The solution to the stumble/yo yo syndromes are neither software nor hardware, but a combination thereof. For software to control a process, that process must be able to be modeled. If the process is erratic it can’t be modeled accurately, and therefore can’t be controlled accurately. Apparently the Cooper S fuel/air/ignition system is erratic at some system rates, hence the stumble/yo yo. The source of the problem is hardware. The fault could be with an individual piece of hardware, or several pieces of hardware, or the interface of several pieces of hardware. Once the offending piece, or pieces, of hardware is/are identified and a system compatible replacement, or replacements, is/are found software can be easily written to control the system.
What has been portrayed as a simple solution is actually quite complicated, hence no quick fix. Makes sense to me.

 

Forgive me for sounding abrupt but... how do you know this? Is it your own personal theory or have you spoken with someone official?

 

...or perhaps he's portraying his hypothesis as fact? ops:

 

I posted what I believe may be a 'workaround' to the dreaded stumble at the end of the original Stumble thread in Performance Modifications. I am wondering if anyone would be willing to try it to see if the results are similar?

Raven

 

The basis of my theory or hypothesis, if you want to call it that, is a fact of every non-linear dynamic hardware/software process control system. MINI has demonstrated it does not have a simple ‘write a few lines of code’ solution. Additionally, posts seem to indicate the severity of the problem changes with time. Hardware dynamics change with wear, software does not. This does not mean they will not find a driver suitable software compromise, but a maximum power/driveability solution will only be found with a hardware/software change.
Still makes sense to me.

 

Keep in mind an additional variable. BMW/MINI may have an easy fix for the stumble, but may not be able to implement it AND keep the MINI within the original EPA mileage estimates.

Just a thought,
JS

 

JS - I appreciate your constructive comment, however, from a practical standpoint a fix that can't be implemented is not a fix.

 

I just thought it stumbles because one day gimlok told me he was thirsty so I poured a bunch of bass through the hood scoop

 

crowstone -- Welcome to MCO. This seems like sound logic and may, in fact turn out to be the real story, as you say.

I don't wish to discourage your contribution but can you tell us a little about your background as it informs this reasoning, please? One of the human dynamics of a forum like this is that you need to establish some credibility before you begin offering advice and solutions. We're pretty open to new thought here but this particular, irritating problem has been discussed up one side and down the other.

 

Ah, but no.

You see, many of us had NO problems until we were updated to a newer version of the software. Then and only then did the stumble rear it's ugly head. So to imply it's mainly a hardware problem coupled with software not adequately controling said hardware is completely untrue.

R

 

As Davbret points out, the stumble is 100% software-related. Many of us drove 10,000 or more miles without the stumble only to have it 'installed' by the dealer with the version 33.2 'upgrade' (and I use this term loosely). This problem is not in the clutch, not in the manifold design or any other complex design-related issue. For those of us who have written engine management software, it is unbelievable that Mini doesn't fix this immediately. Furthermore, the Customer Service Division is unresponsive and sluggish at returning calls. I call once a month and leave a message with a request for a return call from a zone rep and have NEVER gotten a call from the zone rep. Frankly, I'm starting to get pissed. Real pissed.

PerfPow :evil:

 

crowstone....nice thought on the subject, it makes sense to me...my MCS has always stumbled from the start and I have had every ECU upgrade including the last one a couple of weeks ago and the difference is nada, not one iota difference in the low end stumble, I've always attributed it to the low torque of the pentagon engine, especially below 2500 rpms, as above 3000 rpms, no problem....maybe it's not as complicated as the entire engine, but it certainly appears to me far more complicated than just a ECU software issue....thanks for the input and a different take on the issue...it's like Einstein's famous quote..."insanity is trying the same thing over and over and expecting different results".....maybe trying something other than just a software fix might be a start...cheers !

 

I respectfully disagree with the originator of this thread. Ever since I got updates 32.1 and now 33.2 in our 2002 Cooper CVT, the car has been running smooth as silk, which is clear proof that most engine related stumbles and rough idle problems, among others, do originate from the controlling software code used by the Siemens engine management system in the MINI.

If the problems were both, hardware and software related, the stumbling and hesitation would be happening to everyone accross the board, whether it be a 5-speed Cooper, a MCS or even a CVT.

This is the classic case of "One man's poison is another man's medicine". Our Cooper CVT has been working fantastic with the 33.2 upgrade, while many MCS owners have been having a plethora of problems with the same exact software update.


