[PencilGeek]

It always seemed fairly obvious to me that a driver's shift speed would affect his 1/4 mile drag race ET and trap speed. To me, this makes natural sense. So when my own meager "BMW Performance Steering Wheel" 1/4 mile drag results (about as inaccurate as it gets) came under fire, I quickly pointed out that I'm a slow shifter on my 6MT, and this would surely account for my mediocre results. To my surprise, my explanation was vigorously attacked by a self-proclaimed BMW know-it-all and self-proclaimed drag racing expert -- Sticky. He said my car was weak and that's why it performed so poorly. That was a drum he's never quit beating, in spite of all of the data to the contrary.

But I'll admit, I'm a novice with drag racing...in fact I really don't know anything about it. But I do know that if two identical cars take off from the same point at the same time and one of them waits two seconds between every shift, then while it's coasting in neutral it will not be accelerating. Therefore, it will be impossible for that car to have the same 1/4 mile trap speed of the same car that shifts as fast as it can.

I sat on one end of the debate -- as a newbie who never drag raced a car in my life, and at the other end was this self-proclaimed know-it-all of everything. He insisted he was right...and I was wrong. So more than a year ago, I asked him to prove it at the drag strip by running some of his runs in D2 mode instead of D5 (or S6). No surprise to anybody who knows how he hates to be proven wrong...he refused with some flimsy excuse that he was out to set records...not run experiments. Back in my corner, I produced CarTest physics-based car modeling software results that showed exactly how shift speed affects trap speed. Back in his corner, he laughed every time I used car test. Back in my corner, I laughed everytime he laughed at CarTest because CarTest has proven dead-nuts accurate for every performance prediction on my own car . In fact CarTest has proven so accurate that it's been within a few hundredths of a MPH on my Mojave Mile results.

I believed this would be a very easy test to run...all I needed was a car owner with DCT and a vBox. The test to figure this out would be quite simple. Run the DCT in D2 and compare the vBox results against the same car running in D5. Even though the shift speed difference between D2 and D5 is only a few milliseconds, I was hopeful it would prove the results one way or the other. Now all I needed was to find the car and test it. I ended up waiting more than a year until I got my answer.

This weekend, I borrowed a guy's DCT car and took it to a remote road in the desert that was nearly perfectly flat in both direction. There are no side streets, except the ones I would use for the tests. I planned to place the car in D2 and turn on the vBox. I would run from 0-130 MPH -- as this would give me 1/4 and 60-130 MPH results. I would run the car in both directions on the same exact piece of road for three runs in each direction (six runs total), then I would switch to D5 and run the same test all over again.

The car I borrowed was nearly bone stock. Bone stock is ideal because if the results are obvious with a bone stock car, they will be even more obvious and exaggerated on a highly modified car (such as a supercharged car). So taking a near bone stock car for the test was also going to make it the most difficult for me, and give the naysayers all of the benefits.

I found the perfect road, hooked up my Video vBox and tried one practice run in each direction to make sure the road was safe. After I was satisfied, I turned on the vBox and ran in one direction. Then I ran in the opposite direction. Back and forth three times for a total of six runs. I switched to D5 and repeated the runs. I was running low on gas, so after two runs in each direction (instead of three), I abandoned the third set of runs and went home.

The results were overwhelmingly conclusive. Even though I was only changing the shift speed by a few milliseconds, the effect on 1/4 mile trap speed was obvious and I'd say even profound. Wasting a few milliseconds shifting affected my 1/4 mile trap speed by as much as 2MPH. Even the best D2 run couldn't beat the worst D5 run. Sticky was wrong...again. That's right: the self-proclaimed expert of drag racing and all automotive knowledge was proven dead wrong again.

As a final test, I input all of the actual car data in CarTest to see if it could predict the same results. If it did, it would be (yet) another slap in the face to the self-proclaimed expert who regularly mocks the use of this tool. I just happen to have the Dynojet dyno file from this exact car taken only a few weeks ago. CarTest allows the user to input the actual dyno results. After doing so, I adjusted the launch methods to match how I used the DCT. For D2, I selected a 0.20 second shift time (200 milliseconds). I chose this shift time after some research on three or four different car forums and other web sites who claimed that 0.20 seconds was about the slowest the DCT would shift (D2 mode). For D5, I selected 0.08 seconds shift time (80 milliseconds). In input the exact weather conditions I collected during my run, including elevation and wind speed and wind direction. Then changing nothing but shift speed, I let CarTest estimate the 1/4 mile ET and trap speed, along with a 60-130 MPH time as well.

