|
 |
| To Shift, Or Not to Shift |
|
Author & Screenshots:
|
Hanchey
|
Add'l Content:
|
Fair
|
Date Posted:
|
08.27.06
|
Last Edited:
|
12.18.08
|
To Shift or Not to Shift, That Is the Question... As a sports car enthusiast you probably know all of the key aspects you've been told over the years; weight is bad, slow in/fast out, smooth is fast. Did you ever stop to think about banging the rev limiter versus banging that next gear? In autocross or road racing, there are often short sections on course where you see significant speed changes, and frequent up- or down-shifting can occur within only a handful of seconds. Where your car's gearing and the course speeds work out to put you in that "next gear up" speed range for a very brief amount of time - too fast for the lower gear, but the following section of course puts you back down into that gear anyway. This article aims to prove (in certain conditions) when it is better to hold on in that one gear versus the time spent shifting up and back down.
Note: After racing in a car now with a semi-automatic gearbox (2008 EVO X) we now see that a lot of these shifting choices are no longer an issue - the computers can match revs perfectly for every downshift, and both up and downshifts are smooth and lightning fast (and can even be sone automaticly in "S-Sport" mode, such as the Getrag equipped EVO X MR). This can be a significant advantage in autocross use, where the engine can keep pulling at any speed and never be left "sitting on the limiter" for seconds at a time, not accelerating, and with no hurried up/down shifts need to be made. Is it an unfair advantage? Not until they are banned. :)
 |
Our willing test mule:
SCCA STU class prepped
1997 BMW M3
|
|
Time is money, and in autocrossing or road racing, the more time you spend on each lap, the worse your finishing position will be. The biggest waste of time in racing is coasting - anytime you are not into the throttle or brakes, you're wasting time. There are situations when a series of corners produces speeds that are at or near the end of a particular gear for a car, and the driver has to ask himself: ''Do I hold a gear longer or short shift up then back down?'' And in autocrossing, the time between corners is usually shorter, meaning you have much less time to spend shifting up and back down, and these choices become more difficult. Then the other major question in the driver's mind, "How much do I trust those valve springs? Will that oil pump stay in one piece bouncing off the rev limiter at 7000 rpm??" Time and Money, indeed.
The following graphs outline an autocross event hosted by the Lonestar Region of the BMWCCA. Lonestar region autocrosses are typically more Solo 1-like in nature. They run over 100 seconds long and have several fast sectors including a 91mph straight. This is atypical for Solo2, but the results are still applicable to slower courses. Luckily everything is bigger in Texas including the Mineral Wells event site.
The test was conducted under dry conditions approximately 80 degrees with the same driver. The test car was a 1997 BMW M3 prepared for the SCCA's STU class, but running Toyo Proxes RA1's per BMW class rules. The car currently has a rev limiter of 7000 rpm and maintains its stock 3.23 rear gear ratio.
The first sector demonstrates a short straight where the driver must choose to stay in 2nd gear or upshift to 3rd for a brief period of time. Five laps were run with a 3 minute break between runs. The first and last laps have the driver shifting to 3rd to minimize any discrepancies from the driver learning the course. A Race Technology DL-1 Data Acquisition System (with GPS and accelerometers) was used to collect the data.
Figure 1: Course Map used in testing
Figure 2: Sector 1 Map |
Legend:
Lap 1: Black
Lap 2: Red
Lap 3: Green
Lap 4: Blue
Lap 5: Purple
|
The entrance to this sector starts out at approximately 45 mph while the end of the sector is approximately 62 mph. Two seconds into the sector, the driver must make a decision on whether to shift or maintain the current gear. On runs where the driver chooses not to shift, the rev limiter is held on the car for approximately 1 second.
Figure 3: Sector 1 Speed vs. Time
Looking at the time vs. speed graph you'll notice that the driver brakes earlier when holding second gear and brakes possibly too late when shifting to third. You can theorize that the driver is either trying to make up for the lost time shifting or his timing is off because of the shift. Since you can not look at a single corner without considering the "knock on" effect of other corners, we'll look at the data in the next sector to determine if he lost or gained time with the two techniques.
Clearly looking at this one sector by itself, the driver is consistently faster holding the rev limiter versus the time required to shift. This makes sense. The driver spends at least 400ms upshifting and downshifting is approximately the same time if not more. Add the values together for 800ms. The rev limiter was held for approximately 1000ms (1 second), therefore holding the rev limiter for 1 second is the equivalent of a shift! Add to this the chances of a mistake in the up or down shift and the conclusion is simple. If you hold the rev limiter for approximately 1 second (or less), you're making the right decision.
Figure 4: Sector 2 Map
Sector 2 is a fast second gear turn with no shifting. The driver must maximize this 180 degree increasing radius turn since it leads to the 90+ mph straight away.
