Login


10252007, 10:16 PM  #67 
Lieutenant
11
Rep 489
Posts 
Here's some simple math I did, and tell me how this is incorrect. I set up an equation where X/295 = 8400/6800 Pretty much we are solving for X. When you solve you get X to equal 364.41 torque. This is what you would need to equal the power of a car making 295 torque with an 8400 rpm. The Lexus makes more than 364.41, so if I am correct, if both cars have gear ratios in which their redline makes them shift at the same mph, the Lexus would be putting more power to the ground. I think ur program is flawed, no way the m3 should have more peak thrust than the lexus in first gear when according to your graphs it seems that they both shift at the same speed.

Appreciate
0

10252007, 10:37 PM  #68  
Lieutenant General
236
Rep 10,257
Posts 
Wrong again
Quote:
Furthemore it is not true that a formula can not beat actual test data. This may sound unbelievable but think about it for a minute. How much variation do we see in real world testing. Just look at the performance figures thread where we track M3 and competitors performance on this board. In real world test it is impossible to precisely control all variables (temperature, pressure, tire conditions and inflation, individual car to car variations, driver skill, equipment accuracy, etc. etc., we could go on forever here). just look at the real spread of any performance spec of a car from a number of sources! What really matters is an average number and some idea of the statistical spread of those numbers. With simulation you should be able to get a good average number, and sometimes when clever, you can even simulate the distribution of results. Scientists do this regularly  one common technique here is called Montecarlo analysis and guess what  it simply "works". I am encouraged that you are tracking the ISF numbers well enough to have observed this effect. These pieces of data will help me to compare my simulations to an average data point rather than the possible outlier point I noted in my first point. Your comments that the wheel hp is 20% less than the crank is a drastic over simplification. The parasitic losses in a vehicle are from a few main sources, some speed dependent and some rpm dependent, other yet still temperature dependent. Losses come from friction, air resistance, bearing losses (friction) tire heat dissipation, etc., etc. An accurate quantification of these losses is essential in any simulation. These parameters are generally well known and modeled. Using a fixed number here, say 20% is just as bad as talking about peak hp and not the entire torque across the whole rpm range. It will tell you some things but by far not the entire picture. Note my previous posts showing how CarTest takes all of this into account! If you would have read my posts more carefully you would see I have never once "dogged" Lexus. I am thrilled (yet very suprised) the car is performing as well as it did in the C&D test. Competition is simply great for everyone! If you take any time to read some of my posts here on this forum you will see I am quick to ciriticize the M3 and BMW as well as sing its praises. I just am not a fanboy. I am not sure how many times I have to say it (adding a point here actually and refining a bit) but this is my conclusion. One or more of the following are INCONSISTENT: 1. The C&D performance figures acurately representing the average or even the peak performance of the car 2. The weight of 1690 kg 3. The reported hp and torque figures 4. The dyno results This has nothing to do with brand loyalty, fanboyism, favoritism, bias, or any such thing. The "facts" as they are reported to us JUST DO NOT ADD UP. The C&D results for the car are too good given the specs and knowledge we have on it. 

Appreciate
0

10252007, 10:51 PM  #69  
Lieutenant General
236
Rep 10,257
Posts 
Hint  final drive
Quote:
Quote:
Car, 1st gear ratio, final drive ratio, gear ratio product M3:...4.055...3.846...15.6 ISF:.4.596...2.176...10.0 That is a 50% advantage in torque multiplication for the M3! When you want to determine instantaneous accelerative wheel torque (and force and then F=ma for vehicle acceleration) the wheel torque is simply multiplied by the product of the gear ratios (there is some scaling as well but that is not important to do a simple AB comarison). So M3 torque in 1st gear x gear factor = 295 ft lb x 15.6 = 4602 ft lb. ISF torque x gear factor = 371 ft lb x 10.0 = 3710 ft lb. No surprises here! Add in the much higher redline and you will see more of the picture. Hope that helps. 

Appreciate
0

10252007, 11:20 PM  #70  
Lieutenant
11
Rep 489
Posts 
Quote:
Thank you for the info. Here's another question, does that program actually predict 1/4 mile times??? This is the thing, when Car and Driver was only able to pull 12.9 I was suprised they couldn't get it faster. As I have stated in other posts I feel as though the m3 can pull a 12.7 at an actual drag strip, but if your program is correct, and Lexus isn't lying, maybe it's possible for C&D to get a 12.6 out of the car? 

Appreciate
0

10262007, 12:24 AM  #71  
Lieutenant General
236
Rep 10,257
Posts 
Whatever you like
Quote:
It predicts about anything you like (and more), 0distance X time, 0  speed X time, rolling starts, launches, in gear acceleration times from speed X to speed Y, works in US or metric units, wheel forces, parasitic losses, air drag force, graphs vs. time, acceleration vs time (g's), etc. Have a look at some results where I used it to predict the MDCT results and MDCT M3 vs. R8 here. I had some subtleties to work out with the time stepping the in program that I did not get really correct until my last post in that thread. The conclusion of this effort was that MDCT M3 will best the R8 in most speed contests and that we should see sub 10 sec. 0100 mph in the MDCT car. You can also see some more nice screen shots from member T Bones use of the software here. I was a bit quick to criticize him back then for the difficulty in modeling the actual launch event (friction, tire slipping, etc.). But you know what, I was wrong, and although it is not perfect it is darn good. 

