Cars is too quick per specs...

[reece146]

I've been working on a North American spec '98 Subaru Impreza RS.

I've got decent dyno data - taken off a dyno before the car was modified into a turbo monster (baseline).

This data is likely 20%-30% less than the factory flywheel rating for the engine (dyno says ~120 lb-ft peak vs 165 lb-ft IIRC).

But the car in vdrift is something like 1.5-2.0 seconds faster to 100 km/h (~6.5s vs ~8.5s). I think I have the total weight right.

I think I have the tire radius correct also.

The clutch is not correct but I'm certain that the 1-2 shift in my own Impreza is only 0.3. seconds or so; that doesn't account for the difference.

Is there something I'm missing?

It's not all negative - the car is a blast to drive but it is not accurate.

I can post the .car file if someone wants to take a close look.

[thelusiv]

You can go ahead and post the car file, but disable smilies and put it in a code block when you post it, so the formatting won't get screwed up. Are you sure you're converting your torque values from ft-lbs to N-m (Newton-meters) and do you have torque data points at RPMs up to and above the max RPM?

[reece146]

Yep, values are in N-m (1 N-m = 1.344818 lb-ft)

Here's RS2.car. The supension, aero and geometry stuff is not quite right yet. Everything else has been tweaked though. For straight line performance it should be close enough?

Code:

drive = AWD

[ steering ]
max-angle = 33.19

[ engine ]
position = 1.04, 0.0, -0.24
mass = 150.0
max-power = 1.07e5
peak-engine-rpm = 5500.0
rpm-limit = 6800.0
inertia = 0.35
idle = 0.02
start-rpm = 1000
stall-rpm = 350
fuel-consumption = 1e-9
torque-friction = 0.0006
#
# The following values are a stock RS as meassured at the wheels.
#
torque-curve-00 = 1000, 143.7
torque-curve-01 = 1200, 139.6
torque-curve-02 = 1400, 138.3
torque-curve-03 = 1600, 138.3
torque-curve-04 = 1800, 139.6
torque-curve-05 = 2000, 141.7
torque-curve-06 = 2100, 143.7
torque-curve-07 = 2200, 139.6
torque-curve-08 = 2300, 138.3
torque-curve-09 = 2400, 138.3
torque-curve-10 = 2500, 138.3
torque-curve-11 = 2600, 141.0
torque-curve-12 = 2700, 143.0
torque-curve-13 = 2800, 141.7
torque-curve-14 = 2900, 141.0
torque-curve-15 = 3000, 142.4
torque-curve-16 = 3100, 143.7
torque-curve-17 = 3200, 146.4
torque-curve-18 = 3300, 148.5
torque-curve-19 = 3400, 149.8
torque-curve-20 = 3500, 151.2
torque-curve-21 = 3600, 150.5
torque-curve-22 = 3700, 150.5
torque-curve-23 = 3800, 149.8
torque-curve-24 = 3900, 149.8
torque-curve-25 = 4000, 149.1
torque-curve-26 = 4100, 149.1
torque-curve-27 = 4200, 149.1
torque-curve-28 = 4300, 149.1
torque-curve-29 = 4400, 148.5
torque-curve-30 = 4500, 147.8
torque-curve-31 = 4600, 146.4
torque-curve-32 = 4700, 145.8
torque-curve-33 = 4800, 145.1
torque-curve-34 = 4900, 144.4
torque-curve-35 = 5000, 143.7
torque-curve-36 = 5100, 142.4
torque-curve-37 = 5200, 141.7
torque-curve-38 = 5300, 140.7
torque-curve-39 = 5400, 139.6
torque-curve-40 = 5500, 137.6
torque-curve-41 = 5600, 133.5
torque-curve-42 = 5700, 130.8
torque-curve-43 = 5800, 127.4
torque-curve-44 = 5900, 123.4
torque-curve-45 = 6000, 120.7
torque-curve-46 = 6100, 116.6
torque-curve-47 = 6200, 114.2
torque-curve-48 = 6300, 112.5
torque-curve-49 = 6400, 108.5
torque-curve-50 = 6500, 103.0

