Drag Coefficient

[Blooze]

Well, I found it sort of...

But where I found it, out at the end of a search for the terminal velocity, will do us no good.   At least, that's why I wanted to know it, so I could calculate the TV and not wander round and round Nissan until I had it dialed in.

Anyhow, I was challenged the other day about a torque statement I had made so I went to check it out.  I was making a dyno run in 4th gear of a car that I hadn't dialed in the FD yet.  I usually run 'em out until they bounce off the rev limiter, but this car wouldn't hit the rev limiter.  I had to turn it around and come back in the third to hit it.  I filled in the torque table and checked the torque value at the engine rpm that stymied the poor thing.  840 ftlbs at the drive wheels.  At that time that torque value showed up again in the middle range of 5th and at the first of the range in 6th.

So I trailered the dood up and headed out for Nissan Roundy Round.  I tried it out, and sure enough... in 4th when I got to the engine rpm that produced the required air speed the thing washed out.  Same in 5th, and in 6th.  The same speed, the same torque value.  It occurred to me that I was looking at the result of the drag coefficient right there.

So, all a feller has to do is make an accurate record of the torque value, and the speed at which he hits the air wall... then see if he can find the calculations that produce work or torque values from drag coefficient (my Automotive Math book doesn't have the calcs in it), weight (maybe), and velocity and work the equation backwards - yahoo...  drag coefficient.   

In other words, we are still right where we were - wishing they published the value.

$

[Tonka Crash]

Drag in lbs is given by:



where:
ρ is air density in slugs/cubic foot around .002367 for the Standard Atmosphere at sea level.
V is velocity in ft/sec
C_dA is the drag coefficient multiplied by the frontal area.  This is the value to look for.  With the tools available I don't think C_d can't be isolated independently of the frontal area

Wikipedia even has some numbers that match up to some of the cars in the game.

CdA ft²    Automobile model
3.95    1996 GM EV1
5.10    1999 Honda Insight
5.71    1990 Honda CR-X Si
5.76    1968 Toyota 2000GT
5.80    1986 Toyota MR2
5.81    1989 Mitsubishi Eclipse GSX
5.88    1990 Nissan 240SX
5.92    1994 Porsche 911 Speedster
5.95    1990 Mazda RX7
6.00    1970 Lamborghini Miura
6.13    1993 Acura NSX
6.17    1995 Lamborghini Diablo
6.27    1986 Porsche 911 Carrera
6.27    1992 Chevrolet Corvette
6.35    1999 Lotus Elise
6.40    1990 Lotus Esprit
6.54    1991 Saturn Sports Coupe
6.57    1985 Chevrolet Corvette
6.77    1995 BMW M3
6.79    1993 Toyota Corolla DX
6.81    1991 Subaru Legacy
6.90    1993 Saturn Wagon
6.93    1982 Delorean DMC-12
6.96    1988 Porsche 944 S
6.96    1995 Chevrolet Lumina LS
7.02    1992 BMW 325I
7.04    1991 Honda Civic EX
7.10    1995 Saab 900
7.14    1995 Subaru Legacy L
7.34    2001 Honda Civic
7.39    1994 Honda Accord EX
7.48    1993 Chevrolet Camaro Z28
7.57    1992 Toyota Camry
7.69    1994 Chrysler LHS
7.72    1993 Subaru Impreza
8.70    1990 Volvo 740 Turbo
8.70    1992 Ford Crown Victoria
8.71    1991 Buick LeSabre Limited
9.54    1992 Chevrolet Caprice Wagon
10.7    1992 Chevrolet Blazer
11.6    2005 Ford Escape Hybrid
11.7    1993 Jeep Grand Cherokee
16.8    2006 Hummer H3
17.4    1995 Land Rover Discovery
26.5    2003 Hummer H2

I'm still thinking about how to relate this to data available from Telemetry.  Rolling resistance of the tires and friction losses still need to be accounted for and I don't know if you can get those.  They probably are modeled as constants so it may be possible to isolate them. 

[Blooze]

Yeah - having the data after the fact isn't of much value, having only part of the data after the fact is worth even less...

Interesting little bit there with the CRX, maybe a clue as to why it was so much trouble last year...

Good job on finding all that so fast!

$

[bimmerlovere39]

Hm...

a couple of note that I thought I'd throw in:

Does the game properly manage rolling resistance?  If so, (theoretically), wouldn't you want the hardest, narrowest tires possible?

The C_dA chart looks fishy... I don't ean to insult ya here... but I have a hard time believing an M3 has a lower C_dA than a 325i.  The '92 and the '95 are both pre-facelift E36s.  The M3 has a bigger frontal area thanks to the front bumper and wider tires.  I seriously doubt the C_d is lower on the M3 either.

[Tonka Crash]

Blooze, I know you're a retired programmer.  How tough do you think it would be to extract the telemetry data from a replay file?  It would be a whole lot easier for me to look at the telemetry of my laptop.  That's my job, I analyze flight test data and do a little simulation of aircraft.  I also do a little coding as part of my job.  Any idea if the Xbox is big endian or little endian?   I've got one the Datel Xport kits.  I bought it to backup my saves to my PC, but because of the last patch I don't think I can use it for that and really haven't used it at all other than to see it worked.

