[Mr. 5]

Many have talked about the perfect RPM to shift and I thought I would put something out there to explain about how gear ratios have a lot to do with dictating when to shift.

Some people (including myself) have thought that when the horse power starts to drop then that is when one should start shifting, but it has more to do with that.
Horse Power is actually derived from torque so we'll focus on just the torque.

This is what you need to know:

1. The Torque curve throughout the RPM band with your car. Someone else's curve may not necessarily be the same as yours so their shifting points may be different.
2. The gear ratios of each gear.

I'm going to use a Torque curve from my GIAC Stage 1 flash.
I used the Dynograph to determine the torque at each RPM that I randomly chose. I tried to pick increments of 250 RPM.



RPM TRQ
2500 327.7
2735 333.15
3009 360.32
3264 357.89
3498 359.51
3767 363.05
4008 367.43
4270 368.73
4504 369.67
4737 369.7
5006 367.14
5290 354.53
5502 347.33
5743 334.45
5998 320.76
6246 307.97
6501 291.56
6763 271.74
6997 229.74

Since I have an MT, I'm going to use the MT gear ratios.

1st--4.06
2nd--2.4
3rd--1.58
4th--1.19
5th--1
6th--.87

Now, take the torque numbers and multiply the numbers by the ratio of each gear.

1st gear numbers will look like this:

1330.462
1352.589
1462.8992
1453.0334
1459.6106
1473.983
1491.7658
1497.0438
1500.8602
1500.982
1490.5884
1439.3918
1410.1598
1357.867
1302.2856
1250.3582
1183.7336
1103.2644
932.7444

2nd gear numbers will look like this:

786.48
799.56
864.768
858.936
862.824
871.32
881.832
884.952
887.208
887.28
881.136
850.872
833.592
802.68
769.824
739.128
699.744
652.176
551.376

Now, graph each curve. I used excel.
Get the peak number of the gear you're shifting into (in this case it's 2nd gear) and draw or make a horizontal line ratio numbers.
The peak number is 887.28.

When 1st gear numbers, 2nd gear numbers, and 2nd peak number is graphed, it looks like this:



Notice that even at redline (7k RPM) you will never reach the amount of torque that 2nd gear puts out so there is no need to shift early.

When doing the same thing with 2-3 shift, you’ll see that at approximately 6800 RPM is where the peak of the next gear torque hit’s the curve. This is the optimum shifting point.



Do this with each gear and you’ll find the following:

3-4 Shift 6600 RPM



4-5 Shift 6200 RPM



5-6 Shift 6000 RPM



So, to summarize, the best ultimate shifting point for this car making this amount of power is the following:

1-2...7k
2-3...6800
3-4...6600
4-5...6200
5-6...6k

[Nedim]

Ah man, thanks for the info. Should i follow the same procedure for auto?

[Doug007]

Great post!

I did a similar excel analysis using stock torque curves from dynos on the internet. Having your own torque curve is a big advantage since there will be variation.

The only detail that must be considered is the rpm in the next gear after the shift. That is, if you shift at 7,000 rpm in 1st, you'll be at 4,140 rpm in 2nd. The equation is shift rpm * next gear ratio/current gear ratio. (7000*2.4/4.16 =4138). Of course, the next gear rpm moves up as you're in higher gears.

The breakdown for 7,000rpm shifts is:
1:2 4,138
2:3 4,608
3:4 5,272
4:5 5,882
5:6 6,090

I'm not sure if this skews any of your shift points without sitting here and doing the math but it is something else to consider besides peak torque in the next gear....

[Mr. 5]

Originally Posted by Doug007

Great post!

I did a similar excel analysis using stock torque curves from dynos on the internet. Having your own torque curve is a big advantage since there will be variation.

The only detail that must be considered is the rpm in the next gear after the shift. That is, if you shift at 7,000 rpm in 1st, you'll be at 4,140 rpm in 2nd. The equation is shift rpm * next gear ratio/current gear ratio. (7000*2.4/4.16 =4138). Of course, the next gear rpm moves up as you're in higher gears.

The breakdown for 7,000rpm shifts is:
1:2 4,138
2:3 4,608
3:4 5,272
4:5 5,882
5:6 6,090

I'm not sure if this skews any of your shift points without sitting here and doing the math but it is something else to consider besides peak torque in the next gear....