[ clutch ]
sliding = 0.27
radius = 0.1
area = 0.75
max-pressure = 30079.26

[ transmission ]
gears = 5
gear-ratio-r = -3.416
gear-ratio-1 = 3.545
gear-ratio-2 = 2.111
gear-ratio-3 = 1.448
gear-ratio-4 = 1.088
gear-ratio-5 = 0.780
shift-delay = 0.05

[ differential ]
final-drive = 4.111
anti-slip = 600.0

[ fuel-tank ]
position = -0.26, -0.17, -0.56
capacity = 60
volume = 60
fuel-density = 1.0

[ suspension-front ]
spring-constant = 23000
bounce = 5000
rebound = 4500
travel = 0.24
max-compression-velocity = 10.0
camber = -1.0
caster = 0.0
toe = 0.5
#anti-roll = 24000
anti-roll = 10000

[ suspension-rear ]
spring-constant = 28750
bounce = 5000
rebound = 4500
travel = 0.26
max-compression-velocity = 10.0
camber = -0.5
caster = 0.0
toe = -1.0
#anti-roll = 21000
anti-roll = 20000

[ tire-front ]
radius = 0.31595
rolling-resistance = 1.3e-2, 6.5e-6
rotational-inertia = 10
tread = 0.0
# Lateral force
a0=1.4
a1=-35
a2=1550
a3=2400
a4=6.5
a5=0.014
a6=-0.24
a7=1.0
a8=-0.03
a9=-0.0013
a10=-0.15
a111=-8.5
a112=-0.29
a12=17.8
a13=-2.4
# Longitudinal force
b0=1.6
b1=-70
b2=1600
b3=23.3
b4=350
b5=0.05
b6=0.0
b7=0.055
b8=-0.024
b9=0.014
b10=0.26
b11=0.0
b12=0.0
# Aligning moment
c0=2.2
c1=-3.9
c2=-3.9
c3=-1.26
c4=-8.2
c5=0.025
c6=0.0
c7=0.044
c8=-0.58
c9=0.18
c10=0.043
c11=0.048
c12=-0.0035
c13=-0.18
c14=0.14
c15=-1.029
c16=0.27
c17=-1.1

[ tire-rear ]
radius = 0.31595
rolling-resistance = 1.3e-2, 6.5e-6
rotational-inertia = 10
tread = 0.0
# Lateral force
a0=1.3
a1=-45
a2=1700
a3=2500
a4=6.5
a5=0.014
a6=-0.24
a7=1.0
a8=-0.03
a9=-0.0013
a10=-0.15
a111=-8.5
a112=-0.29
a12=17.8
a13=-2.4
# Longitudinal force
b0=1.6
b1=-90
b2=1600
b3=23.3
b4=375
b5=0.05
b6=0.0
b7=0.055
b8=-0.024
b9=0.014
b10=0.26
b11=0.0
b12=0.0
# Aligning moment
c0=2.2
c1=-4.1
c2=-3.9
c3=-1.36
c4=-8.0
c5=0.025
c6=0.0
c7=0.044
c8=-0.58
c9=0.18
c10=0.043
c11=0.048
c12=-0.0035
c13=-0.18
c14=0.14
c15=-1.029
c16=0.27
c17=-1.1

[ brakes-front ]
friction = 0.73
max-pressure = 2.0e6
bias = 0.57
radius = 0.14
area = 0.015

[ brakes-rear ]
friction = 0.73
max-pressure = 2.0e6
bias = 0.57
radius = 0.14
area = 0.015
handbrake = 1.0

[ driver ]
position = -0.21, -0.32, 0.0
mass = 102.0
view-position = -0.21, -0.45, 0.43

