rec.autos.simulators

High End Force Feedback On The Way...

David L. Coo

High End Force Feedback On The Way...

by David L. Coo » Sat, 26 Jan 2002 12:12:57

Dave - who do you program for?


Paul Thurst

High End Force Feedback On The Way...

by Paul Thurst » Sat, 26 Jan 2002 22:36:41


> I think the test could be something like send a continuous stream of "zero"
> force, followed by a stream of "left" force, all the while continously
> sampling the steering input, and seeing how long before you can read a
> change in steering input due to the motor... my suspicion is that a lot of
> the latency will be in the hardware and the USB connection, but if I get
> some numbers I'll post them here.  It'd be great of there was a standard
> controller latency test, kind of like a Winmark, then the controller guys
> could compete on latency.

Dave you're a hero.  I asked Act Labs for something like this two
years ago but nothing came of it.

There's actually two tests I'd do.

Start wheel moving right then switch to moving left and measure the
time taken until the change reaches the input.  That's the total
latency.

Start it moving left again and measure the time until the signal
reaches the input.  As the wheel had already been moving left it would
have taken up any slack in the mechanism.  Subtract that from the
total latency figure and you have the slack in the wheel.

I'd also try it at different velocities to see if that affects the
latency.

I'm very interested to see what sort of figures you come up with.
I've been convinced for ages that the latency should be measured in
tenths rather then 10,000ths of a second.

Paul
:o)

Goy Larse

High End Force Feedback On The Way...

by Goy Larse » Sun, 27 Jan 2002 00:43:20


> Dave - who do you program for?



> > I suppose I have a more direct comparison of the "MS" component in the
> > equation, because the Playstation2 can talk to the Logitech wheels via USB
> > and Logitech's custom API... our last PS2 game just used pretty much
> > identical code to our PC FF code (just tweaking the constant force
> > continuously based on force at the steering linkage); I started working on
> > my own hacked-together USB driver for other wheels, I'm curious to see
> what
> > kind of latency you can get sans Windows.  Of course you still have to go
> > from the CPU to the IOP sub-processor to get to the USB via DMA, so it's
> not
> > as good as an ISA card or something.

Microprose, his real name is Geoff Crammond.....

Beers and cheers
(uncle) Goy

"The Pits"    http://www.theuspits.com/

* Spam is for losers who can't get business any other way *
"Spamkiller"    http://www.spamkiller.com

Pat Dotso

High End Force Feedback On The Way...

by Pat Dotso » Sun, 27 Jan 2002 09:48:35

While you guys are all paying attention, what do you think about...

Instead of basing FF effects on steering mechanism forces, simply tie FF
to the lateral g-forces acting on the driver?  Then you get the
seat-of-the-pants feel that is missing from a computer simulation.
F1-2001 does it this way, and for my money, it's the best feeling FF
currently available in a PC driving sim.
--
Pat Dotson

Doug Millike

High End Force Feedback On The Way...

by Doug Millike » Sun, 27 Jan 2002 17:11:31


> Instead of basing FF effects on steering mechanism forces, simply tie FF
> to the lateral g-forces acting on the driver?  Then you get the
> seat-of-the-pants feel that is missing from a computer simulation.

If I understand what you are saying, I think this would add more lag.
Steer torque comes from front kingpin torque and this in turn is a function
of the lateral force (and tire aligning torque) at the front wheels.  The
only lag in this system (assuming simple rigid suspension) is the tire lag
(relaxation length), because lateral force at the front axle immediately
starts to "push" the car into a turn.  

When you move aft along the wheelbase to the driver position, then the
onset of lateral force lags a bit, because the whole car has to rotate.
----------

Re original thread -- I'm looking forward to reading about some latency
measurements, could be very interesting.  Might even be embarrasing for the
original authors of any bottleneck(s)!

-- Doug Milliken

Ruud van Ga

High End Force Feedback On The Way...

by Ruud van Ga » Mon, 28 Jan 2002 03:14:19



>> Instead of basing FF effects on steering mechanism forces, simply tie FF
>> to the lateral g-forces acting on the driver?  Then you get the
>> seat-of-the-pants feel that is missing from a computer simulation.

>If I understand what you are saying, I think this would add more lag.

Hm, does indeed look like it. I'll add an option to use this kind of
FF (although it will need tweaking, as the influence of lateral
acceleration vs. tire grip is probably mostly a matter of personal
feel) in Racer.
It's not that hard to build in, so that might give a nice testbed.

...

Isn't the lateral acceleration on every part of the car the same?
(obviously the lateral velocity isn't, but the acceleration seems
untied to a position in the car).
That would suggest no more lag than usual, since even though
rotational *velocity* might need a buildup, the *acceleration* of the
car body doesn't (and that's where you derive the FF from).