Appreciate
0

10262007, 01:25 AM  #72 
Colonel
33
Rep 2,291
Posts 
C/D numbers are very questionable. Look at the numbers that they tested on the 06 330i:
060mph 5.6s ? 1.0 faster than spec on BMW website? 0100mph 15s So 060mph = 4.2s for ISF? I would add .4.6s on top of that.
__________________
It's a great time to be alive
Last edited by Pretender; 10262007 at 03:06 AM. 
Appreciate
0

10262007, 02:01 AM  #73  
Brigadier General
109
Rep 4,992
Posts 
Quote:
Jason
__________________


Appreciate
0

10262007, 02:06 AM  #74  
Captain
26
Rep 707
Posts 
Quote:


Appreciate
0

10262007, 02:12 AM  #75 
Captain
26
Rep 707
Posts 
As Jellis have stated, all manufacturers performance figures are usually if not always, very conservative.

Appreciate
0

10262007, 03:06 AM  #76  
Lieutenant General
236
Rep 10,257
Posts 
You are not reading nor thinking
Quote:
From that tiny bit of data you would know that something was clearly and majorly wrong. The case here with the ISF is a bit more subtle (which you cant seem to appreciate) but the concept is the same. Get it, can't you abstract just a tiny bit? If you had more data on the car you could start to figure out what was wrong. "Stock" would be my first guess which includes power, weight and tires. Timing errors or simply going off a cliff might be other less likely possibilities . Look, it is basically F=ma, force equals mass times acceleration, also called Newtons Law. That is what governs 90% of the behavior of the acceleration of a car. If you know F and m you know a, you can not specify all three to you liking. Know why it is called a "law"? Because it can not be broken. Perhaps not even a strong enough term as people can indeed break legal laws. Do you know how much variation exists with dyno tests as well? Apparently not. They are quite good for AB comparisons, say after a mod with all else exactly equal, but for an absolute prediction they have large variation and they test the vehicle not the engine. I guess you know that different brands of dyno give different results as well? Clearly the car is not changing the measurement accuracy is! I think it is also safe to assume that you are not aware that this software does have as on option and ability to repeatedly launch a car under a variety of conditions (clutch drop and clutch slip at any rpm for both and "power braking" as well for automatics, again at any rpm). Then the software automatically (in about 2 seconds) optimizes the launch to get the best acceleration. You then instruct the software to accept the optimized launch schedule before proceeding to get simulation results. Neat eh? This is not simply a formula or spreadsheet you plug numbers into. It is physics based numerical simulation that numerically integrates (do you know what that means?) in real time the behavior of many systems of a vehicle all coupled together. Sure each magazine does multiple runs but again you did not read nor understand my comments about variability. I am not talking about that kind of variation, I am talking about the variation among all magazines (and all drivers) who by the way use journalists who are not professional drivers but are certainly good enough to get a good 060 or 1/4 mi time or whatever. I can just keep going and going... I never said "my data" nor my calculations are the final conclusion nor did I say any reported results are "wrong". Each piece of data is a piece of what I consider to be a puzzle and all I am saying is something does not fit up amongst the data we have and SIMULATION can further provide insight into things not adding up. I think I am going to stop wasting my time and do some more simulations which are interesting, fun, educational and useful. However, I am finding it less and less of these trying to discuss this with you. Of course I will continue to accept new data including more magazines results and other dyno results. 

Appreciate
0

10262007, 03:08 AM  #77 
Colonel
33
Rep 2,291
Posts 
Yes, but 1.0s on a 06 330? No way. I haven't seen any 06 330 060mph result from members < 6.0s yet. Is there any?
__________________
It's a great time to be alive

Appreciate
0

10262007, 08:05 AM  #78 
Moderator
364
Rep 14,377
Posts 
Something else to ponder: Why does it seem the ISF is so much faster than the E39 M5, a car which it is very similar in weight, power and torque (and not TOO far off in redline and final drive either)? I wonder what C&D's test numbers were for the old M5? Because just looking at these early tests, if you lined them up, the M5 will get killed.
Either the ISF is underrated, or that is one very amazing transmission. Or both. Either way, Kudos to Lexus for taking us all by surprise  at least in the straight line department. It will be very interesting to see the comparo tests vs. an E90 M3.
__________________
A genuine BMW effeightzero with them tandem clutches in the transmission and that dad gum sun roof on the topa da cawr.

