Jean

Cross post from 993TT board.

This topic is not discussed as much as it should and I was hardly able to get any information when I asked. I have run a search on Rennlist and although there is some data out there, nothing really comprehensive. I read claims about the Carrera ROW gearbox how better it is than the US gearbox, and how it "feels" like xx hp more etc.. But what are the FACTS?

Now that I have the chassis and engine sorted out, it is time for gearbox choices, to get the last fraction of a second or inch possible.

I have run over 200 computer simulations (993TT) off all possible ratios and then tried to find a Porsche gearbox that would be the closest to the result so as to try to reduce costs. In my search, I kept the final diff ratio at 3.44 and as much as possible stuck with original Porsche's two only options for the first gear in 993s, 3.15 and 3.82. I will not bore you with details and numbers as I have hundreds.

Obviously a computer simulation is not the reality, but it works when you are comparing between different gearboxes with the same engine base. Also, a N/A engine might yield different results as the torque curve is so dramatically different.

Engine and gearbox builders do not have time to answer a zillion questions nor run hundreds of simulations, they will build it based on their experience, and they typically deal with race cars, on which the powerband is quite different.

Some difficult questions remain to be answered and I am puzzled by what is the optimum setup and the balance between 0-whatever and acceleration between gears..

Objectives:
1) Best acceleration possible 10-150 mph, rolling quarter mile runs, and fast track.
2) Solid gearbox to withstand 800NM of torque
3) Daily drivability

Some of the standard gear ratios I have analyzed are the following:
Carrera '96 USA (G50/20)
Carrera' 96 ROW (G50/21)
Stock 993TT (G64/51)
RUF 1997 CTR2
993GT2 (G50/54)
993 RSR 3.8 (G50/34)
993 Clubsport (G50/32)

The results are surprising, as I always thought that shorter gears are faster in straight line acceleration, they are not. Longer gears tend to improve the times related to reaching a certain distance, both in the quarter mile as a reference, and 0.625 miles (1000Mtrs.) where speeds reach around 175mph with the engine I am testing.

The winner in reaching both distances was the G50/20 Gearbox (infamous among the Carrera owners!) together with RUF, and both have the longest ratios.

0-60 mph , they are all quite close, with the stock gearbox beating and the others close behind. Difference between 1st. and 6th. is around 5 ft.
0-125 mph, again G50/20 beats everybody else, leaving the 6th. ranked, by around 20 ft. behind.

What is interesting is the speed between gears such as 30-50 mph, 50-70 and 50-90 mph in the higher gears , where the RUF is clearly behind, while the G50/20 is not too far from the 1st. (The Carrera Row and Stock 993TT).

From the above, I should be concluding that for my objective, the Carrera USA gearbox offers the best performance, and reaching top speed of around 200 mph, which can be taken to 220mph with a longer 6th. gear like the RUF.

The only caveat being the strength of the gearbox, which would need to have a stronger shaft and maybe steel synchros. All these tests were performed with the same engine and car specs (weight, tires, torque curve etc..)

Anyone wonder why RUFs are so quick? Other than a great engine, they do know how to get the best of a gearbox, great acceleration with little compromise for performance between gears.

Am I looking in the wrong direction and is there something important that I am missing in making my choice? I could not identify RPM drops between gears, but since the acceleration numbers are there, the drops do not mean much to me.

Let's hear the experts and also the less experts, share their thoughts and opinions!

graham_mitchell

Are you including the time taken to shift gears in your simulation? If so, how much time?

Jean

Graham

The time between shifts is 0.5 seconds and a gear engage time of 0.2 sec. All the scenarios are based on the same assumptions so the time does not really make a difference for comparison purposes. This is the time it would take an average/good driver to shift.

graham_mitchell

ok, but it does make difference if you are comparing two gearboxes to a certain speed and there is a different number of changes involved.

SundayDriver

And that is the defining aspect of you model. The physics tells us that with zero shift time, the shorter gears WILL be faster on straight line acceleration. The time you gain with shorter gears will be offset with more shifts required, and the time losses. Change the shift times and you will change the answers.

When you add the track to that equation, you have much more to consider. You want to hit redline in the tallest gear on a particular track and want to choose intermediate gears to ensure you are in the power band exiting corners and try to eliminate shifts, if possible. The answer changes for each track.

Simulation software can give you the answer for one set of conditions, but is not going to tell you what is best for straight-line, quarter mile, various tracks and street driveability. Too many variables for simple software to handle.

