[nonsequitur]

To capture this in our thread on objective information, here is a site that is apparently sponsored by MINIUSA where you can buy technical details by the piece (with PayPal):

http://www.minitechinfo.com

My OS and Browser combination are blocked by the site, so I can't check it out. It would seem that if we had solid, clear, objective testing and performance information on the CVT, we might be able to push some serious progress into the quest for understanding.

Can anyone check this out? If it's worth its own thread, then start one on MINI2 and let us know here as a note. Keep this focused on the CVT.

[nonsequitur]

Tooling around tonight with our new 33.2 software. At under 1600rpm, there are no behavior differences between D, SD, or Steptronic that I can see - none at all.

If true, then this may be the "Fast Off" or "Low" program mentioned in several articles I've seen. It might also explain the Launch we experience in the cars?

[nonsequitur]

A user in Owner's Lounge on MINIUSA.com just responded that www.minitechinfo.com does not currently have MINI information? I couldn't get in, but that seems weird under that URL?

On Objective Testing:
I now have an old-model G-Tech Pro, a Cooper CVT, AND the CD#33.2 just installed this week, so I am seeking a time and place to Objectively Test each of these cars on the same road, same ambient temp, same conditions.

Anyone here in Puget Sound want to help?

What parameters should be tested?
0-60mph time
1/4mile time and speed
60-0mph distance
Some "rolling start" comparable (1/4mile from 20mph?)
Horsepower at the wheels

What are the control limits for each test?

MINI One Manual (don't know of One in the states?)
MINI Cooper Manual
MINI Cooper CVT in D
MINI Cooper CVT in SD
MINI Cooper CVT in Steptronic
MINI Cooper S Manual
MINI Cooper JCW Manual

I will also make runs with different wheels and tires (but we'll need to swap around to get the different combinations to work).
15x5.5"
16x6.5"
17x7.0"

Can you give any guidance on the process, any suggestions to make it useful, any help to do it?

[RobFriedman]

Johnw,
I got my MINI Cooper on 5/22 and drove the enxt day up to the White Mtns in NH..< ok.. I wont make the big trip on 6/21>, but in stop and go traffic on the parkway I felt the lurching on my CVT (in D) when it was less than 10 mph..
brought it in a week later < a windshield had cracked.. it was a rock... not by itself.> and Alex @ Hassel MINI, Freeport. updated the software..
there is a little lurching.. (mostly due to the lower torque I've been reading on the engine) but no where like it was before.
Before.. D lurched at low speeds.. SD went much smoother.. and Step was.. well.. Step.
Now, D is much much smoother.. SD still shoots like a rocket. and Step is.. well.. Step still (grin)

Is there a way to tell what software is installed????

Rob

2003 MINI Cooper, All Pepper white, Prem Pack, Cold Pack, Sport Seats, Chrome Bumper Inserts, CVT (I call this the Touring package)

[cooper4us]

Quote:

Originally posted by RobFriedman
Johnw,
I got my MINI Cooper on 5/22 and drove the enxt day up to the White Mtns in NH..< ok.. I wont make the big trip on 6/21>, but in stop and go traffic on the parkway I felt the lurching on my CVT (in D) when it was less than 10 mph..
brought it in a week later < a windshield had cracked.. it was a rock... not by itself.> and Alex @ Hassel MINI, Freeport. updated the software..
there is a little lurching.. (mostly due to the lower torque I've been reading on the engine) but no where like it was before.
Before.. D lurched at low speeds.. SD went much smoother.. and Step was.. well.. Step.
Now, D is much much smoother.. SD still shoots like a rocket. and Step is.. well.. Step still (grin)

Is there a way to tell what software is installed????

Rob


What does your service ticket repair copy says? It should state "DME software updated to version 33.2". If it doesn't or you don't have a copy of the service ticket receipt, head back to the dealer and don't leave until they'll give you a revised copy.

