e36 M3 6500 rpm Blues

[PT356]

My new to me E36 M3 doesn’t like to be redlined in 3rd gear, which will trigger a CEL. The last test drive and resultant CEL revealed codes for cylinder misfires in 1, 2, 4 and 6. Some car history:


-79K miles; 1998 E36 M3 5 speed Sedan
-New INA lifters, NGK plugs, Bosch coil packs, Fuel Filter, Fuel Injectors and rebuilt Vanos
-Fresh oil change with 5-40W Mobil 1.
-Leak down test shows 5-6% for all cylinders


The car runs fantastic and has a very strong engine. However as I discovered shortly after purchase that if I hold 3rd gear to 6500 rpm the CEL will turn on and one or more cylinders shut down. This is resolved by stopping the car, shutting down the engine, waiting a few seconds and then restarting. It is interesting that I can make the shift from 2nd to 3rd at 6500 without issue.


From doing a lot of searching there seems to be a bunch of things that could cause this problem......my shortlist:


-Crankshaft Position Sensor
-Pre Cat Sensors
-Fuel Pump or Regulator
-Vacuum Leaks


I would like to be somewhat systematic about fixing the car as I would really like to know
the specific cause of the problem. I am considering removing the intake manifold to replace all of the vacuum lines but perhaps I should be looking somewhere else first? Would a dyno run while connected to diagnostic software be a good idea? Other suggestions on fixing this problem would be greatly appreciated. I have owned the car for 6 months and have only driven it a few times as I am concerned about not being able to drive the car like I would like to.

[bmwdirtracer]

Vacuum leaks don't cause problems at 6500 rpm; they cause issues at ~1000 rpm.
O2 sensors don't cause issues at WideOpenThrottle, because the DME ignores O2 sensors then.
CkPS doesn't make a habit of failing at high rpms, either.

Fuel delivery / pressure....most likely.

Does your DME have a chip/tune? You might well have a fault in that programming.

Most likely bet will be fuel delivery / fuel pressure -volume inadequacy at extended high load, high rpms.

Easiest way to tell without buying parts is to hook up a fuel pressure gauge, clip it under the wiper or tape it to the windshield, and then go re-create the issue, see if pressure is dropping off while symptom is manifesting.

[MauiM3Mania]

You may have other issues, but as I recall the redline of the S52 is 6800 in 1st and 2nd gears (and neutral). Once in 3rd, 4th or 5th, the factory redline is 6500.

Notice how the red line is not a hard solid line, rather it is gradated?

[PT356]

Chris,

Thanks for the information and I agree with the fuel delivery being suspect, which was why I replaced the fuel filter. Was not aware that vacuum leaks would primarily affect low RPM issues.......good to know! I went ahead and ordered a new fuel pump as they are not too expensive and the replacement process seems straight forward. By the way the engine is stock, no modifications. Will follow up with results once I make the swap and I am able to test.

Phil

[aquadave]

#1 you shouldn't run a stock engine like that. Next try something very simple switch from NGK to Bosh or Champion. I found years ago NGK foul quickly. Are you running white gas? Amoco and BP ultimate are premium gas if you're trying to blow your motor no better gas to do it with. When was the last time you rebuilt the motor? 98 is pretty old. peak performance requires a peak engine. Exhaust? Are you running a performance header system with cross overs? Old fuel has to leave for new gas to enter. Do you have AAA? I have a feeling you'll be in need of a tow truck soon. I'm pretty sure I have the tuning software for that car. buy the adapter and plug it into your laptop remap the system. I'll look and see

[PT356]

Aquadave,

Thanks for the response! I do have a AAA card however I do not plan to blow the motor just yet. The car has 79K miles and 5-6% leak down on all cylinders and has been very well maintained. I would like to think that a low mileage stock engine can be safely shifted at 6500 rpm without scattering engine parts on the highway. It seems that quite a few folks have relatively stock engines with tuning software that raises the maximum RPM's to 7000 RPM or more without harming anything. The NGK plugs (BKR6E ) have approximately 100 miles on them and are currently not fouled so I do not think they are responsible for triggering the CEL. I think I will start by making sure I am getting adequate fuel delivery, and then will try some other plugs if the problem doesn't go away.