TheOtherEric

Your conclusions generally sound about right. But I'm wondering about your objectives; it sounds like you're most interested in street/drag racing. If that's the case, I'd agree that the US gearbox is probably your best bet. But if you want to improve lap times, I would contend you should modify the objectives. I would argue that X-Y mph runs are NOT the way to evaluate whether one gearbox is "faster" than another. Here's an example: the US gearbox will get you to 60mph faster than the others because you can hit 60 in 2nd gear, where you must shift to 3rd in shorter gearboxes like G50/21. However, in a 20-100 run, the G50/21 is faster (I think). So your conclusion depends mostly on your X and Y speeds and how many times you have to shift to get there. In short, it's an irrelevant measure.

The reason that shorter-geared boxes can give faster lap times is that you have more choices in keeping the RPMs in the peak torque range. For a given car and racetrack, there ARE optimal gear choices. The shorter-geared G50/21 is apparently a lot closer to that optimum setup on most tracks than is the tall-geared US box.

You may want to sit down for a while with the torque and gear curves on your car to help determine what setup would work for you.

Jean

Graham,

I understand what you mean. Thanks.

Mark, lucky to have you here! I understand your comments and of course it makes sense if there is no shift time involved in certain acceleration runs. I tried to keep this as close to reality by choosing shift times relevant to an average to good driver. I have also taken different shift RPMs for different gears. My redline is set at 7,200 RPMs and my peak torque will be almost flat between 4,500 and 6,500RPMs, and my peak hp almost flat between 5,000 and 7,000 with a peak at 6,000RPMs, so whenever I shift I am maintaining my revs at good levels, assuming 2000 or so RPM drops.

Concerning the tracks, I agree, I simulated on the Nurburgring, which has a mixture of slow, fast and very different corners, and the results were not substantially different. The tracks I will run are fast tracks (Bahrain and Dubai).

I will change shift times and see what happens. Thanks for your input.

TheOtherEric

If your torque doesn't drop after a shift (which is what you are saying) regardless of the gearbox, then it doesn't matter which gearbox you chose, now does it?

i.e. if your torque curve is flat over your operating RPM range, then gearing doesn't matter.

Jean

Eric,

The simulations were done based on my actual expected dyno charts, torque and hp. The G50/20 beats all the other in any acceleration times up to top speed.

Just to clarify, I understand obviously that a computer simulation is not real life, but it gives an interesting perspective worth exploring. Once on the track, all the optimum shift points etc.. change from the theory, and I am not too consistent in my driving either

Acceleration coming out of the corner will be reflected in the times seen between gears (35-50, 50-75, 50-90mph etc..), 2nd. 3rd. and 4th.etc.. were the G50/20 also did very well.

Edit: I am not sure I understand your second post. I believe that even if your torque is the same, the gearbox will indeed make a difference. This is Flywheel torque. The dyno is performed on an engine dyno.

Thanks

TheOtherEric

Hmm, you lost me. If an engine's torque output is flat across the entire operating RPM range, then the fastest setup would simply be the one that minimizes the number of shifts, no? In this case, a different gear ratio wouldn't help because you'd still have the same amount of torque delivered.

p.s. I don't purport to be an expert, I just find this topic fascinating due to the extreme hype.

SundayDriver

Thanks. Clearly what you need to do, Jean, is buy a sequential gearbox and a SWOL (Shift Without Lift) system. That will take care of those long shift times.

Jean

Unless I am totally confused (which is very likely) same engine torque will throw different torque at the wheels depending on the gear ratio, wheel size etc.. Reason why with chassis dynos you tend to do your runs at the closest to a 1:1 gear ratio.

Jean

Mark, and you think I did not consider them? Until Greg told me how much they cost!

SundayDriver

Consider this (but I have not thought this out fully so I may be wrong):

What you care about is torque at the rear wheels. An engine that makes 60% of the torque of another, but at double the RPM is going to be stronger when you gear reduce it to the same wheel speed. In my car, for example, if I push the engine and run an extra 1,000 rpm, I can lower the gear ratio and get better acceleration from the same engine (at the expense of reliability).

I think you pick gears (ignoring shift time) by looking at the HP curve. This is an inverted J with peak HP coming somewhere below redline. When you upshift, you will move from one place on the curve to another spot at lower RPM. I THINK that you want the line connetcing the high rpm to the lower rpm (rpm in each gear just before and after the shift) to be a line paralled to the bottom of the graph.

Where I may be wrong is that the ideal might be maximizing the area under the curve between those rpm points. I would have to think more about it to understand which is better.

In either case, you are optimizing HP for shift points because that is directly related to rear wheel torque. You would only need to know the rpm drop between gears and that would be much simpler to model than using engine torque where you would have to deal with more conversions and gear ratios to get the same answer.

graham_mitchell

That doesn't sound right to me. Power is proportional to Torque x revs. So even with a flat torque curve, you will still develop more power with increased revs.