gavriil

Eric wrote:

"Actually, there is a "POINT" that is getting missed by everything I've seen here so far.

Its not the HP or the TORQUE but the area under the curve. If we had a CVT, then what I just said would be totally false, and I'd get my assumption kicked up my rear. A CVT would keep the engine on that 30+ HP sweet spot for the bulk of a 1/4 mile run, and it would help quite a bit more.

Since we only have an Auto tranny with 5-speeds and they are not exactly close ratio (to say the least), the car is not sitting in the "meaty" part of the curve (headers and AEM) for very long.

If you want equivalent power over the whole band, get a:

Super charger -- torque and HP up for the whole curve.

Stroker kit -- torque and HP up for the whole curve.

Weight loss kit -- same force, but less non-rotational mass to push through the whole curve

Lighter wheels/pulleys/etc -- same force, but less rotational mass to spin-up through the whole curve.


If you look at Mike's dyno (and others), the flat part of the torque curve gets extended by about 1000 rpm and the significant HP gains are from 5000+ RPM to 7100 RPM.

With the setup we have, we are only getting an approximate 10% power boost over a very narrow range.

I did notice that Mike got better times from the lightened wheels -- they work through the entire range.

So, given the gearing WE currently have, we shouldn't forget that it is the area under the curve that means the most (Integrals come to mind).

(This is just my opinion -- depending on FWD, # of gears, and type of transmissions, this subject could be debated and argued like crazy...)

This is just some additional food for thought -- please just consider it as such -- no more no less [and I didn't check all the hp graphs today...]"

_________________________________________

Eric I have been trying to find out how to calculate the area UNDER THE FRICKING CURVE with no success. In my opinion, that is how HP and Torque should be calculated. That area is what matters ONLY in my opinion.

You mentioned integrals come to mind. Is that how we calculate that? We have got to ask a math professor or something for those that are still at school...

Needless to say: I agree 100% with the above post of yours and would like to expand on the prinicple. How are we going to do that? Guys?

------------------
Gabriel
CL Type S

CLpower

torque still wins races

------------------
my mom and bestfriend have told me to tell you guys to stop telling me to spend my money; they are tired of my automobiles getting pampered and them getting whats left over which usually consists of change I find in my car by doing things you guys encourage

typeR

the best solution would be IMO to find away to raise redline/fuel cut...with all the nessecary changes liferts spring valve work to make 8K redline 8.1 fuel cut ...then when you shift drop you right back into the sweet spot

------------------
'01 3.2 CL typeS
satin silver metalic
Llumar platinum plus
tint 20% all around

-----------------
'98 ACCORD V-6 sedan
16" coupe wheels
215/55/16 nitto's
20% smoke titanium
black chrome kit

CLpower

acura should of already done that for you. IN vtec cars the reason the tach out so high is so when you shift you will land back into vtec which is your strong part. If you punch it and max out the tach to redline and don't shift into the next gear already into vtec then i would have to say acura screwed you. You would still benefit from raise redline, but i don't think it would put you in any different of a sweet spot.

EricL

Ok, since torque converts to hp, hp to torque [at a given rpm] (yes/no)?

Without using linear regression or other curve "fit" methods, calculating the area under the curve would need to be done with Trapezoid Rule, right or left rectangles, Simpson's Rule, (or pick your numerical method -- there are a ton of them) in a piece-wise manner. As the points along the x-axis get closer and closer the change in calculated area approaches a "limiting" value. When there is sufficient "lack" of change, the area can be considered "close enough".

There are some "automotive" calculator programs that "exist" for taking in dynos with Hp or torque readings (at given rpms) and along with the weight of the car and (CD and frontal area) give someone a "rough" idea how much they gained or lost from an acceleration stand point.

I think DTEW found one of these programs, and they come in more and more advanced versions.

The point I was trying to make is this:

since torque converts back to hp (at a given rpm) and hp converts back to torque, given a similar car (or an Acura CL-S to be exact), if the "HP" peak is very narrow, it only is going to "help" the car at top speed or it is only going to help for a tiny percent of the time.

Example:

If a car (like ours) make 40 more HP, but it only makes it between 6400 and 6800 rpm, then the area increase that one would get compared to the original curve would be very small.

If someone was to make 20 more HP, but over the entire range from 2000 rpm to 6800 rpm, the area increase would be greater and represent a greater "gain".

Finally, before I get bashed to death here. Once the car is moving, the increased area of interest is in the area from shift-to-shift. If the cars shift points never get down below a certain level (once the car has left the line), it would be silly to compare the power in that area.


