Holy ****, I feel very unintelligent right now. My head was spinning after reading that. But after I regained consciousness, and read it again, it made perfect sense, even to a dumb *** like me.

 

It will be back in a day or two... you are welcome to it if you like. Kevin

 

LOL seriously, amazing post

Kevin i truely appreciate it, let me know!

 

Dayum, I say dayum!

 

Hi Andrew

Good to see you posting. Very interesting explanation, what I like the most about it is how clearly it explains in words what one sees on the ground. Thanks! I just wanted to post a few questions or comments.

I fully agree and never debated it, formulas to predict HP (both ET and trap speed) are never accurate for obvious reasons of different car dynamics, and a thousand other variables. With ET as you said, you are adding the traction loss variable on top of it so it certainly is more bound to error.
However to be fair, Hale did use real life examples of a database of more than 50 very different cars from Actual R&T tests and actual physical weight of the cars so the statistical confidence is quite well secured. http://www.stealth316.com/misc/performance_lrt.xls
His formulas are not always right but one can take them and build on them, the bulk of the research was already done.. For instance, an adjustment of 6.5% on his formulas show the following results for Porsche cars with similar dynamics.
Here is an example of the ET to trap speed relationship based on the top 14 or 15 listed times on this board. Using his formulas and the realtionship between ET and trap speed @ ET=1450/Trap speed you get a 1-2% margin of error, which is very good I believe. Of course there are exceptions but a link between ET and trap speed is certainly proven, not a straightforward relationship, and not very meaningful, but it is there and could be the difference in a couple or more mph.


Ouchh! I certainly do not agree with this. Street tires have traction of 0.5 on wet surfaces, on dry pavement up to 1Gs are not uncommon, on a drag strip, a 1.1 G should be feasible.
One key parameter that you are not mentioning is the weight transfer during launch, which is clearly THE biggest contributor to traction. A typical Porsche rear engined car with stock height can easily see 85% of its front weight, 5-600lbs more on the rear tires upon launching. This increases considerably the traction available beyond the 1G. obviously and ever more the higher the center of gravity of that car.
To be less theoretical and more practical, here is an example of a datalog from 3 different cars and 4 different runs. It shows the first 70feet or so of a standing acceleration and how the long Gs are impacted. All of the cars had above 0.8Gs maximum, the 996TT is a 4WD car and had 0.98G. Guess which one had the slowest 60ft? The one with the highest Gs and best traction. Despite the better traction the blue car had the slowest 60ft time of 2.8 seconds.

Where you are right is upon the first ****** off the line, all the cars hover around 0.6G and then quickly recover increasingly the Gs by 40-50ft. The 2 GT2 lines (green and red) show the difference between a 35+ PSI tire pressure and a 15PSI tire pressure on a bone stock GT2. Notice there is between 10-15% additional grip with the tires at lower pressure.



Is there traction loss? You are 100% right, of course there is, huge traction loss but the threshold is not 0.6 G as shown earlier.. Concerning the thrust generated by th ecar, I did a calculation of what sort of drivetrain torque is being put to the ground by the 993TT, based on accurate engine dyno torque data, actual weight, tire circumference, weight distribution etc... from there I calculated the wheel thrust being put to the gournd in ft.lbs and then arrived to the Gs available by the engine power to the ground by involving th emass of the vehicle. The results are quite interesting. The 993TT car puts down more than 1.5Gs of thrust for 2 seconds! 0.7Gs of those go to the ground, there is a 0.3G or so loss of acceleration potential....This car is the only one that did the run on the street, not the drag strip. The driver could have gotten an incremental 0.3G to the ground in theory. To give you guys an idea, we are talking about 6000lbs (thousand) of wheel torque to the ground once you multiply engine torque by gear ratios, and diff.!!! Anything beyond the redline is traction loss, you can see that the second gear still looses traction at full torque RPMs or so.



Agreed. Trap speed is a better way to estimate HP, trap speed however IS influenced by the ET to a larger or lesser degree, if the terminal speed at the end of the 60ft mark is impacted positively or negatively by the launch and the car was not able to recover the speed loss in the remaining balance of the quartermile which most likely is not the case (especially that we are comparing two different cars).
In the GT2 example above, the run with better grip had a terminal 60ft speed of 1mph faster than the other run, this could, or could have not had a trap speed impact ( I am the only one to know ). More importantly with 2 different drivers, different torque cars, and a substantially different 60ft time and speed, the trap speed of one of them could have been impacted by quite more than 2mph due to the launch.....Which is what I was only saying

Your last statement above is very important, as it also validates why a flying 60-130mph or any other speed vs time test as measured by an accurate device, is more reliable to predict HP than trap speed which has (or could have) launch variables involved in the ending speed number. Provided slope is verified and the GPS device is accurate obviously.

Andrew, excellent write up and clarity! Thanks

Edit: Divexxtreme is there a way to move all these OT posts to another thread out of courtesy to Vivid.
Oh and to remove a few of the bad rep points I received?

 

I agree with Jean that 60 to 130 mph runs are a good way of measuring HP... especially if you have something like a AX 22. you only shifting once maybe twice.. less room for driver error... very accurate...

 

........Done

 

Thanks Scott.