[ drag ]
position = 0.0, 0.0, 0.2
frontal-area = 2.37
drag-coefficient = 0.35

[ wing-front ]
position = 1.9, 0.0, 0.60
frontal-area = 0.2
drag-coefficient = 0.0
surface-area = 0.3
lift-coefficient = -0.5
efficiency = 0.95

[ wing-rear ]
position = -2.14, 0.0, 0.37
frontal-area = 0.05
drag-coefficient = 0.0
surface-area = 0.5
lift-coefficient = -2
efficiency = 0.95

[ wheel-FR ]
position = 1.14, -0.70, -0.48
roll-height = 0.31265
mass = 20
restitution = 0.1

[ suspension-FR ]
position = 1.14, -0.66, -0.15
hinge = 0,0,0

[ wheel-FL ]
position = 1.14, 0.70, -0.48
roll-height = 0.31265
mass = 20
restitution = 0.1

[ suspension-FL ]
position = 1.14, 0.66, -0.15
hinge = 0,0,0

[ wheel-RR ]
position = -1.12, -0.70, -0.505
roll-height = 0.3179
mass = 22
restitution = 0.1

[ suspension-RR ]
position = -1.12, -0.65, -0.13
hinge = -0.6,0,0

[ wheel-RL ]
position = -1.12, 0.70, -0.505
roll-height = 0.3179
mass = 22
restitution = 0.1

[ suspension-RL ]
position = -1.12, 0.65, -0.13
hinge = -0.6,0,0

[ contact-points ]
# used to find the bounds of the car (for collisions)
# also play into the weight distribution and balance
mass = 0.05
position-00 = 2.06, 0.40, -0.20
position-01 = 2.06, -0.40, -0.20
position-02 = 1.85, 0.72, -0.20
position-03 = 1.85, -0.72, -0.20
position-04 = 0.19, 0.89, 0.24
position-05 = 0.19, -0.89, 0.24
position-06 = -1.99, -0.72, -0.26
position-07 = -1.99, 0.72, -0.26
position-08 = -2.23, -0.46, -0.26
position-09 = -2.23, 0.46, -0.26

# these are used to calculate the weight distribution and balance
[ particle-00 ]
mass = 60.0
position = 1.008, 0.0, -0.45

[ particle-01 ]
mass = 178.5
position = -1.48, 0.4, 0.1

[ particle-02 ]
mass = 178.5
position = -1.48, -0.4, 0.1

[ particle-03 ]
mass = 348.0
position = 1.08, 0.4, 0.0

[ particle-04 ]
mass = 348.0
position = 1.08, -0.4, 0.0

[thelusiv]

Do you also have a model to go with this? If so we might as well add it to SVN and make it easier for everyone to test it out.

[reece146]

Do you also have a model to go with this? If so we might as well add it to SVN and make it easier for everyone to test it out.

Unfortunately, no. I am currently just using the TC body on it for the moment.

I could stick the model-pack body on it and comment the hell ut of it so that it is apparent that it is a "beta" car.

[thelusiv]

I'll try it out with the TC.

[thelusiv]

Alright I tested it out and checked it in. I like how it drives, but I agree that it seems a bit fast. Can you post the dyno charts you made the torque curve from?

[reece146]

Alright I tested it out and checked it in. I like how it drives, but I agree that it seems a bit fast. Can you post the dyno charts you made the torque curve from?

Saw that... but it doesn't come up as an option - at least not in Build 1393. Been meaning to ask how to add cars now - it seems to have changed since 10-06?

Can I svn co/ci the car? I'll tweak it as I get more info. I want to address the suspension sooner than later.

Do I need an account to check stuff in or is it a free for all? 'Figured better ask before trying - svn newbie.

I'll post (check in?) the charts once I get a chance to convert them from XLS to OD. I want to document the guts out of this car.

[thelusiv]

It's listed as the TC (I forgot to change the name in about.txt). Just look for the second time the TC is listed. It should be one of the last ones.