Ruud van Gaal
Free car sim  : http://www.marketgraph.nl/gallery/racer/
Pencil art    : http://www.marketgraph.nl/gallery/

Ice

High End Force Feedback On The Way...

by Ice » Mon, 28 Jan 2002 08:12:17

My god, we are still talking about wheels that shake a bit as you drive,
right? Never mind, carry on... ;)

Ice D





> >> Instead of basing FF effects on steering mechanism forces, simply tie
FF
> >> to the lateral g-forces acting on the driver?  Then you get the
> >> seat-of-the-pants feel that is missing from a computer simulation.

> >If I understand what you are saying, I think this would add more lag.

> Hm, does indeed look like it. I'll add an option to use this kind of
> FF (although it will need tweaking, as the influence of lateral
> acceleration vs. tire grip is probably mostly a matter of personal
> feel) in Racer.
> It's not that hard to build in, so that might give a nice testbed.

> ...
> >When you move aft along the wheelbase to the driver position, then the
> >onset of lateral force lags a bit, because the whole car has to rotate.

> Isn't the lateral acceleration on every part of the car the same?
> (obviously the lateral velocity isn't, but the acceleration seems
> untied to a position in the car).
> That would suggest no more lag than usual, since even though
> rotational *velocity* might need a buildup, the *acceleration* of the
> car body doesn't (and that's where you derive the FF from).

> Ruud van Gaal
> Free car sim  : http://www.marketgraph.nl/gallery/racer/
> Pencil art    : http://www.marketgraph.nl/gallery/

Pat Dotso

High End Force Feedback On The Way...

by Pat Dotso » Mon, 28 Jan 2002 10:57:16

You must own a hyperstimulator :)

> My god, we are still talking about wheels that shake a bit as you drive,
> right? Never mind, carry on... ;)

> Ice D





> > >> Instead of basing FF effects on steering mechanism forces, simply tie
> FF
> > >> to the lateral g-forces acting on the driver?  Then you get the
> > >> seat-of-the-pants feel that is missing from a computer simulation.

> > >If I understand what you are saying, I think this would add more lag.

> > Hm, does indeed look like it. I'll add an option to use this kind of
> > FF (although it will need tweaking, as the influence of lateral
> > acceleration vs. tire grip is probably mostly a matter of personal
> > feel) in Racer.
> > It's not that hard to build in, so that might give a nice testbed.

> > ...
> > >When you move aft along the wheelbase to the driver position, then the
> > >onset of lateral force lags a bit, because the whole car has to rotate.

> > Isn't the lateral acceleration on every part of the car the same?
> > (obviously the lateral velocity isn't, but the acceleration seems
> > untied to a position in the car).
> > That would suggest no more lag than usual, since even though
> > rotational *velocity* might need a buildup, the *acceleration* of the
> > car body doesn't (and that's where you derive the FF from).

> > Ruud van Gaal
> > Free car sim  : http://www.marketgraph.nl/gallery/racer/
> > Pencil art    : http://www.marketgraph.nl/gallery/

Pat Dotso

High End Force Feedback On The Way...

by Pat Dotso » Mon, 28 Jan 2002 13:39:16

[FYI, I mistakenly replied to Doug through email instead of
through RAS.  I've forwarded his reply back to RAS - hope
you don't mind Doug]

More interesting stuff, but confusing.  The #1 complaint
I hear from people who *should* be interested in PC sims,
but arent' (people who race in real life) is that they
can't feel what the car is doing.

In my opinion, current FF comes pretty close to making up
for the SOP feel.  But if real drivers don't necessarily
use the g-force feedback they get, then what should PC
sim FF be based upon?  In my real experience, which is
95% go-kart, actual wheel feedback isn't much of a factor
in driving.  Maybe it should be, but it generally isn't.
That's why I like the idea of lateral force-based FF.
--
Pat Dotson



> Interesting point Doug.  What % of feedback to you estimate
> a driver gets from the steering wheel, and what % from
> seat of pants (SOP)?  I haven't thought too much about
> the differences in feedback, and just assumed that SOP
> would be more desirable.

No idea.  People can be very sensitive, I know I can sense .01 g lateral
with my eyes closed...on a smooth freeway.  But I doubt that I could
even
tell .1 g difference if I was getting shook in a race car with
hard-mounted
engine, cornering at 3 g's.  Once drivers learn a track (a few laps for
most places), I think that they drive from memory with small changes
to see if the change will improve things.  They must be good at
filtering
out most of the SOP info coming in because it's huge (body damaging?)
vibration in most race cars.