Here's the results.

Weather Conditions:
Temperature: 55.9 degrees F
Pressure: 30.04 inHg
Humidity: 27%
Density Altitude: 30 Ft.

CarTest Estimates for D2:
1/4 Mile Elapsed Time (ET): _ 12.65
1/4 Mile Trap Speed: ________ 112.55
60-130 MPH: _________________ 12.63

Actual Results for D2:
_______________________________Run-1, _Run-2, _Run-3, _Run-4, _Run-5, _Run-6
1/4 Mile Elapsed Time (ET): _ 13.208, 13.612, 13.412, 13.204, 13.177, 13.521
1/4 Mile Trap Speed: ________ 112.93, 112.23, 112.81, 112.84, 113.26, 112.13
60-130 MPH: _________________ 12.595, 12.435, 12.512, 12.504, 12.411, 12.456

Average Results for D2 (average of three medium runs):
1/4 Mile Elapsed Time (ET): _ 13.276
1/4 Mile Trap Speed: ________ 112.86
60-130 MPH: _________________ 12.537

CarTest Estimates for D5:

1/4 Mile Elapsed Time (ET): _ 12.48
1/4 Mile Trap Speed: ________ 113.83
60-130 MPH: _________________ 12.20

Actual Results for D5:
_______________________________Run-1, _Run-2, _Run-3, _Run-4, _Run-5, _Run-6
1/4 Mile Elapsed Time (ET): _ 13.135, 12.987, xx.xxx, 13.034
1/4 Mile Trap Speed: ________ 113.35, 114.09, xxx.xx, 113.62
60-130 MPH: _________________ 12.147, 12.104, xx.xxx, 12.128

Average Results for D5 (average of three medium runs):
1/4 Mile Elapsed Time (ET): _ 13.052
1/4 Mile Trap Speed: ________ 113.69
60-130 MPH: _________________ 12.13

Conclusions:
To me, the results seem pretty conclusive. Even changing the shift speed by 120 milliseconds (or whatever the actual value may be between D2 and D5) created a pretty significant change in all 1/4 mile drag results, and 60-130 times as well. As I believed at the very beginning, shift speed does affect trap speed, no matter what the self-proclaimed drag race expert seems to think.

CarTest results proved to be very accurate as well. The D2 predictions vs. actual results were about as close as they could get. 1/4 mile ET was predicted to be 12.65, actual was 13.26 (4.7% difference). 1/4 Trap Speed was predicted to be 112.55, actual was 112.86 (3/10ths of 1% difference). 60-130 time was predicted to be 12.63, actual was 12.537 (less than 1/10th of 1% difference).

The D5 predictions vs. actual results . 1/4 mile ET was predicted to be 12.48, actual was 13.052 (4.4% difference). 1/4 Trap Speed was predicted to be 113.83, actual was 113.69 (1/10th of 1% difference). 60-130 time was predicted to be 12.20, actual was 12.13 (1/2 of 1% difference).

Trap speed and 60-130 MPH results were both less than 1/2 of 1% difference from the predictions for both D2 and D5. CarTest predicted 1.28 MPH trap speed difference, and the actual was 0.83 MPH difference. CarTest predicted an 0.15 second difference for 60-130 MPH runs, and the actual difference was 0.407 seconds. And even though 1/4 mile ET was 4.4% off, Cartest predicted a 0.17 second difference, and the actual results were 0.224 second difference.

VBox files will be made available for analysis to anybody curious enough to dive into the results.

VBox Results (I've highlighted the shift points):



Car Test Configuration:





Additional Information:

I did a little more digging into the results, and it's very clear that clutch engagement time is different between D2 and D5. I looked at the 60-foot times, and it's clear that the clutch engagement is faster in D5 than D2.

At the same time, I exported as much data from the runs as I thought made sense. I tried making some graphs of the data, but nothing was nearly as clear as simply presenting the data itself.

1/4 Mile results:
Even though these aren't actual drag strip results, I tried to simulate them as best as I could using the VBox tools software. I did my best to look up the NHRA rules and figure out how times, distances, and trap speeds are measured. Then using what I learned, I programmed the VBox software to simulate the same thing. For example, I used a 1-foot roll-out, and the trap speed is the average of the speed at 1321ft, and 1257ft (Q-60 ft, with 1-foot roll-out).