 

I think Jean is on to something with this stuff. I know the 1/4mile is important to us guys over here (even though I'm a dumb brit) but what he's saying is sound. I kinda thought it was a good idea to say "screw" all the theories and just run what ya' brung on the road course, braking, 1/4mile, dyno etc... all in one big event, which is why I organized the European Car shoot out and helped with the Excellence one. Now, I know neither of those panned out as "scientific" as I'd hoped but it was clear on that day with the same drivers which cars could do what. Numbers to me don't tell the whole story

 

I think what Jean is saying is sound as well. Especially the piece about AX-22 or Racelogic measurements from a roll.

Our primary disagreement is based on the accuracy of using Trap Speed/weight to show a car's true HP (as well as roll-on acceleration capability). Jean doesn't believe it to be very accurate...while I believe it's extremely accurate; a method that's already been proven mathematically and empirically.

 

Just something else I would like to add to confuse things even more.How you stage (shallow or deep) can also have an effect on trap speed and e.t.Trap speed is a great indicator for "power under the curve"where a broad power band will trap higher than a peaky one because your average horsepower is higher during the timed run.Great thread guys!

 

Hello Jean,

Thanks for sharing your viewpoints.
I have some comments, in response to yours, that will help explain what I was saying and shed more light on the subject.

It depends on your definition of accuracy. The formulas that I re-wrote for Scott are within 2% for the majority of all vehicles I tested them on. Although there are many variables, a 2% accuracy window justifies their use for most people I would think. The Hale formulas were not this accurate in their native form, for most vehicles, and needed adjustment. In fact, the adjustment needed was to add additional depth to very basic formulas which seemed to be based on a theoretical foundation. The reason I say this is that the points, slope and curvature were far enough off that simplicity and theory were the only reasons, that I could have understood for him, to leave the formulas without correcting them. My comment about a lack of statistical confidence was based on the large variance or std. dev. of the data point's expected values......looking at the performance specs., it was clear that many were wrong. In those cases I went to other sources to obtain correct numbers, but by removing them from Hale's data, his set became too small for statistical confidence at that point imo. This is part of why he may have had difficulty equating an accurate formula, while after re-examining the numbers and obtaining more accurate data and results, accurate formulas or graphical representations could be generated.

That's great that a conversion works for turbo 996 Porsches. My goal was to show that with enough correct data, that a relationship could be made for most cars. I think my formulas show that for most vehicles, not just a specific set. But that's good that you have something that is representative for this group or subset. As I said before, if we could always do the same launch with the same variables/technique etc. then you could relate et to trap speed.......we actually see that with some magazines as my et calculator is fairly good with specific name magazine times. But when you start comparing different magazines, private time slips, factory data etc., the et's start varying quite a bit. But the trap speeds don't. Which is another example of trap speed being more dependent on wt/whp and not on the launch.

Jean, I said .5-.6g force/tire.....that's 1-1.2g on a rwd vehicle....
Obviously that balance changes for awd, although not as much as one would think due to weight transfer limiting front tire load transfer capability.

You're right, which was my point. 1-1.2g (.5-.6/tire) and in this case a total of 1.5g of thrust....and the car is still losing traction in second gear. This is what causes et to be launch dependent, or responsible for a time loss, and not responsible for a trap speed loss. Consider for a second the main point that I illuded to but didn't explain well enough in my previous post. Speed curves over distance look like 1/2 parabolic curves ie. they start off steeply vertical and than begin to shallow and converge at an asymptote as distance passes, with speed on the Y axis and distance on the X axis. One more way of saying this is that the whp of a vehicle is most influential to a car's acceleration as speed increases ie. as a greater percentage of the distance of the 1/4mi. elapses, the launch becomes a smaller percentage of the performance, and drag, and as a result whp, becomes the most significant factor. This favors a confluence or matching of trap speed, or speed at a given distance, for a given wt/whp and a lessening of the effect of the launch. This effect continues to build, throughout the run, obviously, since on a percentage basis the launch distance/total distance traveled, is decreasing at an increasing rate as speed increases while time of launch/time of run is decreasing at a fixed rate as time elapse rate is constant. This causes the et differences at the launch to be maintained while the trap differences at a given point or speed at distance tend to merge. That is your key.
That should keep some people thinking for awhile...

Well here we are agreeing and disagreeing. I agree with you that if the launch is significantly different enough to reduce the car's speed significantly at a substantial distance down the 1/4mi., which would be a huge launch difference, then yes, it will affect the trap speed. But for common launch differences without major, long distance mistakes, the trap speeds for a car will remain the same as et varies due to the launch. This is why for common launch variances, trap is consistently proportional for equal wt/whp as their et's vary. This can especially be seen in magazine to magazine et variances while trap remains very close.

You're absolutely right, a rolling accel. using a large range of speeds gives a very accurate picture, as long as it is done right. Consistency and a normalizing of the variables is key though. The only things I would add to your statement in addition to slope would be to consider winds, temperature, absolute pressure, a graphical analysis to equalize starting points and to make sure that the devices are the same model in addition to being checked for accuracy.
Thanks Jean for your thoughtful comments.

 

Joe that's right.. it does make a difference to the ET and not everone stages the same Me? I stage shallow.. heh.

 

This thread is fun! Reminds me too much of college though (which I guess isn't necessarily a bad thing!

 

Wow! Now I'm really lost So let me get this straight....K16 can make 600HP?