[bmwdirtracer]

#1 you shouldn't run a stock engine like that. Next try something very simple switch from NGK to Bosh or Champion. I found years ago NGK foul quickly. Are you running white gas? Amoco and BP ultimate are premium gas if you're trying to blow your motor no better gas to do it with. When was the last time you rebuilt the motor? 98 is pretty old. peak performance requires a peak engine. Exhaust? Are you running a performance header system with cross overs? Old fuel has to leave for new gas to enter. Do you have AAA? I have a feeling you'll be in need of a tow truck soon. I'm pretty sure I have the tuning software for that car. buy the adapter and plug it into your laptop remap the system. I'll look and see

The stock Bosch FGR8KQC or NGK BKR6EQUP or BKR6EK are the ONLY plugs that should ever be used on this engine. The BKR6E should be fine too. Champion and Autolite and AC Delco sparkplugs should never be allowed within twenty feet of any BMW.

You found years ago that NGK plugs foul quickly? Were they the CORRECT NGK for the engine involved? BMW and the vast majority of the techs and owners at these forums insist on the correct NGKs or Boschs. BMW has NEVER approved any Champion, and they never will.

And as for "a stock engine shouldn't be run like that"? Oh, Bullcrap. Tell that to all the BMWCCA members who beat living hell out of their stock cars on racetracks across the country.


Saying the OP is going to need a tow truck because he's using the stock , low mileage engine at stock, permitted rpms is absurd.

And then you're going to offer unknown tuning software for an engine that you say shouldn't be run hard with STOCK programming?

Sorry, this is the mechanical forum -- that nonsense doesn't fly here.


PT356, Maui M3Mania is probably the best researcher I've ever known. He's owned your engine, and is very familiar with it. Based on his post, I think it's likely you're bouncing off the rev limiter

[shadowpuck]

chris, i agree - i can't say i've heard of issues with NGK plugs....especially in m5x/s5x engines; it's a factory spec plug that has served me well for many years...
i'd also agree that being aggressive with the engine is fine - at least, in my experience it's been just fine! my m50 was at 208k when i swapped it out for the s50, and the m50 was definitely run at redline a bunch (after 140k, even!) that engine is still alive today in a friend's e36!

admittedly - i had no idea on the rev limiter and gearing - that's interesting....any idea how the ECU knows the gear in use? the inner geek/nerd in me is curious!

[PT356]

As MauiM3Mania has pointed out and Chris is in agreement with perhaps I am hitting the Rev limiter in 3rd gear. The car runs great, feels very smooth with lots of power up to that point. However it does not seem right that if the engine hits the rev limiter that it would generate error codes and shut down one or more cylinders. My current commute is a windy ass road (HWY 17 in Santa Cruz, CA) where in certain places there is no shoulder where I can pull over, turn off the engine and restart to restore all the cylinders. I realize that I probably won't be going 6500 rpm in 3rd gear on my commute home but if it were to happen on an uphill the engine would probably struggle big time forcing me to pull over immediately. When the CEL goes off the engine runs horribly which makes me feel that more than one cylinder is deactivated. What is suppose to happen to the engine when the Rev limiter is hit?

[presterone]

You just replaced crank sensor? did you use a factory sensor? they are pretty damn sensitive to RFI and the harmonics at that RPM could be doing something fucky. I would remove the connector and make sure its clean and seated fully. I just heard of a car with a similar issue and had a loosely fitted CKPS connector. Its worth a shot.

[bmwdirtracer]

You've got a good point here, Phil. Bouncing off the limiter does not usually engage a CEL or cylinder cut off. I'd consider sending your injectors off to RC Engineering for analysis, blueprinting, and re-analysis, computer reports before and after, $25 per injector, one day service. But I'd likely do the fuel pressure test first, on the road, symptom occurring.

I've just ordered a new FP test set, to arrive next week. If you want to PM me with your address and email, I'll loan you my older gauge, with a long enough hose to clip under your wiper. No charge, return it to me when you've done your tests.

You need to read the codes, then clear them, then re-read the first code set. Tell us those results, and we can better point the way.

