M-DCT vs 6 speed - simulated performance improvements

[swamp2]

Here goes some more simulation. For those who think simulation is rubbish, I guess I will be a little harsh and just say right up front "go away". I am willing to discuss simulation but tire quickly of beating the dead horse and being defensive. For those of you who realize the accuracy, relative accuracy and benefits I'm sure you will enjoy this - both the results and expected accuracy.

OK in my last post about simulation and M-DCT (comparing to C63 AMG and/or R8) I had a very rough optimization done to the gearing. M-DCT will certainly have different ratios than 6MT and they will be optimized in some sense. The gearing optimization feature built into CarTest is very cool, nonetheless it does not really work. And even though it works half way, understanding what exactly BMW will choose to optimize involves some guessing. If you fully optimize for time to get to Vmax, lower speed acceleration times suffer and vice versa. What I chose was a broad based optimization improving all of the times for 0-(speed X) and all of the times for distances 0 up to 1/4 mi as compared to the 6MT. There will be gearings that will best these for any one category, but all around these are fairly nicely globally optimized. I think they will be fairly close to what BMW chooses.

In gear accelerations improved in most cases. I did check a variety of rolling start acclerations other than the ones posted more typical of a track and with higher speeds involved such as 40-120, 50-90, etc. both allowing and not allowing shifts. Again the DCT ranged from a few tenths better to 1+ second better on the wider ranges.

Also "driveability" as defined by "lazy acceleration" improves substantially. Here I mean the time to accelerate from speed X to speed Y in a single gear (and not the optimal gear but one with a gear number numerically higher than the optimal gear). Low rpm big torque engine fans rejoice - more gears does offer some of the same type of benefits as this type of engine. As contradictory as that may seem, as in "hmmm 7 gears, I'll have to be working those darn paddles like crazy to go fast" which simply is not true. See the results to convince yourself.

The 1/4 mi times did not end up as good as I guessed and the reason is that the 6MT can finish the 1/4 in 3rd gear whereas the M-DCT must be in 4th and in the latter half of 4th on the DCT the 6MT does its most drastic catch up due to better relative acceleration (see lower right plot of acceleration vs time). Also notice here the clear differences between M-DCT and MT shift times. Soon after the 1/4 when the 6MT goes to 4th the DCT makes back significant groud in the time to speed graph.

If some normal (i.e. metric system) folks want other distances or in gear accelerations using km/hr let me know an enitre list and I will run them. All of the hard work is done now. Hope you enjoy this and find it useful.

Note on shift times: My simulations use DCT shift times of about .03 s (30 ms) and MT shift times of .30 s. This means that the MT has a very good driver, but perhaps not as good as THE BEST driver. I think these values are reasonable and fair to both cars performance. Note also that M-DCT times arbitrarily end at 15 seconds in the graphs, another small bug when using small time steps to capture DCT physics properly.

[m_bazeepaymon]

this is M divisions secret weapon if this simulation is correct

Lets just hope

[e36jakeo]

Car and Driver's online article comparing the C63, M3 and RS4 has the M3 doing 0-60 in 4.3 seconds and 1/4 mile in 12.8 seconds at 113 mph.

http://www.caranddriver.com/comparis...-m3.html?al=98

At 113 the M3 manual is in 4th gear, so I believe the DCT will have an even greater advantage since it will be more in the "meat" of 4th gear and will have gained the gearing advantage of a shorter 3rd and 4th gears. Combine that with the improved shift time and I think the M3 with DCT should hit 60 in 4.1 or better (assuming warm tires) and the 1.4 mile in 12.5 at 115.

Your sim is VERY close to these numbers - VERY NICE!

[footie]

If you haven't experienced DSG (Audi talk) then you will be amazed. The acceleration decrease between gears in nil, off my head is 20ms or in other words your don't notice any drop off in acceleration.

I see your 0.66s difference to 100mph is close to my 0.75s I suggested else where, but I still feel the gap will be bigger as you close in on the 150mph mark. The shift from 1st to 2rd is quicker than the cross shift to 3rd and again to 5th, I don't think you are taking this into account in your calculations.

[Sedan_Clan]

Good work Swamp. Valuable effort well done!

[swamp2]

Originally Posted by m_bazeepaymon

this is M divisions secret weapon if this simulation is correct

Lets just hope

Yes and no. Will it be great, I'm fairly sure. Did Audi, Bugatti, VW and Mitsubishi beat BMW to the market with the technology, yes as well. So hardly a "secret". They are one of the front runnners to offer a DCT for sure but not leading with innovation as well as I would like. Heck at least they will be beating MB and Porsche!

[phoenixbmwlife]

Nice. I don't know if I should wait for MDCT or just get the manual up front. Maybe I should just wait to save up some more money. But it will kill me to see the M3 out for months with me still waiting for DCT.

[swamp2]

Originally Posted by e36jakeo

Car and Driver's online article comparing the C63, M3 and RS4 has the M3 doing 0-60 in 4.3 seconds and 1/4 mile in 12.8 seconds at 113 mph.

http://www.caranddriver.com/comparis...-m3.html?al=98

At 113 the M3 manual is in 4th gear, so I believe the DCT will have an even greater advantage since it will be more in the "meat" of 4th gear and will have gained the gearing advantage of a shorter 3rd and 4th gears. Combine that with the improved shift time and I think the M3 with DCT should hit 60 in 4.1 or better (assuming warm tires) and the 1.4 mile in 12.5 at 115.

Your sim is VERY close to these numbers - VERY NICE!

Thanks. As it often the case with simulation; the differences may be closer than the absolutes in any one particular A-B test comparison. So in this case existing best test data for MT + gains of simulation deltas will be better than existing simulation absolutes for DCT. Does that make sense?

