So, my MCS (like many others) has decided that it doesn't like to start on the first try. This problem has appeared off and on over the past almost 3 years of ownership, but with a few consistent patterns:

1) It's always in the Summer or early Fall, never other seasons.
2) It's always on the first start of the morning. I don't encounter this problem at lunchtime or in the evening.

Other than that, the problem hasn't pinned itself down to brand of gas, mods present, or version of software (V.32 and V.36 both have done it, with stock, MTH, and GIAC programs). I really wouldn't expect aftermarket chips to change anything related to startup behavior anyway.

I hooked up our prototype BiM-COM logger this morning to try to document the event. I logged as many variables as possible, at the expense of sample rate, but what I found is that ignition timing is 27 degrees BTDC while cranking. That stayed the same for each cylinder, both in the NO-START and in the YES-START scenario. Tomorrow, I will try logging with fewer variables, specifically, throttle angle and injector pulse widths to see if I can catch variations between the two scenarios.

 

how is your fuel pressure when it won't start?

 

No sensors for fuel pressure, unfortunately.

 

a while back i had the same problem...advice was given to me to put the key in the ignition, let it sit in the ON position before starting for a few seconds, let the fuel pump do its thing. It always strted up quickly after that. So maybe it is fuel pressure

 

I tried doing that the last few days, but with no change in behavior. In the morning, my MCS will not start unless I turn the key ON, turn it to START to crank it, turn it OFF, then turn it to ON and START again. It then fires right up. Here are the variables I logged today:

 

Funny you should say that because Andy was a person who said to do this only a couple months ago. I also had someone tell me to try this almost 3 years ago and it never change over the years.

It happens at random on cold days, hot days, winter, fall, summer, spring, but only at random. When it does happen it is only in the morning after sitting for the night, but never at night even if it has sat for two days without being started.

It only happens after midnight - sounds spooky kids, but it's true.

 

Yep, that scheme didn't seem to work over time.

 

I did some testing this evening, logging various stuff while (successfully of course) starting the engine.

According to the logs:
It only takes about 1/2 second to start the engine.
The engine calls for 11.49 ms of injector time during the start.
After it is started, the MIL stays on for about 4 seconds.
After the engine is running, injector pulse widths drop to about 2 ms.
It takes over a second to shut off the engine.

More research tomorrow.

 

I suffer this problem almost every day, especially in the warmer months. First try stalls out, second try starts up just fine.

Believe it or not, my dealership's head tech told me to try using "regular" unleaded (87 octane), versus the 91 octane super unleaded that my MCS's manual demands.

NO, I have not tried 87 octane gas in my car, and refuse to, even if the prices keep climbing thanks to Hurricane Katrina. I'm inclined to trust the engineers who created this car when they demand 91 octane, versus one dealership who suggests using 87 octane gas is the only "solution" for this problem.

Has anyone else been told this by their dealer? Does this suggestion have any merit?

BTW, my dealer is Towne BMW/MINI in Buffalo, NY...

 

ive been using 93 octane...thats whats available here

 

The fuel pump will stop running after a second or 2 if you don't crank the engine. Try turning the key on, off, on, then crank. If that works then install a fuel press gage at the end of the fuel rail & see if the press drops overnight.(50psi. norm.) May be the fuel press. reg. or fuel pump leaking slightly. A falty coolant temp sensor is another possibility (leaving the mix. too lean.) Winter & spring it starts fine? Not the coolant sensor then.

 

Andy,

Great investigation there!

Once you'll get the exact problem's source, do you thing the cold-start problem could be solved by an aftermarket ECU software (MTH, Giac, Unichip,...) ??

 

That makes zero difference on mine.

 

Yes some clueless dealers have said that, in a lame-assed attempt to just get the customer to go away. It has nothing to do with it.

 

No. It happened on mine when it was 100% stock, as many others have also reported.

These theories have all been gone through already on the other huge thread about it. Just wait until Andy finds something new, there's no point in thrashing through all the old stuff again and again.

 

I too had this problem in the first few weeks of my 05 Hardtop, and have always taught myself and my wife to turn the key to the on position for a few seconds, and it always fires flawlessly. On her Drop Top, when we got it, we did the same from the beginning, and never have had an ignition problem thus far. Just for kicks I tried starting it w/o letting it sit in the ON position and get the sputter and immediate dying.. I know Andy is the KING of small problems like this, so if this hasnt rectified his problem yet, he will soon have an IGS (Ignition Gain System) Mod for us to use...

 

Okay, here are some preliminary findings. Thankfully, my engine didn't start on the first try this morning.

There is one notable difference in my logs between the Fail and Pass start attempts. There is a value called " Air Conditioning Compressor Relay Active". It can be either 0 (not active) or 1 (active). During the Fail start attempt, the value stayed at 1. During the Pass start attempt, the value was 0. It looks like it then went to 1 several seconds after the start. I'll have to do more testing to see if this is consistent. If the A/C compressor were engaging while the engine is trying to start, I could see that potentially creating problems. Now, why the value and behavior are different between the first and second starts of the day is unclear.