Finally, if A CVT gets added to the picture, you can toss every thing I just said out, since the CVT has the ability to stay on the HP peak and the engine with the highest peak wins.

I remember saying, "There is more to all of this..."


Notes:

Curve fitting methods:

"LaGrange Polynomial Interpolation" -- simple with limitations

"Cubic Spline Interpolation" -- harder but more accurate.

Here is a link to a "Mathematica Page" on both:

http://www.ma.iup.edu/projects/CalcD...s/splines.html

Notes part 2:

If someone uses one of the interpolation, curve fitting, or nonlinear regression methods to obtain a formula in the form of f(x) = ".....", here is a page that do numerical calcs:

http://147.4.150.5/~matscw/RealWorld.../integral.html

------------------
Silver 2001 CL-S with NAVI

• Mud guards
• Wheel locks
• Toyo T1S 235/45ZR17-97W* Proxies on 17x8" SSR Competition wheels (48lbs less than stock)
• Kevlar/SS Brake lines w/Brembos?
• Comptech headers & sways
• Silver AEM CAI
• 9 coats of Zaino magic

[This message has been edited by EricL (edited 05-01-2001).]

moomaster_99

My head hurts from reading your post!!


------------------
2001 Cl-S Black on Black with Navigation
Black Tint all around (15%~45%)
Weapon-R Dragon Filter
Weapon-R Coil-Overs
Koni Adjustable Struts
Comptech Sway Bars
Comptech Headers (hehehe...I got something for you drchang!!)

4pumpedCL

A(squared) + B(squared) = C (squared)

...that's all I remember from ASU mathmatics.


------------------
2001 CL-P, new project undergoing complete reconstruction. As of now: Custom CAI fabricated by SRD, Polarg M6 bulbs, 4 ProHopper Chrome hydraulic pumps with whammy tanks, 4 ProHopper dumps, 4 Optima marine deep-cycle batteries, 4 switches. Much much more to come...

1997 Accord coupe EX-L. Chrome suspension, Chrome motor, every bolt-on imaginable. One pump, 8 dump hydraulics with accumulators. Leather, Billet grills, 18" Neeper Shadows...etc..etc http://www.artofnoize.com/Cars/aaron.htm

1990 Toyota standard cab. Bodydropped 3". 17" KMC Nemesis', extensive suspension work. 4 wheel airbags.

EricL

BTW -- there is a lazy mans way to do all this without all of the math mumbo jumbo.

I'm sitting here -- laughing.

Just scale a couple of the graphs to the same size -- use your favorite graphics program and be sure to scale two different plots to the same scale and baselines.

Then put in some quad rule graph paper in the printer. Print out the graph in Landscape or Portrait (whatever) and then count the # of squares inside the area.

Now compare the # of squares from graph 2..n

Dirty integral method for HP and/or Torque. Just divide the area by the width, and you have the mean (average) HP or Torque over the curve of interest.

(just a thought...)


------------------
Silver 2001 CL-S with NAVI

• Mud guards
• Wheel locks
• Toyo T1S 235/45ZR17-97W* Proxies on 17x8" SSR Competition wheels (48lbs less than stock)
• Kevlar/SS Brake lines w/Brembos?
• Comptech headers & sways
• Silver AEM CAI
• 9 coats of Zaino magic

fbazakos

ok you said the magic word. i have a numerical analysis class, and lagrangian polynomials are cake, and cubic splines can be done with matlab. email me if you need something solved, i'll do it for you.

ps. you could possibly use divided differnces, bezier curves, least-square approximations, etc ... my final is a week from tomorrow, and as you all know how it goes, you forget that stuff immediately.



------------------
Black 2001 Type S
*Eibach Pro-Kit (Waiting to be installed)
*Xephyr Cold Air Intake
*PIAA 19169 Road Lamps
*Rockford 500a2 [FOR SALE]
*Pioneer 12" @ 400w (x2) [FOR SALE]
*77237-SOK-A02
---Still 2 rims not scratched---

EricL

Ok -- tag your it. When Gavrill comes a knocking, you get to pull out your Matlab.



------------------
Silver 2001 CL-S with NAVI

• Mud guards
• Wheel locks
• Toyo T1S 235/45ZR17-97W* Proxies on 17x8" SSR Competition wheels (48lbs less than stock)
• Kevlar/SS Brake lines w/Brembos?
• Comptech headers & sways
• Silver AEM CAI
• 9 coats of Zaino magic