[PT356]

I decided to order a new pump which I got installed but the terrible weather prevented me from wanting to test the car until this weekend. I took the car up to 6500 in third and I felt the rev limiter kick in (or what I think is the rev limiter) I backed out of the throttle and immediately the CE light came and predictably cut a cylinder off. The stop and restart sequence fixed things. The car felt great and was accelerating hard up to that point. The codes I pulled are as follows from the freeze frame function:
DTCFRZ = PO301 (Cylinder 1 misfire)
FUELSYS1 = CL
FUELSYS2 = CL
LOAD_PCT(%) = 80.4
ECT(F) = 177
SHRTFT1(%) = -21.1
LONGFT1(%) = 2.1
SHRTFT2(%) = 3.1
LONGFT2(%) = 3.1
RPM = 4987
VSS(mph) = 86

Hopefully this will help identify something. I feel somewhat better as last time the codes indicated 4 cylinder misfires. I am thinking I will switch Coilpack #1 with #2 as it is easy to do.

Thanks in advance for any help you might be able to provide.

Phil
1998 BMW M3/4/5
1965 Porsche 356 Vintage racecar

[bmwdirtracer]

It would be very bad, if the computer removed ~21% of the fuel from one bank, at full throttle. In fact, at WOT or thereabouts, the computer shouldn't be adjusting fuel at all. it should be following the stored maps. Not sure why you're seeing closed loop either.

All that said, the -21% fuel trim on bank 1 is very likely to be the issue here. I'd personally scratch the cylinder numbers into the injectors, and then send them off to RC Engineering, for testing and blueprinting and retesting; $25 each. This would seem to be a good baseline to establish. in view of the data you've found.

EDIT: Oh hell; I just reread the former recommendations I made....and I see I gave you the same recommendation, before the latest info/ RC is in California, you know?

I've been wrong, many times, but that's where I'd start.

[ross1]

Are you sure this isn't as simple as a plug "wire" that's marinating in oil down in the plug hole?
Sounds like the freeze frame is catching it on the over-run.
How many misfires needed to set the code?

- - - Updated - - -

[QUOTE=bmwdirtracer;29578930], Not sure why you're seeing closed loop either.

QUOTE]
My first thoughts were that it's catching the scene of the over run.
Are these symptoms a plausible scenario of the DME NOT seeing a WOT signal, as in a failed throttle position sensor?

[PT356]

Ross1 - I do not think I have a valve cover oil leak into the cylinder spark plug hole as I have a new VC gasket but I will check for that....thanks!
I cleared the codes from yesterday's test run and got another CEL on the way to work. The codes were quite a bit different:

DTCFRZ = PO301 (Cylinder 1 misfire)
FUELSYS1 = OL
FUELSYS2 = OL
LOAD_PCT(%) = 79.2
ECT(F) = 156
SHRTFT1(%) = -18.8
LONGFT1(%) = 0.0
SHRTFT2(%) = 18.0
LONGFT2(%) = 0.8
RPM = 5005
VSS(mph) = 64

Of note was I am showing Open Loop and the SHRRFT2 is @ 18.0. Also for both occasions the codes are being generated around 5K RPM and not at redline.

Does this provided any new information that might point to something other than the fuel injectors? I plan to go ahead and send the injectors to RC Engineering for testing/calibration.

Any additional input is appreciated.

Thanks!
Phil
1998 BMW M3/4/5
1965 Porsche 356 Vintage racecar

[bmwdirtracer]

ECT is probably too low for closed loop operation; I believe the threshold will be ~165-175F.

Short term fuel trim shouldn't really be highly involved, at these rpms....and again, not in open loop. Yet, we see the DME pulling fuel very heavily from Bank 1 in both instances....and dumping the same amount, extra, into bank 2, in the last example.

(Ross, I don't tend to think of this as overrun - both freeze frames show 80% load). But Phil; note that both occasions also happened at ~5000 rpm....not at max revs...)

So I'm going back and reviewing this thread -- please correct me or fill in where I've missed:


At first, your misfires happened near redline, but you got codes for cylinder misfires on 2 cylinders on each bank.
You've already replaced lifters, coils, plugs, and injectors. (You don't need to send new injectors to RC, assuming you bought quality OEM parts....I missed the new injectors, previously.}
And you have a new fuel pump and filter (and I'm pretty sure the regulator is part of the filter, in your car). So, a fuel pressure test when symptoms are manifesting is still officially needed, but unlikely, since both banks are doing different things...at least at the code-setting. [That said, this must remain , officially, a question mark until proven]

And now, under closer scrutiny, and after your many replacements, the misfires SEEM to only happen on number 1 cylinder.