Figure 5: Sector 2 Speed vs. Time
| Driver Runs |
Sector 1 |
Sector 2 |
Total Time |
| Lap 1, 3:48.21 |
4.67 |
13.21 |
17.88 |
| Lap 2, 3:45.12 |
4.47 |
13.11 |
17.58 |
| Lap 3, 3:35.05 |
4.5 |
13.08 |
17.58 |
| Lap 4, 4:30.11 |
4.38 |
12.88 |
17.26 |
| Lap 5, 3:34.82 |
4.42 |
12.99 |
17.41 |
Table 1: Sector 1 + Sector 2
|
As mentioned, it is not possible to view one sector of a track without considering what happens to the next sector or sectors. In a series of turns, for example, increasing your speed in the first of three turns, but leaving yourself in the wrong line for the remaining turns does not decrease your lap times overall. The next sector is a short second gear straight (no shift required) followed by an almost 180 turn that leads to the long straight. The turn before the long straight is one of the most important turns on the course and determines the maximum speed on the straight. Interestingly enough, the driver's first run was the slowest entry speed into the corner and highest exit. Remember slow-in, fast-out? The remaining four runs are all similar. The Sector 1 runs where the driver remains in second gear are faster, but only by 0.4mph but a maximum of 0.11 seconds in the sector. The results indicate that staying in 2nd gear sets up the remaining corners better. The time spent shifting is obvious, but what the results show is the "knock on" effect of that time spent in the next sector.
Figure 6: Sector 3 Map
Sector 3 is the starting line of the course. It is immediately followed by a tight 90 degree turn. Often in Solo 2 events, the starting line precedes a sharp turn; no big hairy launches in Solo 2. For the test car, the straight ends in the top of first gear, approximately 30 mph.
Figure 7: Sector 3 Speed vs. Time
| Driver Runs |
Sector 3 |
Sector 4 |
Total Time |
| Lap 1, 3:48.21 |
7.68 |
8.44 |
16.12 |
| Lap 2, 3:45.12 |
7.41 |
8.24 |
15.65 |
| Lap 3, 3:35.05 |
7.56 |
8.26 |
15.82 |
| Lap 4, 4:30.11 |
7.44 |
8.08 |
15.52 |
| Lap 5, 3:34.82 |
7.5 |
7.81 |
15.31 |
Table 2: Sector 3 + Sector 4
|
Looking at the data, this might just be a toss up. The driver is .04s faster (lap 2 vs. lap 4) when the car is held in first gear and shifted on the front straight in Sector 3. Lap 5 is much slower (the driver did not upshift in Lap 5 until the next straight). However, overall lap 5 is faster. Again, look at the next sector to determine the "knock on" effect. Looking at the next sector, lap 5 wins by 0.21s! Unfortunately it appears there are several variables at work. The driver in lap 5 appears to have taken a better line. Notice how the driver goes "slow in/fast out" on lap 5 and therefore is accelerating more quickly at corner exit. This could be a situation where the driver took a better line in Sector 4 on Lap 5 for the overall win. Each event will vary, in this case there are two runs very close to each other with and without shifts. The driver will have to decide which is better.
By shifting early, the driver is able to look ahead and set up for the next corner. Consequently, different cars may not find this to be true. A rev-happy Honda S2000 in the same situation would probably not short shift to second since it will likely kill its time due to lack of torque at lower revs. A ZR-1 Corvette (tall gearing, large spacing between gears) almost never shifts in a typical autocross. Know the strengths of your car.
One of the first things to check on a given car before prepping it for Solo2 is the ultimate speed in 2nd gear - a combination of transmission gearing, final gearing and tire diameter. This is a great tool: http://www.f-body.org/gears/ Look up the ratios and tire sizes for your car beforehand and you can find 1st and 2nd gear terminal velocities. If you have sufficient mid range, do not hesitate to try the "short shift" in these between gear situations, but if your car is a "revver", it might pay off to hold the lower gear - within reason (and the warranty period!).
Figure 8: Sector 4 Speed vs. Time
|
Each course will be different and each car is different. For this test, it was concluded that shifting should be avoided if possible and shifting in less critical areas (first turn) does not hurt the performance of the run. Also, holding the rev limiter for 1 second or less proved to be faster than an upshift followed quickly by a downshift. If you're lucky enough to be driving a car with paddle shifters, your results may vary. You may also drive a car with a 9000 rpm redline, don't shift! The next time you have the choice remember, bang that rev limiter on a straight instead of upshifting, if its 1 second or less at the limiter.
At Vorshlag, we don't make the car you drive. We make the car you drive faster.
|
|
|