Good point but based on the shifting points on the above example will be the following:

1-2 4137
2-3 4476
3-4 4970
4-5 5210
5-6 5220

[themyst]

Mr. 5 - excellent post. My shift points are very similar to this actually, except I don't wind out any gear to redline. I suspected due to the taller nature of 3rd gear compared to 2nd that winding out 2nd gear would be optimal, based on your findings, turns out I was right.

[zsapphire7]

yup!! I figured this out over a year ago . The reason why it isnt at redline is because our torque curve drops a lot in the upper end.

[GAaaAR]

Might be a stupid question, but can you explain why its optimal to shift once the previous gear's graph line drops below the next gears peak? I always thought its about the area under the curve? Even though it passed the optimal peak, its still producing more torque than other areas on the curve?

[boostd92]

You also have to consider shift time, if you're 6MT. If reducing the RPM you shift at in every gear causes you to have to use an extra gear, it's not worth it.

For example, in my LS1 RX7, if I shift at redline in every gear, I can go through the traps just at the very top of 4th gear. If I shift a little earlier to take advantage of a handful more torque, I have to shift to 5th gear at the end.

The few ft/lbs of torque does not make up for the .1-.2s it takes to shift from 4th to 5th - I've tested it.

Sorry if this does not apply to N54/6MT, the gear ratios may allow more in that example, but it may be a worthwhile point for people thinking about upgrading to a higher ratio M3 rear end, or similar (My RX7 has a 4.10).

[cstavaru]

Originally Posted by GAaaAR

Might be a stupid question, but can you explain why its optimal to shift once the previous gear's graph line drops below the next gears peak? I always thought its about the area under the curve? Even though it passed the optimal peak, its still producing more torque than other areas on the curve?

I agree that this may be the correct way to measure.

[Doug007]

Originally Posted by GAaaAR

Might be a stupid question, but can you explain why its optimal to shift once the previous gear's graph line drops below the next gears peak? I always thought its about the area under the curve? Even though it passed the optimal peak, its still producing more torque than other areas on the curve?

When people talk about area under the torque curve it's usually comparing a very peaky torque/power curve to a flatter one. One engine might make more peak torque, but the AVERAGE torque over a run (0-60, 1/4 mile, etc.) may very well be lower than an engine with a lower peak but flatter curve.

As for shift points, it's all about shifting when the wheel torque in the next gear will be more than your current gear. This is why Mr. 5 used the peak torque in the next gear.

In the higher gears, say 5-6 shift, if you wait until after 6000 rpm you will have less torque to the wheels in 5th than you would had you shifted to 6th for the same car speed.

The goal is to maximize the average wheel torque over the run...

[edo]

Good thread but I would rely on practical tests more.
My Pbox times and logs show it is best to shift at 6.2-6.4k rpm.

Moreover, it depends on a tune, I believe. Shiv wrote some time ago it was best to shift at 6.2-6.4k rpm as there's more boost after such shift.

[Mr. 5]

Originally Posted by edo

Good thread but I would rely on practical tests more.
My Pbox times and logs show it is best to shift at 6.2-6.4k rpm.

Moreover, it depends on a tune, I believe. Shiv wrote some time ago it was best to shift at 6.2-6.4k rpm as there's more boost after such shift.

Remember what I said in the intial post though...

1. The Torque curve throughout the RPM band with your car. Someone else's curve may not necessarily be the same as yours so their shifting points may be different.

I am only providing a tool for the community to determine their optimal shifting point.
The torque curve on a procede may be different than the torque curve of the GIAC tune so this is why one would need to do use this tool with their curve.

Another example would be comparing my GIAC Race tune with this GIAC stage 1 tune. The shifting points of my race tune is different than the stage 1 tune.

[edo]

Originally Posted by Mr. 5

Remember what I said in the intial post though...

1. The Torque curve throughout the RPM band with your car. Someone else's curve may not necessarily be the same as yours so their shifting points may be different.

I am only providing a tool for the community to determine their optimal shifting point.
The torque curve on a procede may be different than the torque curve of the GIAC tune so this is why one would need to do use this tool with their curve.

Another example would be comparing my GIAC Race tune with this GIAC stage 1 tune. The shifting points of my race tune is different than the stage 1 tune.

I do understand that, but as I wrote before I would rely on practical tests to determine the best (for acceleration) shifting point rather than on theoretical calculations.
I do believe one can get a bit different results.

[Mr. 5]

Originally Posted by Doug007

When people talk about area under the torque curve it's usually comparing a very peaky torque/power curve to a flatter one. One engine might make more peak torque, but the AVERAGE torque over a run (0-60, 1/4 mile, etc.) may very well be lower than an engine with a lower peak but flatter curve.