Winston Cup, Sprint cars, F1, probably others.  It's an area
ripe for innovation, imo.

Why off list??
-- Doug Milliken

Pat Dotso

High End Force Feedback On The Way...

by Pat Dotso » Mon, 28 Jan 2002 13:48:25

F1 2001 FF is based on lateral force at the front wheels.
I have a feeling that this implementation actually causes
a kind of lag compared to basing it off of force at CG
or force at drivers SOP.

Here is why:

The front wheels can still be gripping, and still be
generating lateral force, even when the back wheels
have lost traction and are sliding.  It will take a
moment for the front wheels to also lose traction,
whether it's a slide, or a spin or whatever.  It seems
like a mid-car position would give you a better idea
of what the whole car is really doing.  This method
should give a better feel for when the car is getting
loose.

It would be great to be able to test this idea in RACER.
--
Pat Dotson


 > >When you move aft along the wheelbase to the driver position, then
the
 > >onset of lateral force lags a bit, because the whole car has to
rotate.
 >
 > Isn't the lateral acceleration on every part of the car the same?
 > (obviously the lateral velocity isn't, but the acceleration seems
 > untied to a position in the car).
 > That would suggest no more lag than usual, since even though
 > rotational *velocity* might need a buildup, the *acceleration* of the
 > car body doesn't (and that's where you derive the FF from).

Ice

High End Force Feedback On The Way...

by Ice » Mon, 28 Jan 2002 23:13:14

Nah, doesn't have an ashtray...

Ice D


> You must own a hyperstimulator :)


> > My god, we are still talking about wheels that shake a bit as you drive,
> > right? Never mind, carry on... ;)

> > Ice D





> > > >> Instead of basing FF effects on steering mechanism forces, simply
tie
> > FF
> > > >> to the lateral g-forces acting on the driver?  Then you get the
> > > >> seat-of-the-pants feel that is missing from a computer simulation.

Tony Whitle

High End Force Feedback On The Way...

by Tony Whitle » Mon, 28 Jan 2002 21:21:32

"David L. Cook" wrote...
> Dave - who do you program for?



<snip>

Dave was lead programmer for MGI's Viper Racing and as loyalty seems to be a
feature of sim programmers he's probably still there

Tony Whitley
Lead programmer, mobile phones :-(

Dave Henri

High End Force Feedback On The Way...

by Dave Henri » Tue, 29 Jan 2002 03:05:34

"Tony Whitley" <> Dave was lead programmer for MGI's Viper Racing and as
loyalty seems to be a

   You are correct sir.  Dave has been the guy who's pumped out the patches
for Nascar Heat and done quite a-bit of troubleshooting too.  Thanks sir...
ps...any of you seen the mod community springing up around Heat?  Really
taking off thanks to the patch additions Dave included.
dh

Ruud van Ga

High End Force Feedback On The Way...

by Ruud van Ga » Tue, 29 Jan 2002 22:19:47

On Sat, 26 Jan 2002 23:48:25 -0500, Pat Dotson


>F1 2001 FF is based on lateral force at the front wheels.
>I have a feeling that this implementation actually causes
>a kind of lag compared to basing it off of force at CG
>or force at drivers SOP.

>Here is why:

>The front wheels can still be gripping, and still be
>generating lateral force, even when the back wheels
>have lost traction and are sliding.

Yes, but the acceleration you get is based on the total grip on all
four wheels, so I was going to take the body acceleration (not the
lateral forces on just the front wheels).
And using the body (linear) acceleration, all 4 wheels are accounted
for, so there's no lag then.

Ruud van Gaal
Free car sim  : http://www.marketgraph.nl/gallery/racer/
Pencil art    : http://www.marketgraph.nl/gallery/

Doug Millike

High End Force Feedback On The Way...

by Doug Millike » Wed, 30 Jan 2002 00:06:47


> Yes, but the acceleration you get is based on the total grip on all
> four wheels, so I was going to take the body acceleration (not the
> lateral forces on just the front wheels).
> And using the body (linear) acceleration, all 4 wheels are accounted
> for, so there's no lag then.

Yaw rate does not depend on measurement position, lateral acceleration does.

Look at the simple 2-dof equations--this is enough of a hint<grin>.


rec.autos.simulators is a usenet newsgroup formed in December, 1993. As this group was always unmoderated there may be some spam or off topic articles included. Some links do point back to racesimcentral.net as we could not validate the original address. Please report any pages that you believe warrant deletion from this archive (include the link in your email). RaceSimCentral.net is in no way responsible and does not endorse any of the content herein.