strange fueling behavior... BMW 325i E36 + MS3

[JoseMiguel]

I have been tracing a problem for more than 3 weeks and im in a point where I dont know where else to go or what else to look...

Car is a BMW E36, 2.5 engine, turbocharged with a clean MS3 installation. As some of you know im an experienced MS3'er, so I can guarantee you that everything is wired just like in the manual.

Im using the stock vr sensor with the 60-2 wheel, the cam sensor which gives one pulse per 720, injectors wired for sequential, stock COPS wired using two quadspark modules (each module using 3 channels). RPM capture is perfect, no lost sync, cam also gives a good, clean signal (im using the cam input in ms3x with the pullup jumper which I thought was strange considering everyone told me that the sensor was a VR sensor). Everything works great....

this is exactly the same engine with exactly the same coils



Car fires right up, good clean startup. Car still has the idle valve working with the stock computer and it works well. Idle is good, acceleration is clean and strong, except for one very short issue: when the car sees anything over 5-6 psi, wideband shows a stupidly lean AFR but it does not fail (I lift because im afraid of killing the engine).

My first thought was a faulty wideband. I tried 2 more widebands and they show exactly the same behavior. It is strongly showing low 13, high 12 AFR on acceleration and when boost starts to build, it inmediately goes to 18-20 AFR numbers, but no hiccups (it sounds like a cylinder or two are not firing) until I get scared and lift. I tried stupidly high VE values on the boosted part of the map with no luck. More fuel, same wideband reading and when I try stupidly high values it begins to hestitate which makes me think that the car is insanely rich (but the 3 widebands I have tried show lean)....

Then I went with the fueling part. I tried a brand new bosch 044 fuel pump, same issues, I tried more than 4 pumps and same behavior. I also tried three more set of injectors (55lb, 44lb, 36lb and 22lb) and all of them show exactly the same problem. I have tried a pressure dial in the rail and pressure is strong when the car goes full lean.

I have tried all sort of stupid and non stupid stuff, I even wired the injector and coil feeds directly from the battery with high caliber gauge wire and no change. Only thing I have not tried is changing the coils, the ECU or the quadspark modules. I have triple checked the grounding. No luck.

On the tuning part I have also tried all the known and unknown tricks. I have tried batch fueling, wasted spark and ignoring the cam sensor completely. I tried playing with the dwell, injector timing, nothing... exactly the same behavior each time I have tried.

At this point, im completely lost, I really dont know what else to do. Ideas on what to do are welcome. Also, if somebody has worked before with this coils, comments and ideas are welcome too. In my opinion, spark seems a bit weak, but I dont think that is the cause of the problem.

Im attaching the MSQ and a few datalogs. The fueling map on the MSQ im attaching is not very elaborate as I have been modifying it a lot trying to trace the problem with no luck. Also, please remember im at a high altitude, for this car, boost is past 76 kpa...

[93supercoupe]

i had an almost exact situation with my first MS2. It ended up being that the Accel Enrichment was set for 50 MAP and 50% TPS. The MS was seeing minute fluctuations when in boost and was fueling accordingly.

So i set it to 100% TPS and she was good.

[JoseMiguel]

i had an almost exact situation with my first MS2. It ended up being that the Accel Enrichment was set for 50 MAP and 50% TPS. The MS was seeing minute fluctuations when in boost and was fueling accordingly.

So i set it to 100% TPS and she was good.

Thanks for the answer

Unfortunately, my accel is set 100% tps and there is no sign in the datalogs of AE kicking in where it should not

[nuvolarossa]

Don't know if this will solve the issue, but you must ramp up dwell. Carefull as they don't like too much of it, they'll heat etc.

I know some and also read online that many are using around 2.6/2.7ms on the m50 coils doings LOTS of turbo power (700+ whp).
Go for 2.7 and you'll be good.

here it's measured so you know they can take more (but don't know voltage of their measurements): http://forums.bimmerforums.com/forum/sh ... st24806273

--- Why the stock ecu is still wired? idle valve on stock ecu?
why that, are there special reasons/other things stock ecu is handling?

(also afr not showing on log)

[JoseMiguel]

Don't know if this will solve the issue, but you must ramp up dwell. Carefull as they don't like too much of it, they'll heat etc.

I know some and also read online that many are using around 2.6/2.7ms on the m50 coils doings LOTS of turbo power (700+ whp).
Go for 2.7 and you'll be good.

here it's measured so you know they can take more (but don't know voltage of their measurements): http://forums.bimmerforums.com/forum/sh ... st24806273

--- Why the stock ecu is still wired? idle valve on stock ecu?
why that, are there special reasons/other things stock ecu is handling?