Other than that, there is no other way to verify what DME software version you currently have installed. Your dealer needs a special computer interface to access the car's DME and read the software version downloaded in there.
2003 MINI Cooper, All Pepper white, Prem Pack, Cold Pack, Sport Seats, Chrome Bumper Inserts, CVT (I call this the Touring package)

[onefastmini]

Rob, there is no way to tell if 33.2 is in a specific vehicle, only the software package #. The part #'s change when programmed with v.32, but cannot be differentiated after that. sorry guys. If its important to you, you should ask the service writer/tech to put it in his story.(they should do this anyway). ---Dan----

[nonsequitur]

OK guys. I've found a very impressive all-wheel dynometer facility in Redmond that I can rent by the hour with operator and software to test out this car. This exercise and any data from it will be in the Objective Testing thread. I have a 2nd week in July date reserved.

It is capable of very customized loading, so we are considering a "light" (car and driver with fuel load - 2,600#) and "heavy" (car, driver, passengers, luggage, and fuel - 3,300#) set of runs. Without aero forces, the results really don't convert directly to the street but this should be very interesting just the same.

We will find the power curve (torque most important, horsepower interesting) from the engine at the wheels.

I'm most interested in the transmission. A too-simple question is how does it work? In my previous post (4th above) there are some ideas for tests to run. What would be the most useful to us?

I've worried about the 0-60mph or 1/4mile type tests (although I think we should do them) because the CVT seems to perform best in the real world of rolling performance (like 20-70mph).

Does anyone have detailed suggestions for this exercise?

[Arnel]

Please do some power brake launches. Essentially what's normally done at the drag strip. You know, hold the brakes with the gas mashed to the floor. And then lift off the brakes to launch.

Which is the best method? In D, SD, 1, or 2???

I'm a seasoned bracket racer at the drag strip. And I recently picked up a CVT. I know it doesn't have allot of power. And it's a snail at launching compared to my 300ZX. But I'm at least hoping that it's a very consistent 1/4 miler. I may start bracket racing the MINI too. I'm just concerned that the CVT may not be able to deal with the abuse. And obviously, it would be nicer if the launch was just a little better.

I just got thru the break-in miles and I hope to take the MINI to the track this weekend to learn and experiment with the launches. Also, I'm looking to get some baseline times before doing an intake & exhaust.

[nonsequitur]

It's the "very consistent" prospect for this car that seems so exciting.

The mechanical system appears near-bulletproof, but time will tell. I'm hearing no real mechanical issues at all, it's all software integration within the MINI/BMW/ZF collection of choices.

[cooper4us]

I agree with Johnwald. The powertrain (TRITEC and ZF CVT) seems to be bulletproof. The problem all along has been the Siemens electronic engine management system software.

[nonsequitur]

I have a small pile of printouts from my session on the Mustang Dynometer. Preliminary reactions are some surprise, a little disappointment with the process, and confirmation of our seat-of-the-pants feelings.

SD @ Detent (not floored) is the fastest tranny selection - because it delivers MORE torque and horsepower to the wheels than any other selection, and power delivery is continuous - not "shifted" by the software. The computer does find the torque peak, and then holds rpm there for quicker turbine-smooth acceleration compared to the other selections. This is under strapped down, controlled conditions of the test facility, not on the road where other needs like cornering and braking also present themselves.

The engine/transmission is VERY heavily managed by the software. The graphs show clean, totally consistent, repeated by multiple pulls and distribution of power.

The dyno readings imply that the shifting behavior and "floored" throttle position are pure marketing, built into software for effect, as several of us suspected. That says nothing about fun, or the control delivered by a Steptronic "gear" in corners.

Acceleration pulls were done at the end of the session and they stack ranked: SD @ Detent, SD @ Floor, Step @ Detent, D @ Floor, Step @ Floor, D @ Detent. I wasn't allowed to "drive" these pulls and we didn't do enough of them to weed out driver variation so I'm not going to use these with much authority, except to say that it appears SD outpulls Step which outpulls D.

There is no doubt that the setup delivers very consistent performance and bulletproof management of the drivetrain -- very appealing characteristics.

[Griffy]

Has anyone noticed a lessening of performance in normal Drive mode after having the latest software update installed (33.2). Mine seems noticably less responsive in D mode now, like when selecting to high a gear in a conventional manual. SD & Step seem to be ok as far as I can tell. Even after the latest software, the car sometimes wants to lurch when coming to a stop, it hasn't quite done it yet, but you can just feel it trying to, I wont be surprised if it returns.

