HP vs. Torque and the System

[Knestis]

The RX8 and V8 conversations in the ITR thread has me thinking again about the "not enough torque to tighten my lugnuts" issue that some cars face, and the impact of that reality on classing/spec'ing IT cars. I'm wondering if maybe we're missing a bet here...

Dynos actually measure torque, right?

Horsepower is a derived value estimate of "work," that considers torque, RPM (for the time factor), and a constant to chase out issues of units. What about the possibility that the Process should actually consider torque instead of HP, perhaps at particular RPM points, or look at where peak torque is in the rev range and consider not only the peak value but the revs at which it occurs...?

We might be able to get away from the subjectivity that gets injected where torque values are out at the ends of the distribution curve - either high or low, well out beyond the mean, since they may be making their peak torque very low, or very high, in terms of revs.

"Oh, but the math might be complicated!" you say. Well, as it is (since the process considers PEAK HP, at just one point on the rev scale), the influence of the location of the torque peak is completely ignored. So yeah, if we want to do a better job of dealing with outliers in an objective, consistent way, then maybe it's worth the extra complexity.

I might also be very confused since my real understanding here is limited to values and manipulating them, rather than the realities that they reflect...

K

[lateapex911]

I would LOVE for there to be a great little apendage that worked within the process to get a better picture of the conflicting power issues.

lets bounce some different formulas about and see how they work.

How about ITS and the E36 ~ RX-7? Anyone care to weigh in with the dyno numbers?

One thing that will come up here is the actual torque curve on some engines is a real peak, while others have a plateau....

[JohnRW]

I would LOVE for there to be a great little apendage that worked within the process to get a better picture of the conflicting power issues.[/b]

Taken out of context, that statement could be quite disturbing.

[lateapex911]

You know, when I saw the "John RW" as the most recent poster on the thread, I thought, "Oh, good, he'll have some good insight and ideas..."

Sigh..........

[Bill Miller]

I would LOVE for there to be a great little apendage that worked within the process to get a better picture of the conflicting power issues.

lets bounce some different formulas about and see how they work.

How about ITS and the E36 ~ RX-7? Anyone care to weigh in with the dyno numbers?

One thing that will come up here is the actual torque curve on some engines is a real peak, while others have a plateau....
[/b]

LOL, you said 'formula'. :P

[Andy Bettencourt]

One thought process by a very wise man is to use a 'power number' instead of HP.

HP + TQ / 2. (This is using JUST published numbers)

ITR RX-8: 238 + 159 / 2 = 198.5. 198.5 * 1.25 = 248.13. Sets the RX-8 weight at about 2790 before any adders (of which I think there are none).

The E36 325's weight in ITR doesn't move much (if any) if you use 189 / 181 from it's stock specs.

(On edit, this would put the S2000 at about 2765...). Maybe the process fails high HP cars with no torque. But I would never say the process 'failed' the 13B ITS RX-7 as it seems to be right on - it would lose over 100lbs using this method) All food for thought. I will admit that as we get higher in the HP numbers, the possibility (probablity?) for errors will go up.

[GKR_17]

Any single torque value is almost meaningless. What is important is the shape of the power curve. A high torque value usually implies a nice flat power curve, but there are clearly a lot of variables involved. Also, good transmission ratios can make up for a peaky power curve (usually low torque motors). If you want to improve this aspect of the process, you will need a full dyno plot for every car and incorporate the transmission data, certainly not an easy task. I have a program that will take several vehicle parameters (dyno plot, aero data, weight, as well as transmission and final drive ratios) to run a virtual acceleration test, but as with anything involving computers, it's only as good as the input data. I wouldn't dare try to class cars with it.

Grafton

[Knestis]

...so the area under the curve.

It seems like a modern dyno could whip that out, no problemo. It wouldn't even have to do calculus to make it happen, since taking a few hundred thousand data points between 0 and 8000rpm wouldn't even make it sweat.

K

EDIT - FWIW, NASA uses an average of peak HP and torque for some of its classes (e.g., GTS) that class cars based on dyno results and race weight. If nothing else, it splits the error that might be imposed by using just one value or the other. Or at least makes it not much worse, if the cards fall that way.

[Ron Earp]

Grafton and Kirk, I was thinking the same thing too If you had the area under the torque curve you'd have an excellent handle on classing a car with respect to engine output. Be extremely hard to do though since you'd need an agreed upon standard dyno type, prep, and so forth.

[JohnRW]

"Area under the curve" was what I was mulling over, before Andy scared me with his "appendage" notion. Yikes.

Given some time, us geeks can come up with some method of quantifying a "factor" for between "area under the curve" and how "# of gears and their inter-related ratios" impact a "cumulative area under the curve", from idle thru red line in top gear.

It's not just how big, flat and fat that torque curve is, but also how ratios will allow you to inhabit the big, flat fat part of that curve.

[chuck baader]

CRAP Shades of NASA GTS where the average hp and tq to get an artificial number...moved my car to run with ITS and ITR cars.....el toro po po

Gentlemen...the only measure of straight line performance is the hp or tq applied to the ground. If you are going to develop a formula, you must include the gear ratios because, and I use my car as an example, straight tq numbers are meaningless. Example: Mr. Amy's car makes 160hp, uses (for example) a 4.00 rear gear. Multiply the two and you get 640 applied hp. My car makes 140hp and uses a 3.73 gear(because my red line is 5600) so my applied hp is 522. (This is of course assuming equal top gear and tire sizes.) Now use those numbers to arrive at recommended weights and see where those two cars fall.

My point is, NASA went down the slippery slope and was blinded by cranial rectosis....it takes a very expensive all out build to get to the hp/wt minimum of each class.