(also afr not showing on log)

Thanks for the answer and from the help...

We are keeping the stock ecu as it controls a small computer in the dash and the climate control. But we already tried completely disconnecting it. Exact same issue, no matter if it is connected or disconnected..

I have tried dwell levels as low as 0.5 and as high as 5 with no difference whatsoever....

Do you have any more info on this coils? I have searched a lot and I get contradictory info. An experienced person in here tells me that those coils are logic level and should use dwell levels as high as 4. Some other forums say that this need external ignitors to work. Im using them with quadspark modules and they work but spark does not seem very strong and dwell seems to make no difference.

[93supercoupe]

the BMW coils are 100% NOT logic coils.

What interface dsoes the stock DME have with the OBC and climate control? I didnt realize they had any integration in the e36's. And whatever it connection they do have, im sure the MS can take care of it.

I just see alot of potential issues by keeping both computers. I could be wrong.

[jamies]

stock COPS wired using two quadspark modules (each module using 3 channels).

if they are similar to the VAG coils you will need to drive them with a PNP driver

[JoseMiguel]

the BMW coils are 100% NOT logic coils.

What interface dsoes the stock DME have with the OBC and climate control? I didnt realize they had any integration in the e36's. And whatever it connection they do have, im sure the MS can take care of it.

I just see alot of potential issues by keeping both computers. I could be wrong.

thanks for the help

At first I thought that the stock computer could do no harm. After being unable to find the problem, we tried everything, including disconnecting completely the stock computer. The issue remains the same with or without the stock ECU in place!

[JoseMiguel]

stock COPS wired using two quadspark modules (each module using 3 channels).

if they are similar to the VAG coils you will need to drive them with a PNP driver

as far as I know (I could be wrong) they are not like VAG cops. Although I have contradictory info on the coils, they seem to be standard dumb coils.

[93supercoupe]

because you have changed so much and have not effected the issue at all, im inclined to say its in the tune somewhere.

Just out of curiosity,do you have a picture of how the stock DME and the megasquirt are hooked up at the same time?

[JoseMiguel]

I am going to try something today that I have not tried before

The BMW coils have 3 pins. Voltage, Signal and Ground

We tapped the voltage and signal, but we have done nothing with the coil grounds, they are grounded as stock! this has not been touched

I am going to try and ground the coils directy to the block and see if that helps!

[93supercoupe]

our situations are different in alot of ways, but im also dealing with the coil grounding right now. I dont know if that will fix either of our issues, but its probably a good idea to get it sorted out.

Im going to be running pins 15 and 45 to a single mainboard ground, and pin 28 (crank sensor sheilding) to the crank sensor metal strand shielding.

http://www.msextra.com/forums/viewtopic ... 31&t=55620

[JoseMiguel]

Grounding the coils to a "better" ground did nothing for the problem

Next thing im doing is to fight with the FPR... although it seems strange that there is no pressure drop whatsoever when the car goes stupidly lean so my guess is that there is no pump problems and no FPR problems!

[93supercoupe]

i really think its a setting more then mechanical. Have you checked all ur ignition comps, fuel comps, accel enrichments, over boost, ect. ?

[JoseMiguel]

i really think its a setting more then mechanical. Have you checked all ur ignition comps, fuel comps, accel enrichments, over boost, ect. ?

unfortunately yes, I have checked everything.

[grom_e30]

is the fpr connected to the manifold? and able to see vacuum and boost?

[JoseMiguel]

is the fpr connected to the manifold? and able to see vacuum and boost?

I just checked this and yes, it does...

and only to be sure, I decided to tee the signal from the MAP hose so that im completely sure that the fpr is seeing what the MAP sees!

[grom_e30]

hmm, can you get a log with the ecu connected up to the wideband? so we can see what you are seeingh?

[JoseMiguel]

hmm, can you get a log with the ecu connected up to the wideband? so we can see what you are seeingh?

here is the log, freshly taken a few minutes ago



as you can see, the AFR is perfect before 4000RPM and 120KPA, steady 12.3 - 12.1.... When it reaches those KPA, it abrutply goes lean. The log shows 14.7 AFR in there but im guessing it is some kind of limitation with the logging of this WB module (the display shows 18...20AFR in there), but you get the idea...

im attaching the complete datalog

[grom_e30]

spark blow out maybe causing it to read lean?