While your theory is interesting, does not hold much water in this case.

Anyhow, welcome to the forum. BTW, do you own a MINI? :smile:

 

Perfpow, Davbret,

Since you guys HAD working cars before the upgrade, are there other differences other than the stumble/hiccup? Does the car feel less powerful in low gearing after you had the upgrade?

I'm just wondering what the car will be like when it's right, since mine has been a dog performance-wise since I took ownership.

See ya,
JS

 

In my opinion a critique of an analysis should be based on the facts and not the resume of the analyst. However, I am a mechanical engineer…began my career doing independent race team R&D for automobiles, motorcycles and snow machines in the early 1970’s…great fun…poor money. Moved on to the good money in the oil and gas industry…with the oil downturn and decreasing money flow, I transitioned into the commercial construction industry…I am currently a problem project construction management person…negotiating a possible move to the good money in Iraq…enjoyed the good Saudi money during Desert Storm.
I drive a Sept build 03 S. No 2000 rpm stumble, but I have the flat low end and cold start problems…cold start began after approximately 8500 miles…since the software didn’t change one must conclude the change is due to hardware…I believe the fix for these problems was anticipated with the latest round of software, however, and I paraphrase other board contributors, ‘this did not fix the problem, it only moved it up the power band and made it worse’.
The variance of problem severity within a software write also points to hardware…the software doesn’t change so the only thing left is hardware.
I would consider the software in my S to be a compromise fix for the 2000 rpm stumble…doesn’t fix the problem but makes the car more drivable…
All this said it doesn’t mean they won’t find an acceptable software solution…within the problem rpm range they may increase/decrease sampling rates, or increase/decrease dampening rates, etc. etc. and smooth out the hardware anomaly causing the problem…it will however be a performance compromise…the existence of the initial problems indicates they are having a difficult time modeling the low rpm range…unpredictable low rpm hardware interface…the fact that they introduced the huge 2000 rpm stumble with software indicates someone involved is related to the folks that brought us Lucas Electrics.
The yo yo is a different problem altogether but still hardware predictability related.
Still makes sense to me.

 

There are no LUCAS components in the new BMW MINI. In fact the engine management comouter system is being provided by the German folks at Siemens.

 

>>In my opinion a critique of an analysis should be based on the facts and not the resume of the analyst. However, I am a mechanical engineer…began my career doing independent race team R&D for automobiles, motorcycles and snow machines in the early 1970’s…great fun…poor money. Moved on to the good money in the oil and gas industry…with the oil downturn and decreasing money flow, I transitioned into the commercial construction industry…I am currently a problem project construction management person…negotiating a possible move to the good money in Iraq…enjoyed the good Saudi money during Desert Storm.
>>I drive a Sept build 03 S. No 2000 rpm stumble, but I have the flat low end and cold start problems…cold start began after approximately 8500 miles…since the software didn’t change one must conclude the change is due to hardware…I believe the fix for these problems was anticipated with the latest round of software, however, and I paraphrase other board contributors, ‘this did not fix the problem, it only moved it up the power band and made it worse’.
>>The variance of problem severity within a software write also points to hardware…the software doesn’t change so the only thing left is hardware.
>>I would consider the software in my S to be a compromise fix for the 2000 rpm stumble…doesn’t fix the problem but makes the car more drivable…
>>All this said it doesn’t mean they won’t find an acceptable software solution…within the problem rpm range they may increase/decrease sampling rates, or increase/decrease dampening rates, etc. etc. and smooth out the hardware anomaly causing the problem…it will however be a performance compromise…the existence of the initial problems indicates they are having a difficult time modeling the low rpm range…unpredictable low rpm hardware interface…the fact that they introduced the huge 2000 rpm stumble with software indicates someone involved is related to the folks that brought us Lucas Electrics.
>>The yo yo is a different problem altogether but still hardware predictability related.
>>Still makes sense to me.
>>


Nope, it doesn't make sense to me.

 

I have a feeling that the software might have compromised some hardware i.e. requested the hardware to perform over it's limit. It would make sense since the stumble-yoyo didn't happen right away after the upgrade, but in about 50 miles or so, and it's getting worse. I have to go with crowstone on this one.

 

>>I have a feeling that the software might have compromised some hardware i.e. requested the hardware to perform over it's limit. It would make sense since the stumble-yoyo didn't happen right away after the upgrade, but in about 50 miles or so, and it's getting worse. I have to go with crowstone on this one.