[Tonka Crash]

The C_dA chart looks fishy... I don't ean to insult ya here... but I have a hard time believing an M3 has a lower C_dA than a 325i.  The '92 and the '95 are both pre-facelift E36s.  The M3 has a bigger frontal area thanks to the front bumper and wider tires.  I seriously doubt the C_d is lower on the M3 either.

The numbers are from Wikipedia, so your guess is as good as mine as to the validity.  I've seen the number for the Lotus Elise in a couple different places around the web.  This isn't the first time I've thought about drag coefficients, but it was starting to make the game feel more like work than play, so I put it on the back burner and went back to racing.

[Big Mooing Cow]

Could you not save some time using the Benchmark screen?  When we were running the oval challenge I found the top speed number seemed to be pretty accurate in most cases... within a couple MPH at worst.

Seems like it would be faster to use the Benchmark screen to find the top speed like you were ranging artillery than it would to find the top speed on the oval, where there are one or two instants in time where you could know your exact top speed, and only then if you nailed the previous corner.

Where's the damn Test Track "oval" when you need it?  I spent half my life on that track tuning transmissions in FM1!

[Hekalite]

Blooze, I know you're a retired programmer.  How tough do you think it would be to extract the telemetry data from a replay file?  It would be a whole lot easier for me to look at the telemetry of my laptop.  That's my job, I analyze flight test data and do a little simulation of aircraft.  I also do a little coding as part of my job.  Any idea if the Xbox is big endian or little endian?   I've got one the Datel Xport kits.  I bought it to backup my saves to my PC, but because of the last patch I don't think I can use it for that and really haven't used it at all other than to see it worked.

I'm a programmer as well.  How easy it is to pull data from the replay file all depends on the format.  I don't have high hopes for that being an easy task.  They would want the replay data files to be as small as possible and small does not generally lead to readable.

Not to mention, the data might not even be in the replay file.  They may just "rerun" the race through the system for display.

[Blooze]

Not to mention, the data might not even be in the replay file.  They may just "rerun" the race through the system for display.

I can't put my finger on it now, but I've seen some things that indicate this to be true, that the save file is a binary of coordinates and such that feeds the display engine.  We'd have to be privy to a whole bunch of their code to do anything with it.  I think Auto-Cad does the same thing with their files, if you want a picture, you gotz to load it in Auto-Cad to print it.

$

[Blooze]

Where's the damn Test Track "oval" when you need it?  I spent half my life on that track tuning transmissions in FM1!

LOL - didn't we all.  I used to have me a rule, build it to get to the 1500 meter mark the fastest.  That was usually the best racer, except for the TT Oval itself... 

Speak of the deebil his ownself... there's a post in the Excel thread about setting the FD from the tuning screen this very morning.

$

[Big Mooing Cow]

I would bet that the simulation is completely deterministic.  The replays would be run by feeding the driver's inputs (stored in the replay file) into the simulation, and you'd get the exact same results as you got in the race.

I doubt they're storing the state of cars throughout the race, because if they were their rewinding and changing car times would be near-instantaneous rather than introduce the large delays that they do presently.

Also, Turn 10 has said in the past that they would have to actually run replays to see the telemetry from the replays, in regards to being able to tell if people were AWD glitching.  Now that might be typical Turn 10 smoke blowing, but it also fits my hypothesis.

If my hypothesis is true, it would be impossible to get car data from the replay files, because that data lives in the simulation and the simulation alone.

[Hoplee]

if you watch the replays, you'll notice something interesting. I was watching some of the top times from the FCT Challenge, specifically RacingPro97 and bazdc2. Their cars begin the replay with some heavy scraping damage along the right side of the car. If you let the replay run a few laps, you'll eventually see a collision between the car and a bumper left on the track from the wreck that presumably happened before the replay was recorded.

This indicates to me that the replays record more than driver inputs. At the very least, they're recording damage data and even keeping track of parts knocked off the car.

[Big Mooing Cow]

Hm, considering that your fastest lap might be one with damage, to get identical results between the race and replay they'd have to store existing damage, too.  Otherwise a car with aero damage would appear to drive faster in the replay than it did in the real race.

I think you're right; they have to store damage in the replay if they are doing what I think they are.  Starting position, speed, gear, RPM, and suspension state, too.

[Hekalite]

Hm, considering that your fastest lap might be one with damage, to get identical results between the race and replay they'd have to store existing damage, too.  Otherwise a car with aero damage would appear to drive faster in the replay than it did in the real race.

I think you're right; they have to store damage in the replay if they are doing what I think they are.  Starting position, speed, gear, RPM, and suspension state, too.

Consider this though, have any of you ever sworn you beat your ghost and yet you didn't get a new fast time?  That has happened to me a few times and I'm not sure what it says about the replay data other than I doubt we can trust it.

[TheJohnNewton]

Consider this though, have any of you ever sworn you beat your ghost and yet you didn't get a new fast time?  That has happened to me a few times and I'm not sure what it says about the replay data other than I doubt we can trust it.

Just yesterday I was racing a ghost from the leader board to help me get a good line.  The ghost was #100 on the leader board and my best was around #300.  So I start the lap and notice my best ghost is ahead of the leader board ghost through the first few turns.  I'm thinking "great I do have a good line there" but this continues the entire lap and my ghost actually finishes ahead of the other one.  Then the next lap I finish ahead of the ghost with no penalty time and still my time is behind the ghost time.     So there is some issue with the fidelity of the ghost replay.