Appreciate
0

10262007, 08:37 AM  #79  
Captain
26
Rep 707
Posts 
Quote:
Last edited by gbb357; 10262007 at 11:01 AM. 

Appreciate
0

10262007, 11:21 AM  #80 
Major General
117
Rep 8,034
Posts 
Guys, I have lost track of the real difference of opinion in this debate. Swamp has done some detailed modeling of the 060 scenario with CarTest, which does seem like a fairly comprehensive and reliable simulation environment when it comes to that sort of thing. I believe he has validated the accuracy of the environment for several cars by using manufacturers' specs as parameters for the simulation and comparing the outcomes to the performance figures reported by magazines. That said, it is valid to question a simulation outcomeregardless of the comprehensiveness of the model being usedwhich I guess is what gbb357 is doing. For instance, my understanding is that C&D is using rolling starts in the 060 test. Does CarTest model that? Where does CarTest get the necessary data for modeling transmission losses, and how accurate are that data (I am not talking about the conceptual physics model that is being used to simulate the losses; I am referring to the actual parameters). Regardless, I think it is possible to say that, with some degree of confidence, an engine might be underrated if things simply do not add up in the end. Maybe the issue here is what level of confidence one can make such a statement given there might be uncertainties in the simulation model. Is that the basis of this debate? If it is, then I don't see how one can come up with a definitive answer in the absence of a larger statistical database which shows how CarTest simulation outcomes compare to reported realworld performance figures from multiple sources. Maybe the authors of CarTest have published that type of information?

Appreciate
0

10262007, 11:42 AM  #81 
CommanderInChief
163
Rep 7,750
Posts
Drives: 2015 M4 Coupe, 2012 ML350
Join Date: Nov 2005
Location: Lake Oswego, OR

Autoweek IS F article
Didn't see this posted yet. Some snippets
 Each cylinder has two injectors, one port and one direct  Electrically variable titanium intake valves; hydraulically variable steel exhaust valves (lift stays same on valves)  Upshifts take 0.1 seconds  "'The goal is not to make a BMW M3,' Yaguchi said, though many people will see it as such. 'The M3 is fun for a realy good driver, but if you're not a really good driver, it's not fun,' he said. 'This is a car everyone can enjoy; with this car you skill level doesn't matter.'"  Great!  "This one (suspension) is softer than the previous M3's...but, it is a tradeoff we could live with."  "Lexus cleverly chose to reveal its IS F just before BMW revealed its M3 (sedan). Or was that pure coincidence? In either case, everyone's stories about the Lexus IS F will come out before other stories about the M3. This will allow people to view the IS F separatels as a sports sedan unot itself. In such a context it is an unbridled success. Against the M3, we'll have to wait and see."
__________________
Greg Lake Oswego, Oregon, USA 2015 M4 Coupe  Silverstone/Sakhir/CF 2012 ML350 Last edited by GregW / Oregon; 10262007 at 01:26 PM. 
Appreciate
0

10262007, 11:52 AM  #82  
Captain
26
Rep 707
Posts 
Quote:


Appreciate
0

10262007, 11:59 AM  #83 
Major General
117
Rep 8,034
Posts 
I shouldn't have said "rolling start". I meant to say that they start the clock when the car has covered some small distance after an optimized launch. I thought Bruce had posted some info on this a while back, but I might be wrong.

Appreciate
0

10262007, 12:04 PM  #84 
Lieutenant
11
Rep 489
Posts 
Seems only fair, the 1/4 mile doesn't start until you trip the lasers....

Appreciate
0

10262007, 12:29 PM  #85  
Lieutenant General
236
Rep 10,257
Posts 
Good summary
Quote:


Appreciate
0

10262007, 01:01 PM  #86  
Lieutenant General
236
Rep 10,257
Posts 
Still missing the point
Quote:
Quote:
Quote:
Quote:
Quote:
Last but not least when you say the M3, RS4 and ISF "specs" are close you have to dive a lot deeper than hp! So on the surface you are correct but looking at what matters you are terribly and grossly wrong. You need peak torque, the enitire torque curve, the weight and last but not least one of the most important things you keep over looking is the importance of the transmission type and gear ratios (gears AND final drive). Did you happen to read my post about torque multiplication and showing how the M3 has about 25% more torque delivered to the ground than the ISF in 1st gear (and similar but smaller advantages in other gears)? I don't, and no one who knows anything, gives a rats a$$ about the engine peak torque spec. or even rear wheel dyno torque results comparing the these two (or any two cars for that matter...). The torque delivered to the wheels divided by the weight is an absolutely essential factor. It is absolutely not about which car I like better or which has superior data sheet specs.  it is about consistency of all of the specs, testing, data and simulation. If the ISF had the superior torque delivery and poor real world tests I would be just as confused and questioning everything as well! 

Appreciate
0

10262007, 01:10 PM  #87  
Major General
117
Rep 8,034
Posts 
Quote:


Appreciate
0

Post Reply 
Bookmarks 
Thread Tools  Search this Thread 

