biker

Sheesh, I wondered how people got those high Post numbers.

Who cares? Any gains (if there are any) from better gearing are "negligible" meaning any gains would be lost to varying testing methodologies and drivers. I mean you have car rags publishing 0-60 times that vary by more than 1/2 sec so a gain of .1 or .2 sec is meanigless, it's within the error of the measurement. Now if it made you "feel" better (faster) that's another matter.

Buff-Daddy

Just for clarification, it wasn't me, but I do remember reading that thread.

TinkySD

Gearing makes all the difference. Stock tsx first gear multiplies engine torque about 15.5 times to the groud..say you made first gear more similar to that in the auto which is just under a 12 times multiplication. ARe you saying that 500bft+ of torque being delivered to the ground is negligible? 25% less drive power through the entire gear has a minmal effect?

rb1

I think you still might have to simulate this. If you did what you say above, you'd get to stay in 1st gear about 30% longer, while the stock TSX has shifted to 2nd and is now getting a (say) 7.5 multiplier.

Still though, I'd tend to agree that shorter gearing makes a car faster overall as it increases the average delivered horsepower per the earlier discussions.

TinkySD

That's why i use cartest 2000!! Awesome program. The caveat of course is the ability to deliver the power to the ground with traction. If you can't then it's wasted energy(per the beginning of this thread)

rb1

While true in theory, losses due to friction are somewhat higher in current CVTs than in an MT and this offsets the advantage somewhat (basically, at least in today's CVTs, you have a belt-driven car ).

VW/Audi, for example, claims identical 0-60 and fuel economy numbers for the A4 CVT and A4 MT.

TinkySD

CVTs love high rpm/horsepower engines. It let's them start off witha really highly multiplied gear and go from there. My testing was done with cartest which may not be entirely accurate but it indicated a cvt would be slightly quicker than a 6mt. It also nailed down the times for the a4 cvt and a4 mt according to the factory. The new odyssey in japan is going to be equipped with a CVT and the k24 engine...I would love to see that dombo brought over here for the tsx.

drchaos

"d^2E/dt^2 = F dv/dt

which would indicate that v^2 is proportional to the double integral of F (torque)."

that's torque to the ground, which you maximize by having maximum average HP and proper gearing. it's not engine torque.

fdl

thanks again drchaos

TinkySD

== you want maximum area under the curve which is what most of us have been saying.

dnb

Oh, sure ...

You derive something with "power" at the end and you conclude with "maximum HP."

I derive something with torque at the end and you conclude with "maximum HP."

My question for you now is this: Forget about the first gear ratio. Will adding a "Type R" sticker increase horsepower, or will it increase torque?

TinkySD

the key term he used was average horsepower which has nothing to do with peak horsepower but instead is a measurement of area under the curve. It's why I thought it really didn't tell us much as I think most of us know you ant maximum area under the hp curve...but there are several ways to get there.

drchaos

You want to maximize the integral of the power output over time.


Since there is a local maximum in power with RPM, then ideally you would want to upshift at RPMS a little bit above this so it ends up a little bit below this.

Now since you maximize power with time, you have to be able to figure out what P(t) to use. If you had a CVT at always optimal ratio, this would be a constant: set the engine speed at maximum HP. With fixed gears there is a bit more complicated optimization problem and the shape of the curve with RPM will come into play.

First gear ratios of course will control how fast you can get up to the power peak as well and therefore that will also change the integral of the power with time.

Jab31169

You know what this thread should be renamed:

Nerds with calculators!! Beware!!

You guys are killing me!!

dnb

What's interesting is that you want to do the above while trying to minimize time (i.e., the 0 -- 60 time). So

t(0-60) = integral_from_0_to_60 of F(Power, v,t) dv

where F is a can of worms.

Note to mods: vBulletin should be able to process TeX so we can have nice mathematical notation in our discussions. What's a nerd to do without his partial derivatives and integrals?

dnb, who used to write in TeX (LaTeX is for wimps)