Cheers all.

[nonsequitur]

There MAY be some modification of the Drive program in 33.2, although I haven't noticed it.

One attribute of D is a less-aggressive map than SD. That appeared in our tests today with lower torque and whp calculations in D than SD. The shape of the curves for torque vs rpm were some different between D and SD, further modified by throttle position and detent switch engagement.

[Griffy]

Also...

Does it worry anyone that the oil in the CVT box is never supposed to be changed...EVER ?

Surely a Steel belt on two pulleys must dirty the oil somewhat, not to mention two clutches, especially seeing as its all hydraulic.

Is there a way of changing the oil at some point?
My local Mini dealer was the one that told me this is the case, let me know if I'm wrong. I dont have the manual here to check at the moment.

Cheers

[nonsequitur]

The objective for this limited dyno testing was to get multiple pulls for Torque and Horsepower in each tranny mode (DriveCVT@detent, Drive-floored, SportDriveCVT@detent, SportDrive-floored, Steptronic@detent, and Steptronic-floored).

The car was strapped into a Mustang Dynometer, with an exhaust monitor, in 86 degrees ambient temperature, and cooling fans to move heat out of the engine bay. The equipment was committed after me so we cut off the Steptronic (virtual manual 6-speed) as redundant after observing the SportDrive-floored data stream.

We rolled the car at 23mph in the tested mode, then accelerated to 110mph for data generation. We also did some "1/4 mile acceleration runs" at the end just to see what that that looked like.

Here is one example of the data and some notes done on Tuesday this week from two related SportDrive pulls:

SportDrive@detent (steady acceleration)
4000rpm: 83.08ft-lbs, 63.53whp
4100rpm: 85.95ft-lbs, 67.40whp
4200rpm: 88.83ft-lbs, 71.28whp
4300rpm: 90.38ft-lbs, 74.17whp
4400rpm: 90.61ft-lbs, 76.09whp
4500rpm: 90.85ft-lbs, 78.01whp
4600rpm: 92.27ft-lbs, 81.00whp
4700rpm: 93.69ft-lbs, 83.99whp
4800rpm: 94.83ft-lbs, 86.79whp
4900rpm: 95.68ft-lbs, 89.39whp
5000rpm: 96.53ft-lbs, 91.99whp
5100rpm: 100.81ft-lbs, 98.32whp
5200rpm: 105.09ft-lbs, 104.65whp
5300rpm: 107.86ft-lbs, 109.09whp
5400rpm: lost in printout
5500rpm: 110.41ft-lbs, 114.20whp
5600rpm: 66.24ft-lbs, 68.52whp
5700rpm: 22.08ft-lbs, 22.84whp

SportDrive-floored ("shifting" behavior)
4000rpm: 65.45ft-lbs, 49.78whp
4100rpm: 66.09ft-lbs, 51.60whp
4200rpm: 66.74ft-lbs, 53.42whp
4300rpm: 65.99ft-lbs, 54.02whp
4400rpm: 63.85ft-lbs, 53.40whp
4500rpm: 61.70ft-lbs, 52.78whp
4600rpm: 62.59ft-lbs, 54.83whp
4700rpm: 63.48ft-lbs, 56.87whp
4800rpm: 69.19ft-lbs, 63.70whp
4900rpm: 79.73ft-lbs, 75.32whp
5000rpm: 90.26ft-lbs, 86.93whp
5100rpm: 94.57ft-lbs, 92.54whp
5200rpm: 98.89ft-lbs, 98.14whp
5300rpm: 103.54ft-lbs, 101.31whp
5400rpm: lost in printout
5500rpm: 98.51ft-lbs, 102.79whp
5600rpm: 97.47ft-lbs, 103.23whp
5700rpm: 96.43ft-lbs, 103.66whp
5800rpm: 76.72ft-lbs, 83.10whp
5900rpm: 38.36ft-lbs, 41.55whp
6000rpm: 0.00ft-lbs, 0.00whp

These numbers look weird, and graph weird, until you think about what is happening to the car while they are being generated:

Under SportDrive@detent the computer quickly brings rpm up to 5500rpm, AND HOLDS IT THERE, right at peak power, very steady, while the CVT turbine-smoothly manages acceleration. Peak power is employed for the whole pull from 25mph to 110mph.