You can check it out from SVN, if you have already got an svn checkout then just 'svn up' should get it.

Anyone can check out from SVN but only people who have been manually added have check in access.

Currently we don't really have any particular place we keep raw data about cars (since most of it came from the web anyway). I don't think it necessarily has to be checked in. Maybe you could post it on the development mailing list as an attachment, or just copy the data out of the file and post it on the forums.

OpenOffice will open XLS and save OpenDocument, so if you want to do that I suggest OO

[cotharyus]

OpenOffice will open XLS and save OpenDocument, so if you want to do that I suggest OO

Second that. Some of us have not bought anything from certain software companies in 10 years or better, and refuse to burden a computer with such restrictive software

OOo rules.

[reece146]

Here is the spreadsheet:

http://moose.ca/content/RS_DOHC_Dyno_w_dyno.ods

I've cleaned it up a touch and in fact lessened some of the fudged values between 1000 to 2000 RPM (steeper curve descent from 2000 down).

The two columns on the end with the "Factory" heading can be ignored for the most part. They are rough estimates of what the curves will look like with factory released HP/TQ values plugged in. They are a way to estimate drivetrain loss but are probably way out of whack given the scale of the loss and the fact they don't match between HP and TQ.

Anyway, columns A and B are the "real" life dyno curve taken off a stock (?) 1998 Subaru Impreza RS 2.5 (N.A. spec). The car was run by Cobb Tuning in a comparison against the 1999 SOHC Impreza RS. The dyno chart is embedded in the spreadsheet for those interested. The numbers are low so I don't think any "fudging" was done to them.

Column D is the converted torque values in N-m as required by Vamos/VDrift.

I don't know why this car is so quick in the simulator. Maybe Vamos does something "funny" with AWD? Can someone confirm the weight in the RS2.car file? I think it is correct total weight but...

Looking at the dyno numbers taken at the wheels, you'd expect the car to be quite a bit slower than reality. I initially had the numbers multipled by ~20% to compensate but that setup was "way crazy fun" fast so I stuck the actual dyno numbers.

Any insights appreciated.

[thelusiv]

I have just checked in a model for the RS2 that kcid added to the art repository a little while ago, along with some updates to the RS2.car. This one drives pretty nicely now. The textures are crap (just solid colors) but the model looks good. All this is in SVN r1594. Here's a link I used to help figure out some of the specifications: http://www.sidc.co.uk/faq.htm

The car could probably still handle better, if anyone wants to play around with it.

One more thing, reece146, do you know what kind of dyno was used? I have heard that some dynos tend to give numbers a little high, and some others a little low.

[reece146]

I have just checked in a model for the RS2 that kcid added to the art repository a little while ago, along with some updates to the RS2.car. This one drives pretty nicely now. The textures are crap (just solid colors) but the model looks good. All this is in SVN r1594. Here's a link I used to help figure out some of the specifications: http://www.sidc.co.uk/faq.htm

The car could probably still handle better, if anyone wants to play around with it.

One more thing, reece146, do you know what kind of dyno was used? I have heard that some dynos tend to give numbers a little high, and some others a little low.

I'll try to check it out tonight. I'll check my notes about the dyno stuff too.

[reece146]

Managed to get the new build compiled and to test the new RS2.

The transition to oversteer is too sharp. Way too easy to spin. I'll look at the geometry in the .car file and see how it compares to the geometry of the Subaru in the shop. I'm guessing I'll need to calculate the real ICs and such so this may take a few days (finding the time). Maybe by the weekend I'll have a diff for the .car file.

Btw, enjoying Mosport. I really need a new display card...

[thelusiv]

Yeah I just estimated where the weights should be. I didn't look at weight distribution, or try to do a spreadsheet with the weights and try to get a reasonable center of gravity. The hinge points for the suspension are pretty standard (1.0 in front of back wheels, 1.0 behind front wheels). Also the suspension travel was a guess, and I'm not sure what the spring rates or bounce/rebound should really be.