As for shift points, it's all about shifting when the wheel torque in the next gear will be more than your current gear. This is why Mr. 5 used the peak torque in the next gear.

In the higher gears, say 5-6 shift, if you wait until after 6000 rpm you will have less torque to the wheels in 5th than you would had you shifted to 6th for the same car speed.

The goal is to maximize the average wheel torque over the run...

Great points.
Remember what gears do. Gears multiply the amount of torque from the engine and deliver that torque to the ground.
The amount of actual torque being applied to the ground by 1st gear is so much greater than 2nd. Once the torque of the next gear is higher than the present gear, there is no need to stay in the present gear any longer.

[Mr. 5]

Originally Posted by edo

I do understand that, but as I wrote before I would rely on practical tests to determine the best (for acceleration) shifting point rather than on theoretical calculations.
I do believe one can get a bit different results.

Do you have any recent dynos?
Why don't you use the tool so you can compare the theoretical values to real world values? They very well may be the same.
I'm interested in what you find out.

[Ron1n]

This is a rehash of an old thread, actually. The major difference is that you provide a more granular approach, and the original thread spoke in general terms.

[O-cha]

I'm pretty sure I explained this to you like a year ago, you're just getting it now?

For "practical tests" you'll need a G meter for some sort and flat road. Accelerate in each gear to redline and simply compare acceleration #s with the next gears.

[Mr. 5]

Originally Posted by O-cha

I'm pretty sure I explained this to you like a year ago, you're just getting it now?

For "practical tests" you'll need a G meter for some sort and flat road. Accelerate in each gear to redline and simply compare acceleration #s with the next gears.

Yes, I am.
You were right. I didn't understand your graphs back then because I didn't know what each axis was and I didn't know what each curve was.
I remember that I said that I shifted early in 1st and 2nd and I wasn't convinced that going all the way to redline was the best but now I am.

[michaelthepsycho]

This thread is one of those few that make forums worthwhile. Now... If there was only a way to implement this type of user-uploaded shift-logic into a tune to control the TCU for various different torque curves. I'll reckon maybe 50~100 revolutions need to be added to account for shift-delay on a Steptronic (or any other automatic for that matter).

[Clap135]

Like I stated in you FI thread, this is HORRIBLY wrong. Dynojet load is not real load, each gear will make a different curve/number. Your car will no dyno the same in 1st as it does in 2nd or 3rd/4th/5th ect.

[Ironring Racing]

Good first try, and a good exercise to go through the math, but as has been pointed out:

1. Power/"dyno graph" in each gear will be different, especially in this car with the gear specific OEM (and aftermarket) mapping
2. Doesn't take into account the rpm/power you enter the next gear at. By your calcs you should shift at 6k in 5th. Look at the low RPM that puts you in for 6th. Think of it as "power under the curve". You want to maximize both the exit gear shift rpm power, AND the RPM/Power that you enter the next gear at.

O-cha has it, acceleration (G-force) graphs for each gear are the best real world data. Maximize the acceleration forces across your shifts. i.e. you want to maximize the combined acceleration from both the gear you're leaving, and the gear you're entering. Could also be done with dyno graphs, but you'd need them for each gear, and you'd be optimizing for best overall power from the gear you're shifting out of, and the gear you're shifting into.

[Mr. 5]

Originally Posted by Ironring Racing

Good first try, and a good exercise to go through the math, but as has been pointed out:

1. Power/"dyno graph" in each gear will be different, especially in this car with the gear specific OEM (and aftermarket) mapping
2. Doesn't take into account the rpm/power you enter the next gear at. By your calcs you should shift at 6k in 5th. Look at the low RPM that puts you in for 6th. Think of it as "power under the curve". You want to maximize both the exit gear shift rpm power, AND the RPM/Power that you enter the next gear at.

O-cha has it, acceleration (G-force) graphs for each gear are the best real world data. Maximize the acceleration forces across your shifts. i.e. you want to maximize the combined acceleration from both the gear you're leaving, and the gear you're entering. Could also be done with dyno graphs, but you'd need them for each gear, and you'd be optimizing for best overall power from the gear you're shifting out of, and the gear you're shifting into.

I just did some pulls on the way home from work and compared the Gs at redline to the Gs at the same speed in the next gear and it matches what I have said. The acceleration up to redline in 1st (for example) is greater than being in second at the same speed. In other words, if I would have shifted before redline, I would be losing acceleration.
I'll dive more into this and look at the power under the curve but I have a feeling that the results are going to be extremely similar.