Gear Analysis: Time in Gears


Gear Analysis: Distance in Gears


Speed Analysis: Time to Speed


Speed Analysis: Distance to Speed


Video:http://www.youtube.com/watch?v=MAXGBr4ioeA

[captbilly]

I really appreciate when someone attempts to do an objective test (scientific method) and looks at what variable might effect the outcome, and then controls those variables. You did about as good an experiment as one could. I suppose that if it were possible it would have been useful to measure the actual shift speed, but even without that info it seems pretty clear that switching from D2 to D5 has a statistically significant effect on the outcome. If we are being absolutely critical of the data and experimental setup we would have to at least consider the possibility that the software does other things then change shift speed (when switching from D2 to D5). Potentially the software could also change the shift points slightly (I don't think this is the case), or even interact with the throttle (I don't think it does this either) or engine map (also don't think this is the case).

I am surprised by how few people actually are willing to do what it take to conclusively answer these sorts of questions (ie does shift speed effect 1/4 mile time). We are beginning to see this grow to epidemic proportions in our society as a whole, where people who don't have a clue about what they are saying spew nonsense and nobody checks the facts or logic. You premise, that shift speed effects 1/4 mile time seems logical enough without even thinking about it much, but the fact that you took the time to do a critical analysis and collect real meaningful data, speaks volumes about your objectivity. I hope you attempt to do the same thing where you work, when you vote, how you treat your friends (children or spouse, if you have them). Our world would be a dramatically better place if we all critically analysed what goes on around us.

[AMPowerJ]

Nice work as always ... The fact that someone would think shift speed doesn't affect 1/4 times frankly baffles me. Again I appreciate all you did to prove that it does, but it sorta seems like having to prove the pedal on the right accelerates the car and the pedal in the middle (or on the left for you DCT converts) stops the car.

[Cleveland]

Very interesting (and logical) conclusions.

[duk]

Crazy to think that shift speed would not impact overall speed times (maybe albiet how small) in the first place.

[pbonsalb]

Is it possible that with 550 rwhp, the faster shift speed breaks traction more at the shifts and thus negates the gains found at 350 rwhp? Significant wheelspin is fun, but not fast.

[epholler]

He's gonna say the D5 times were faster because there was less gas in the car.

[k1ndafast]

I would be interested to see the 60ft for all the runs in addition to trap and ET.

At what RPM where you shifting? And was it identical for all the runs?

Just playing the devil's advocate, but wouldn't you agree that your tires got warmer (grip increased) and you got more used to launching your car by the time you started experimenting with D5?

[PencilGeek]

Originally Posted by k1ndafast

I would be interested to see the 60ft for all the runs in addition to trap and ET.

Good idea. Trap and ET's were already posted. I'll post the rest of the data below.

At what RPM where you shifting? And was it identical for all the runs?

D2 and D5 are automatic modes. I just mashed the pedal from a dead stop and let the car do the rest. The car should be shifting at Max RPM all on its own.

Just playing the devil's advocate, but wouldn't you agree that your tires got warmer (grip increased) and you got more used to launching your car by the time you started experimenting with D5?

Taking your suggestion, it does seem like run-1 in D2 and D5 are both a bit slower off the line than all the other runs in the group. This cold be caused by tire warmth.

[PencilGeek]

Additional Information:

I did a little more digging into the results, and it's very clear that clutch engagement time is different between D2 and D5. I looked at the 60-foot times, and it's clear that the clutch engagement is faster in D5 than D2.

At the same time, I exported as much data from the runs as I thought made sense. I tried making some graphs of the data, but nothing was nearly as clear as simply presenting the data itself.

1/4 Mile results:
Even though these aren't actual drag strip results, I tried to simulate them as best as I could using the VBox tools software. I did my best to look up the NHRA rules and figure out how times, distances, and trap speeds are measured. Then using what I learned, I programmed the VBox software to simulate the same thing. For example, I used a 1-foot roll-out, and the trap speed is the average of the speed at 1321ft, and 1257ft (Q-60 ft, with 1-foot roll-out).



Gear Analysis: Time in Gears


Gear Analysis: Distance in Gears


Speed Analysis: Time to Speed


Speed Analysis: Distance to Speed