However, I don't think the real life differences in an apples to apples MT vs. DCT transmission will show a significantly larger advantages than my prediction.

Again not to be overlooked is consistency. If we get a decent luanch control program the ability of 95% of DCT drivers to best 95% of MT drivers will be a given, based on both LC and even an automatic mode in DCT!

@footie: Not sure what you mean in comparing 1->2 vs. 2->3. Each successive up shift is on alternating gear sets and when accelerating the next gear is always pre-selected. There should be no time difference between any two upshifts. My simulations use about 30 ms for all upshifts. Perhaps the Audi box works differently? Not AFAIK.

[footie]

Originally Posted by swamp2

@footie: Not sure what you mean in comparing 1->2 vs. 2->3. Each successive up shift is on alternating gear sets and when accelerating the next gear is always pre-selected. There should be no time difference between any two upshifts. My simulations use about 30 ms for all upshifts. Perhaps the Audi box works differently? Not AFAIK.

Sorry mate, I didn't make myself clear on this.

What I mean to say was with a manual gearbox the time taken to shift from 1st to 2rd was less than cross shifting from 2rd to 3rd and again from 4th to 5th, the ref. 'cross shifting' meant crossing the shift gate. It normally takes over one and a half times longer to shift a cross the gate (2nd to 3rd) than straight shifting (1st to 2rd).

Hope this explains.

[swamp2]

Originally Posted by footie

Sorry mate, I didn't make myself clear on this.

What I mean to say was with a manual gearbox the time taken to shift from 1st to 2rd was less than cross shifting from 2rd to 3rd and again from 4th to 5th, the ref. 'cross shifting' meant crossing the shift gate. It normally takes over one and a half times longer to shift a cross the gate (2nd to 3rd) than straight shifting (1st to 2rd).

Hope this explains.

Ahhhh, got it. Would be nice if CarTest allowed different shift times for different shifts but wouldn't that really only amount to maybe .05-.1 seconds every other shift? Pretty small compared to .3 vs. .03!

[footie]

Hey, don't knock the difference it makes. If you are the guy about to race every little bit could win the race for you.

I found the difference between quick shifting in the manual vs DSG of roughly 0.3s to 0.5s, the later being the shift across the gate. By quick shifting I mean keeping the throttle fully pressed while clutching and changing, not the best for the gearbox but very effective. That meant in a TTv6 I found the difference just under a second by the time you had reached 100mph.

I know the TT is AWD which is slower to shift gear with so the RWD M3 might be a little quicker for this.

[ruff]

The DCT version may not be faster off the line depending on launch control. If we get launch control in the states, which I doubt, it will depend on how much BMW wants to safegaurd the drivetrain if they offer it. The stick shift Evo is faster off the line than the DCT version with launch control for this reason.

[mkoesel]

Originally Posted by ruff

The DCT version may not be faster off the line depending on launch control. If we get launch control in the states, which I doubt, it will depend on how much BMW wants to safegaurd the drivetrain if they offer it. The stick shift Evo is faster off the line than the DCT version with launch control for this reason.

ruff, is the 6MT M5 also quicker to 60MPH than the SMGIII M5?

[mesier1111]

Originally Posted by mkoesel

ruff, is the 6MT M5 also quicker to 60MPH than the SMGIII M5?

I know I'm not ruff, but that is probably a bad judge for whether the 6MT for the M3 with be faster than the M-DCT. Why you ask? Because the M5 was not designed to use a 6MT, and they just took the 6MT from the regular 545/550i and beefed is up a little, but not nearly to the degree that the SMG III is beefed up to. However, the M3 got a 6MT designed for the car, not just thrown in there because people were bitching the car didn't have one (as they did with the M5).

[mkoesel]

Originally Posted by mesier1111

I know I'm not ruff, but that is probably a bad judge for whether the 6MT for the M3 with be faster than the M-DCT. Why you ask? Because the M5 was not designed to use a 6MT, and they just took the 6MT from the regular 545/550i and beefed is up a little, but not nearly to the degree that the SMG III is beefed up to.

Yep, I'm aware of that - and that was partly why I asked. I am curious how the 6MT does, despite it being an afterthought.

However, the M3 got a 6MT designed for the car, not just thrown in there because people were bitching the car didn't have one (as they did with the M5).

Might wanna check that. M3 6MT = 335i 6MT. However, what is true is that the car was designed around both this tranny and M-DCT from the outset, unlike the M5.

[lucid]

Cool stuff. Good to see the 0-100 time dipping below 10sec.

[Sedan_Clan]

Originally Posted by ruff

The DCT version may not be faster off the line depending on launch control. If we get launch control in the states, which I doubt, it will depend on how much BMW wants to safegaurd the drivetrain if they offer it. The stick shift Evo is faster off the line than the DCT version with launch control for this reason.

Why wouldn't we get LC? We have it now, so why would that change?

[swamp2]

Originally Posted by Sedan_Clan

Why wouldn't we get LC? We have it now, so why would that change?

The E46 M3 with SMG in the US has what is know as a "crippled" LC. It is LC but is designed to minimize clutch and transmission wear COMPARED to the EU version of LC. Too many stop light drag racers and "too good" of a warranty here in the US.

Unfortunately I suspect the same for the M-DCT. Perhaps something that can be "hacked" with software. Does such a hack exist for the E46 M3

[mesier1111]

Originally Posted by mkoesel

Yep, I'm aware of that - and that was partly why I asked. I am curious how the 6MT does, despite it being an afterthought.



Might wanna check that. M3 6MT = 335i 6MT. However, what is true is that the car was designed around both this tranny and M-DCT from the outset, unlike the M5.

Not completely the same transmission. I can't remember the specifics but the plate and something else that would wear down is different. Shorter throws too I believe, but I'm sure ruff or someone will correct me if I'm wrong.