Under SportDrive-floored the computer runs rpm up to cutoff at 6000rpm, dropping rpm to about 5100rpm while adjusting the belt to a new "gear", to then run it up to 6000rpm again. The rpm waves back and forth along the graphed "power bump" between 90ft-lbs (5000rpm) through 99ft-lbs (5500rpm) to 77ft-lbs (5800rpm) -- all lower power than the 110ft-lbs generated in SportDrive@detent.

This stuff is wild. I have more data coming to me from the 26 pulls we performed. Help me with questions, or challenges, or ideas to steer analysis.

The numbers, and the performance of the car, were VERY CONSISTENT. We went back to an earlier test after running others just to see if results could be repeated. Results always repeated.

The graph lines are very sharp, totally consistent. This car is being heavily managed by the computer. The SportDrive@detent position just put the engine in its maximum power position on the curve and held it there, no wiggles. I'll bet this is why I could pull with the Cooper racecar all the way to braking - no losses while shifting, no clutch disengaged, just grunt "continuously" applied to the road!

[Arnel]

Interesting...

So holding SD-at-detent is the key?

So does this mean, a 1/4 mile pass with SD-at-detent will yield better times than SD-to-the-floor? Sounds like it would which I would never have guessed.

And thanks for the info. Very informative.

[nonsequitur]

That's what needs to come next I think. An evening at some drag strip running against the boys and their Honda Civics just to test out what happens in the real world with data from those controlled conditions too. It would be fun to have two Cooper CVTs, side by side!

My primary interest is road driving and open track days, but we did do some "1/4 mile Sprint Tests" to see what would happen in each mode through a restricted distance.

SD@detent beat out SD-floored. SportDrive was fastest, Steptronic (auto upshift at 6000rpm) was second, and Drive was third.

The thing I couldn't really figure out how to test is low speed, low rpm behavior of the CVT, where I suspect that lots is going on in the CVT and software to make it "smooth". The CVT is not quick off the line, but it is consistent, which points to heavy management (in software) of all the variables.

I felt the whole detent switch thing on the track, but didn't really believe it until we saw the graphs and the results. I want to think about all of this some more, and then ask "How do we drive it better?" over on the Driving Technique thread.

[nonsequitur]

OK, here's something to check out.

What is happening here?
And, why is it happening?

1.
With the car motor running, put the CVT in "N", foot on brake.

VERY SLOWLY add gas so rpm creeps up past 1500, 1600, 1700, 1800, 1900, 2000, 2100, 2200, ... You will hear a "click" at 2000rpm. Let the rpms come back down, again a "click" in the cabin at exactly 2000rpm.

If you do gas quicker, the uptrend click will look like 1600rpm or so, but I'm sure that's only an impression related to the tach needle.

Are we hearing a switch to select or deselect the starting functions in the transmission? Or, are we hearing the transmission or some other component physically adjusting? The click does not appear to be the e-throttle because it is rpm-dependent not throttle position dependent.

2.
Now try this one:

With the car motor running, put the CVT in "N", foot on brake.

Floor it. Or, if that scares you, add lots of gas progressively.

The maximum rpm you'll get is 4000, exactly. No amount of gas will change it. This is obviously EMS2000 control, but what is behind this logic and function?

3.
What happens in the manual cars: One 5-speed, Cooper 5-speed, Cooper S 6-speed?

[Spookyfish]

No. 2: I think to prevent damage. More automatics (Audi's multitronic? BMW's SMG?) do this. Think what will happen if you drop it from N in D at 4000rpm, now think again at 6000rpm.

Would be nice to get a complete inside look on the Mini CVT. We need someone who knows people for that.

[onefastmini]

GRIFFY, You do need to replaced the CVT oil at every Inspection service. That would be roughly 25,000 miles and again at 55,000 miles. There is no filter to replace. The service maintainence book should indicate this interval.

Somebody wanted to see the inside of a CVT. I will try my best to pull some strings and ask someone to send me photos so I can post them here for you.


---------Dan--------