I am still not convinced. How come 33.2 works wonderfully in my Cooper CVT, yet it performs horribly in your MCS?

If the problems were of hardware origin, everybody, I repeat, everybody would be having stumbles regardless of model being driven and how many software updates were done to the DME.

My car had stumbles and very rough idle quality prior to update 32.1. Since then the car completely changed for the better and now runs much better than it did when we first took delivery nearly 11 months ago.

I said this again...interesting theory but doesn't hold water.

 

In the presence of gearheads...

Ok, I'm not a total moron when it comes to cars, but I am so far below the understanding of all y'all that its almost laughable. However I think, as an unbiased layman, I can make some observations, be they relevant or whatever...

Someone asked already, and I'll ask again, if it's a hardware issue, why AREN'T all MINIs everywhere suffering the same fate? For every one person with low rev power issues, cold start and stumble, there are probably 5 people who have MCs and MCSs that run brilliantly.

I would also think that a hardware problem would be more readily identifiable by engineers and techs, no? I mean, you can generally see, hear or feel a hardware problem, particularly if it is something that has occurred as a result of the software asking too much of it (as someone suggested) and you have an understanding of how the car was supposed to run. In any event, wouldn't it be simpler to determine the fundamental issue if it was hardware? Any other time some piece of hardware craps out you just take it to the shop and the tech tells you what needs replaced.

?

 

OmToast,

This is a Cooper S problem…the fact that Coopers may run without problems is not relevant…their hardware system are different…you have however hit upon one of my points…the variation in performance problems between cars of the same type with the same software release…the software is identical between individual cars, but the hardware is not…every throttle body, sensor, etc. they produce is different…the differences may be small but they are different.
Controlling a dynamic process like an automobile engine is very difficult…each piece of hardware may function within design parameters when tested individually, however, when those individual pieces are ask to act in unison to control a process their individual idiosyncrasies may combine to produce a process error which is not predictable and therefore can not be modeled accurately…in the case of the Cooper S a flat low end, stumble, yo yo, etc…now try to fix the problem…every piece of hardware functions within design parameters…how do chose which piece of hardware to change and how do you change it to produce a predictable system which can be accurately modeled and therefore computer controlled…

 

Crowstone posits:

>>The variance of problem severity within a software write also points to hardware…the software doesn’t change so the only thing left is hardware.

Would fuel be included in that definition of hardware?

 

You are right, oh originator of this lengthly thread, software does not change (unless edited/upgraded). However the way software behaves can change over time if it is programmed to do so. No one here is a vehicle software programmer, so unless we find out for certain, no one can say if the MINI software is designed to change (or learn as some may say) after certain occurances such as the passage of a particular amount of time or mileage, or under certain temperatures or other variable inputs.

 

>>OmToast,
>>
>>This is a Cooper S problem…the fact that Coopers may run without problems is not relevant…their hardware system are different…you have however hit upon one of my points…the variation in performance problems between cars of the same type with the same software release…the software is identical between individual cars, but the hardware is not…every throttle body, sensor, etc. they produce is different…the differences may be small but they are different.
>>Controlling a dynamic process like an automobile engine is very difficult…each piece of hardware may function within design parameters when tested individually, however, when those individual pieces are ask to act in unison to control a process their individual idiosyncrasies may combine to produce a process error which is not predictable and therefore can not be modeled accurately…in the case of the Cooper S a flat low end, stumble, yo yo, etc…now try to fix the problem…every piece of hardware functions within design parameters…how do chose which piece of hardware to change and how do you change it to produce a predictable system which can be accurately modeled and therefore computer controlled…
>>


You are beginning to make more sense, but don't forget that there are Cooper S owners that have not experienced the same problems as many others with the stmbling, yo-yo, etc.

Fuel quality is definetely yet another variable as its quality its pretty inconsistent from gas station to gas station and accross state lines.

I still don't see this as the by product of hardware design issues. Many people have driven their MCS for 10K-15K with no problems and suddenly after update 33.2 they started experiencing all the well documented issues.

For instance, several owners (Cooper and MCS) have experienced premature failure of the throttle body control module and the engine wiring harness and that is a problem traceable all the way back to a bad batch of these components. I don't see the yo-yo and stumbling in the same light.

 

Nuvolari,

Fuel would be considered hardware...one of the timing parameters is fuel...could be contributing to the problem but isn't, in and of it's self, a solution...if fuel was a solution the members of MCO would be aware of it.