Here's a graph showing RPM and injector pulse widths for the Fail and Pass conditions:

 

That is certainly consistant with the feel of the problem. Why it would cycle during a start, and why only at certain periods is quite the mystery.

 

I have found my rig will start up with dsc disabled if I start the engine too quickly, rather than key 0n-hesitate 2-3 seconds-start. Implying it may take the ecu a moment to get its shitt together before it starts running things.

 

GBMINI#3 definitely responded to using 91 instead of 93.

See here: http://www.gbmini.net/mtblog/archive...ld_start.shtml

I tried 91 for two tanks, and eliminated the "cold start".
Then I switched back to 93 - back came the problem.
Since then I have run 91 always and never had the "cold start" problem - except for one start this last weekend, which was just after a fill-up which maybe was not-so-91 as claimed ...

Alternatively, perhaps the problem will recur when the weather gets cooler - that's what GBMINI#2 always used to do

Andy - anything you can discover will be great.

 

My 2002 MCS NEVER started on the first try for over 40k miles no matter the temperature or the gas octane. Then I got the MTH flash and it has never NOT started on the FIRST try since and I now have 62k miles.

 

Ah well, the "Air Conditioning Compressor Relay Active" was off during both the Fail and Pass starts this morning. More research needed.

BTW, here's the stuff I logged (out of a possible 150+ variables):

Marker
TimeStamp
Battery Voltage
Vehicle Speed
Engine Speed
Transmission Output Shaft Speed
Manual Transmission Gear Ratio
Auto Transmission Gear Ratio
Canister Purge Duty Ratio
Calculated Load
TimeStamp
Air Temperature Error Active
Coolant Temperature Error Active
Camshaft Signal Error Active
Camshaft Limp Home Mode Active
Cam Falling Edge Active
Crankshaft Signal Error Active
Crankshaft Limp Home Mode Active
MAP Signal Error Active
MAP Limp Home Mode Active
Throttle Position 1 Error Active
Throttle Position 2 Error Active
Injector A Error Active
Injector B Error Active
Injector C Error Active
Injector D Error Active
Driver Demand 1 Error Active
Driver Demand 2 Error Active
Motorised Throttle Error Active
Motorised Throttle Controller Error Active
Aircon Presure Transducer Error Active
LDP Error Active
Upstream MAP Error Active
Cannister Purge Solenoid Error Active
Fuel Pump Relay Error Active
Incoming CAN ABS Frame Error Active
Incoming CAN EGS Frame Error Active
Incoming CAN INSTR Frame Error Active
CAN Protocol Error Active
Fuel Pump Relay Activated
Engine After Start Phase Active
Idle Phase Active
Engine Part Load Phase Active
Engine Trailing Throttle Phase Active
Engine Trailing Throttle Cut-Off Phase Active
Engine Full Load Phase Active
Closed Throttle Active
Main Relay Activated
Brake Switch 1 Active
Brake Switch 2 Active
Clutch Switch Active
EVAPS Pressure Switch Active
PRNDM Fault Flag Active
CVT Vehicle Speed Error Active
CVT Ratio Plausibility Error Active
Transmission Oil Temperature Error Active
CVT Shift In Progress
ASR Request Authorised
MSR Request Authorised
Clutch Locked Active
TimeStamp
MIL Active
Air Conditioning Relay Activated
Cooling Fan Low Speed Active
Cooling Fan Medium Speed Active
Cooling Fan High Speed Active
Ignition Key Is On
Air Conditioning Request Active
Air Conditioning Compressor Relay Active
Air Conditioning Fan Low Speed Request Active
Rich Mixture Detected On Upstream Oxygen 1
Closed Loop Enable Active
Upstream Oxygen 1 Ready For Air/Fuel Regulation
Downstream Operability Active
Engine Is Running
Engine Is Synchronised
Engine Is Stopped
Engine Start Phase Active
Torque Increase Flag Active
Torque Reduction Flag Active
Throtte Adaptation Successful

 

Some BMW (or ex-Rover) engineer is reading this and laughing his *** off right now...you know this.

 

This morning, I tried another test. Prior to starting the car, I checked and cleared fault codes (even though none were present). This should reset some adaptations (additive and multiplicative fuel trim, maybe others). It had no effect at all. I'll try clearing each of the specific adaptations:

Idle Speed Adaptation
Misfire Adaptation
Knock Spark Adaptation
ECT Throttle Adaptation
Lambda Adaptation
Knock Control Adaptation
MAP Adaptation
Dynamic Trim Adaptation

Too bad this problem only manifests itself once a day with horrible regularity, so I can try clearing them one at a time (which will take up to 8 days) or just go hog wild and do all of them at once (which won't narrow down the problem if it proves to be a "cure").

 

EVOtech software hasn't cured the problem yet either as it's an underlying base software problem.

The dealership here in Portland is recommending that effected parties email them a really nasty letter so they can forward it on to BMWNA. (No, I am not kidding)