So we should probably get beyond "freeze frame", and get live data.

You've done a leak-down test, which was fine.

Have you looked at a recent complex "misfire" thread for my own M5? It's called "DME Repair?", in this forum.

Can you watch live data, and do a misfire count, with your scan-tool? (As you'll see in my thread, I was watching ~600-800 misfires an hour, with no codes) If all six cylinders are misfiring, we might be led to different conclusions than if we're really now searching for just 1 cylinder)

Do you have any ABS codes or issues?

Get a can of electrical contact cleaner; clean both sides of all the DME plugs. Check (unbolt and clean) the engine main ground (EDIT: and also any large ground point near the DME box)

[ross1]

. Yet, we see the DME pulling fuel very heavily from Bank 1 in both instances....and dumping the same amount, extra, into bank 2, in the last example.
)

Could bank 1& 2 sensors have the connectors exchanged?
No mention of them being serviced by the OP but he does mention that the car is new to him.

[328 Power 04]

Could bank 1& 2 sensors have the connectors exchanged?
No mention of them being serviced by the OP but he does mention that the car is new to him.

You are quite possibly onto something.

Front connector goes to bank 2 (cyl 4 5 6), rear connector goes to bank 1 (cyl 1 2 3) if I remember correctly.

- - - Updated - - -

But if this is at WOT, then it wouldn't matter, as WOT = open loop, no lambda regulation.

[ross1]

You are quite possibly onto something.

Front connector goes to bank 2 (cyl 4 5 6), rear connector goes to bank 1 (cyl 1 2 3) if I remember correctly.

- - - Updated - - -

But if this is at WOT, then it wouldn't matter, as WOT = open loop, no lambda regulation.

The freeze frame indicates it's in closed loop and the amounts are same, hence my thoughts.
I had similar scenario with a Jaguar after a trans job, the R&R man had swapped the O2 connectors. That car, however, ran like TOTAL crap.

[bmwdirtracer]

One freeze frame is open loop, the other closed. I seem to recall that this engine uses 5-wire primary wide band sensors, and sets codes for swapped sensors, when that happens.

And, one freeze frame shows opposite values for short term fuel trim, the other doesn't...

I absolutely see where you've come up with that likelihood, Ross, but I think we need live data, to see whether there's a possibility of this.

Phil, you need to unplug one of the primary O2 sensors, then look at data, and make sure that the correct bank's sensor shows up with no data.

Can you find a way to look at live data, and use that to count misfires, Phil?


Oops, I see that Phil, you've already seen the misfires aren't at max revs, sorry I missed that, it's hard to keep up, on these extended threads.

[pbonsalb]

Misconnect the 02 and the car runs like crap pretty fast. Been there done that. Not sure what happens if you misconnect knock sensors on obd1 or put 1 into 2's location on obd2.

[ross1]

Any resolution yet?

[PT356]

First I want to thank everyone who has contributed to me finding a solution to my problem, particularly Chris! It is a lot easier to move forward on this when I do not feel like I am tackling this alone. Been very busy and since the M3 is an extra car for me combined with the wet weather I haven't spent much time on the car. I hope for there to be resolution soon though! Although this car is in very good shape it lived its life in the Chicago area so the fasteners on under carriage have a fair amount of corrosion. Per Chris' advice I plan to clean the DME connectors and the clean up the ground. I also plan to unplug one O2 sensor to determine if these got swapped somehow. My OBD reader has Live data capability so obtaining real time data might help as well. I did look at some older OBD data which I had recorded which also showed the that the CEL was being triggered around 5K RPM and not 6.5K. This would indicate that there is a delay occurring from when issues are being detected and CEL lights up. Hopefully this is something the live data will show. Will post again when I have more to share.

[bmwdirtracer]

Just for your info, Phil, you've got the attention of some of the very best BMW guys I know. Pbonsalb, Ross, Abel, shadowpuck, MauiM3Mania.....damn you have got some great people on your side.

[tmswell]

Great to meet you this week--enjoy the springs
As we chatted about, the curveball idea might be to check for a bad cat. I'll come up with more ideas if that's a strike-and-a-miss.

-tmswell