unplugged Bank1 pre-cat O2 sensor => Got code P0030

Test 3
Unplugged Bank1 post-cat O2 sensor => Got code P0036

Test 4
Unplugged Bank2 post-cat O2 sensor => Got code P0056

I can confirm that pre-cat sensors are correctly plugged.
Regarding post-cat ones, I didn't follow the wires all the way to them as it requires to have access underneath the car.

lszlszx, could you please provide details about the O2 sensors connection wire routing ?
I assumed that once it reaches the back of the engine it then goes to the engine wire harness located close to the throttle body and then to the DME inside of the big black cable duct. Do you confirm ?


2006 330ci zhp
2,043 Posts
Isn't the engine in open loop on cold start? Then the O2 sensors wouldn't have anything to do with this issue.

I had the problem a couple of times. Seemed like my issue may have been related to lazy lifters. I tried a bottle of Lubegard Oil treatment for the hell of it. I've never had the problem since. As a matter of fact, the engine seems to start much easier. It roars to life on cold start, initially going over 2k rpm.

39 Posts
Discussion Starter · #15 · (Edited)
Yes that is what I'm thinking to. Even though the post-cat O2 sensor may be shot, it shouldn't have an impact on my issue as engine doesn't use them on cold start.

On test 3 & 4 I unplugged both post cat O2 sensors assuming that they were correctly wired (Bank1 O2 sensor plugged to front connector for bank1 and to rear for bank2). Based on the error code thrown, we can validate the fact that the connectors are in the correct location but we can't validate the fact that bank1 sensor is truly connected to bank1 connector and the same for bank2, can we ? Is there a way to validate that looking at the graphs ?

Another question, why the LTFT are +7% at startup. Is it a common value or a proof that engine is running lean for quite a while ?

Thanks for the tip Dave, I will have a look at what similar Oil treatment product I can find here in Europe. How many miles did your car have when you had the issue ?

Premium Member
28,872 Posts
Do not have much time today or over the next few days, I may drop in from time to time as I get a chance.

A few things/thoughts:

Post Cat O2 sensors appear to be fine, the problem is they are really operating in either extreme, which is very unusual for Post Cat O2 sensors.

Almost seems like the wiring is switched bank to bank?? The Post O2 Cat sensor wiring is a PITA to follow, it might be easier to label and swap the Post O2 Cat wiring and see what happens. See how the sensors graph out and if they appear to be a bit more on the normal side.

One Post Cat sensor is reading that the fuel mixture is very lean, could be due to misfiring or injector fuel shut off, however, I would expect to see the same signature in the Pre Cat O2 sensor for the same bank.

Again, does this have a direct impact on the cold start issue??? Many will say no, but we have an extreme case here where the car does not run properly and there is no specific clue or item that is obvious at fault. Also almost all of the expected/known problem areas have been checked and corrected.

I do not think this is an engine mechanical fault, however, this may need to be revisited.

There is a lot of hydraulic lifter comments out there in the field, not sure any of this has really be substantiated. Also if this was a lifter related problem, I would expect the problem to clear long before the engine is fully warmed up.

Long before I would be considering any lifter replacement, I would be replacing something like the intake manifold gasket. But not sure I would go there yet.

Although problem starts when the engine is in open loop, the engine is into closed loop fuel control in one of the graphs in around 70 seconds. The rough running continues until about minute 4.

Once the engine goes into closed loop, the cold start enrichment phase is no longer active.

I would also be looking at something strange like maybe???? the #3 & #4 fuel injector wiring swapped. Not sure this is possible, but this should be checked.

The other thing to check is the signal and ground wiring on one of the Post Cat O2 sensors may be swapped, keep in mind this could be either on the cars wiring harness or could be on the O2 sensor connector OR if an O2 sensor was replaced or wiring damaged at some point, someone may have done a repair between the sensor and the connector on the pigtail.

Could be a universal sensor was installed, could be someone repaired damaged wiring, could be someone ordered the wrong sensor with a short pigtail and make a field splice for the O2 sensor wiring.

Remember, you have to assume nothing is correct on the car, otherwise you may fail to find a very simple and obvious issue.

I do not expect anything like a fuel pump relay to be a problem, however, it is not a bad maintenance item to consider replacing.

The more data we have the better.

Although this misfire starts in open loop after about 10 seconds of running, some things like the leaned fuel trims can cause issues even with cold start or at least give some clues to cold start issues. Also many do not think the Post Cat 02 sensors do anything other than monitor the catalytic converters, this is not true on some platforms.

Was working with someone the other day that claims Torque my be interpreting the wrong bank on some of the O2 sensor codes. I have not been able to compare and verify this, however, keep this in mind and if you have another OBDII tool, you may compare results.

One other thing that seems a bit strange is I do not see the SAP running?? I would have expected to see a gradual drop in all 4 O2 sensor Voltage while the O2 sensors are warming up and the SAP is running, then once the SAP shuts off the O2 sensor Voltages usually peak at their highest Voltage before going into full toggle mode during closed loop operation.

At least from the first data that we have, the LTFT seem to be running in the 6-7% range, which means the engine appears to be running lean.

So I understand you want to get this issue resolved, but be patient while you gather data and this is issue is logically looked at.

335 Posts
Following this, this is great troubleshooting by All involved, had similar issues and mine ended up being my FPR hose was non-existent, i replaced quite a few thigns before the fuel filter, jumped under car and there was NO line running back... I have both post cat o2's bad currently as well, I cleaned the front two and they are fine, the two rear I may try to clean as well but may just get some new ones, they are just expensive...

Premium Member
28,872 Posts
Following this, this is great troubleshooting by All involved, had similar issues and mine ended up being my FPR hose was non-existent, i replaced quite a few thigns before the fuel filter, jumped under car and there was NO line running back... I have both post cat o2's bad currently as well, I cleaned the front two and they are fine, the two rear I may try to clean as well but may just get some new ones, they are just expensive...
Cleaned the O2 sensors?? Not sure what you do to clean them.

O2 sensors are considered consumable and need to be replaced, similiar to spark plugs, but they can also last a very long time under good operating conditions.

You mention your rear O2 sensors are bad, what is the clue/situation with these. Rear O2 sensor wiring routing is a PITA as well.

But usually the rear O2 sensors fair better than the fronts on a well maintained vehicle.
1 - 20 of 77 Posts
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.