Chuck Baader

[Andy Bettencourt]

"Area under the curve" was what I was mulling over, before Andy scared me with his "appendage" notion. Yikes.


[/b]

???

CRAP Shades of NASA GTS where the average hp and tq to get an artificial number...moved my car to run with ITS and ITR cars.....el toro po po

Gentlemen...the only measure of straight line performance is the hp or tq applied to the ground. If you are going to develop a formula, you must include the gear ratios because, and I use my car as an example, straight tq numbers are meaningless. Example: Mr. Amy's car makes 160hp, uses (for example) a 4.00 rear gear. Multiply the two and you get 640 applied hp. My car makes 140hp and uses a 3.73 gear(because my red line is 5600) so my applied hp is 522. (This is of course assuming equal top gear and tire sizes.) Now use those numbers to arrive at recommended weights and see where those two cars fall.

My point is, NASA went down the slippery slope and was blinded by cranial rectosis....it takes a very expensive all out build to get to the hp/wt minimum of each class.

Chuck Baader [/b]

Gonna be hard to use a 'free' item (rear gear) in any calculation.

[JeffYoung]

When working on teh Mustang proposal, Ron ran a very interesting linear regression that I initially thought to be completely irrelevant -- it showed where on a predicted weight linear graph based on hp a variety of cars fell.

The linear progression worked perfectly for ITR cars using weight. It "proved" the process correct.

The linear progression is ALL OVER THE PLACE for torque for R cars.

To me, it showed that the process doesn't adequately account for torque and that the subjective 100 lbs here, or 50 lbs there is failing us some.

Ron, post those regressions and let's discuss them. The bottom line is I think Kirk is right. We need to be very careful to avoid a torque monster slipping through the cracks in R and being an overdog.

[JoshS]

FWIW, NASA uses an average of peak HP and torque for some of its classes (e.g., GTS) that class cars based on dyno results and race weight. If nothing else, it splits the error that might be imposed by using just one value or the other. Or at least makes it not much worse, if the cards fall that way.
[/b]

They only do that when torque is greater than HP. Otherwise torque is ignored.

[Bill Miller]

For a long time, I've said that peak numbers don't tell the whole story. People have talked about how the open ECU rule isn't going to really increase peak numbers. What it is going to do, is change the shape of the curves, and the subsequent area under those curves.

It would be interesting to see how it would play out w/ published, stock data. While I haven't picked up a copy in years, IIRC, C&D, MT, etc. used to have power plots for cars that they did reviews on. Would be a decent place to start.

At the very least, I think both torque and hp need to be factored in.

Andy,

Maybe there should be an adder for cars that rev significantly higher than most (e.g. RX7/8, S2000)?

[Ron Earp]

The linear progression is ALL OVER THE PLACE for torque for R cars.
[/b]

Now now, the process is working pretty good.......Linear Regression, Linear Regression!! Stat 101!!

While I haven't picked up a copy in years, IIRC, C&D, MT, etc. used to have power plots for cars that they did reviews on. Would be a decent place to start.
[/b]

Whew, Bill, don't date yourself! C&D etc. gave that up many, many years ago. I think the only magazine that actually dynos the test subjects is Cycle World. I've always found that refreshing and it cuts down on the BS claims from the bike makers. they know they'll be rolling bikes on the CW dynos for tests.



The R squared for the first plot was accidentally left off. It is 0.68, which in the hard sciences isn't so great, but in the humanities is probably a pretty good fit and model. 0.07 in the second plot is poor.

And, not all the R cars are on this plot. I only went for cars near the 3L mark and that were likely to be on track in a short time. In part of writing my Mustang proposal I wanted cars that had somewhat similar torque figures, therefore, the Type R, Celica, S2000, and cars like that I didn't use here. These were more of the heavy weights in R that were likely to see some track time.

[Quickshoe]

Area under the curve with the start and end rpm points to be defined as the RPM at which peak HP occurs less the rpm drop between the closest two ratio gears (3rd-4th in many cars, 4th-5th in some) and the RPM at which peak HP occurs.

The issue then becomes what engines have broad power curves with close ratio transmissions, and which ones have peaky power curves with wide ratio transmissions relative to more ideal situations?

[Bill Miller]

Whew, Bill, don't date yourself! 018.gif C&D etc. gave that up many, many years ago. I think the only magazine that actually dynos the test subjects is Cycle World.[/b]

Told you I hadn't picked up a copy in years! :P

And I guess I am an old fart. I still get a little freaked out by the fact that the 80's were 20+ years ago!

[JeffYoung]

Tell me about it. I practice law with kids BORN in the 80s. That be scary sheet man.

[GKR_17]

Gentlemen...the only measure of straight line performance is the hp or tq applied to the ground. If you are going to develop a formula, you must include the gear ratios because, and I use my car as an example, straight tq numbers are meaningless. Example: Mr. Amy's car makes 160hp, uses (for example) a 4.00 rear gear. Multiply the two and you get 640 applied hp. My car makes 140hp and uses a 3.73 gear(because my red line is 5600) so my applied hp is 522. (This is of course assuming equal top gear and tire sizes.) Now use those numbers to arrive at recommended weights and see where those two cars fall.
[/b]

That's not quite right. Available torque is changed by gearing, but the power stays the same (minus drivetrain losses). Power to the ground is what matters, not torque. Gearing is extremely important however since it dictates where you'll be on the power curve in the next gear. A car that only drops 1000 RPM when shifting from 4th to 5th (or whatever gear) has a huge advantage over one that drops 2000 RPM. My point is, gearing is just as important as the power curve, and any single torque value does not fully describe that curve.

Grafton