Car physics; driving backwards

Hi all, I've been distracted by some fancy stuff, so now I've got some engine sound and skid sounds working on the PC (using DirectSound3D, very nice for the feel), but I'm back at some physics again. (although thinking hard too about some basic My-First-Sony track format) I have a problem when the car is moving backwards, and my wheels starting out not rotating (car is heading straight forward, no turning). Hm, it's getting late here, and my brains are stalling more and more. ;) To get a longitudinal force (in tire coords, which for this situation is the same as car coordinates and even world coordinates), I do: if(slipRatio>0) coeff=tractionCurve->GetValue(slipRatio) else coeff=-brakingCurve->GetValue(-slipRatio) This assumes that if slipRatio>0, a tractive force is present, and when slipRatio<0, a braking force is present. That seems ok. (and all my curves are in the 1st quadrant; with positive X and Y values; see my website if you want, the curve editor is downloadable at this point). But when I start out with a car that is thrown backward (vz=-5, would be vx=-5 in the SAE axis system), here's what happens: t=0; SR=-1, because the wheels are not rotating, so according to the Calspan TIRF slip ratio: SR=(ang.vel.*loaded_radius/velocity)-1. Which result in SR = (0*.25/-5)-1 = -1. (my tires have a radius of .25m for example) I don't think that's very intuitive, since the wheel is basically moving forward faster than the wheel's actual velocity (so I would expect a slipRatio>0). Even though the wheel is not rotating. Because of SR=-1, I get a force that pushes the car even harder backward. So the car endlessly accelerates backwards. Ofcourse, that isn't right. But to make it right, I'd have to come up with a slip ratio >0, but that means modifying the slip ratio formula for the situation where the wheel is moving backward. The wheel should start spinning backward. I would say something like the relative speed somewhere, hm. Or is the slipratio right, and must I look at the wheel's longitudinal velocity to decide which way the (long.) force will point? I'll have to sleep on it; getting sleepy after some long hours of getting out some other problems and getting the skidding sound in. But if you see the problem offhand, that would help. :) Thanks, Ruud van Gaal, GPL Rank +53.25 MarketGraph : http://www.marketgraph.nl Pencil art : http://www.marketgraph.nl/gallery/ Car simulation: http://www.marketgraph.nl/gallery/racer/

Quoted text (7 lines)

>> I had to use a few IF statement to really get the slip ratios to calculate >>right. I'll read your post again and get back to you later. It's party >time >>at First Avenue, Minneapolis, MN tonight!! > >Have fun partying :) >

I'm back. Thanks, I had a fun time, and woke up with a nice hangover :0P !! This is what I did to solve it. Of course, calculating the rpm of the tires is inefficient and unneccessary, so use rotational speed instead. This code is in PowerBasic format, so you'll have to convert to C++. ABS() is absolute value. To figure this out and get the slip ratios correct (I think this is right ;-)), I drew a bunch of wheels and manually gave them free_rolling and true rotational velocities of .1, 1, -.1, and -1 in different combinations, but yes, calculating slip ratio differently based on the situation fixes the problem. All I do is use some ABS() or -ABS() operations on the slip ratio for the most part. Of course, you'll want to use your slip ratio definition instead of mine :-) IF TIRE_RPM(TIRE) < 0 AND TIRE_FREE_RPM < 0 AND TIRE_RPM(TIRE) < TIRE_FREE_RPM THEN SLIP_RATIO(TIRE) = -ABS(TIRE_RPM(TIRE)/TIRE_FREE_RPM - 1) END IF IF TIRE_RPM(TIRE) > 0 AND TIRE_FREE_RPM < 0 THEN SLIP_RATIO(TIRE) = ABS(TIRE_RPM(TIRE)/TIRE_FREE_RPM - 1) END IF Todd Wasson --- Performance Simulations Drag Racing and Top Speed Prediction Software http://PerformanceSimulations.Com

My previous post has nothing to do with actual tire use or operation, only with the equation that is used to determine Slip Ratio on the TIRF test machine -- I believe that the equation is the thing that is causing problems... Now, for Real tires: Some tires do not like to be mounted "wrong way 'round", because the tread splice may open up. These (should) have directional arrows on the sidewall. The raw tread rubber is typically cut from a long extrusion and wrapped around the tire, the two ends butted with an angled splice or 'scarf' joint. The ends of the tread rubber are vulcanized together when the tire is cured, but at least in some cases, the joint may not be quite as strong as the rest of the tread rubber. Many years ago we had some go-kart slicks that came without arrows, and luck of the draw, we got one mounted the wrong way...and the splice started to open up after some very hard acceleration (200cc shifter kart) and sustained cornering tests on a circle. We re-mounted it the other way around and the splice didn't open up any further, so the tire was still useful (but we never ran at really high speeds). -- Doug Doug Milliken <[email hidden]> On 30 Sep 2000, David wrote:

Quoted text (49 lines)

> > > Doug, > > I have a question, why would tires slip be different from moving from one > direction to the reverse. I could understand due to some items like castor > adjustment being negitive instead of positive and the toe out now being toe > in, but it that what would effect the car, but I was just wondering why it > would be that way. I would not think its due to tire construction, but > maybe I am wrong here. I do know the Hoosier slicks I am running do not > care to be remounted and ran in a different direction, but I always thought > that was due to the heat cycling of the tire and well the rubber migration > you would get when they are up to temp when racing. > > You just have me wondering on this. > > Dave Robinson > Just a little Solo ll Racer > Doug Milliken <[email hidden]> wrote in article > <[email hidden]>... > > > > > > > TIRF Slip Ratio may not be the correct choice. My understanding of the > > history: (I know something about this machine...I was over at TIRF > yesterday) > > > > TIRF SR was developed because the TIRF test machine has the capability to > > vary slip ratio in real time (variable-speed hydraulic-motor-drive to the > > belt-speed and also to the wheel-rpm). The designers used loaded radius > > because it was easy and quick to measure for real-time control. The TIRF > > machine is never used to perform SR tests with the roadway in reverse, so > > the definition was never required to work in reverse. > > > > The roadway can reverse, and this is used to measure at tire assymmetry > > (conicity and plysteer), but only at very low speeds. High speed in > > reverse would probably cause trouble with the tracking of the steel-belt > > roadway. More on TIRF at <http://www.tireresearch.com>. > > > > -- Doug > > Douglas Milliken <[email hidden]> > > Milliken Research Associates Inc. > > > > ps. Thanks to all of you that bought a copy of RCVD it has now > > gone into a 5th printing, a new record for SAE text books. > > > > > >

I had to use a few IF statement to really get the slip ratios to calculate right. I'll read your post again and get back to you later. It's party time at First Avenue, Minneapolis, MN tonight!! Todd Wasson --- Performance Simulations Drag Racing and Top Speed Prediction Software http://PerformanceSimulations.Com

On Fri, 29 Sep 2000, Ruud van Gaal wrote: .....

Quoted text (10 lines)

> my curves are in the 1st quadrant; with positive X and Y values; see > my website if you want, the curve editor is downloadable at this > point). > But when I start out with a car that is thrown backward (vz=-5, would > be vx=-5 in the SAE axis system), here's what happens: > > t=0; SR=-1, because the wheels are not rotating, so according to the > Calspan TIRF slip ratio: SR=(ang.vel.*loaded_radius/velocity)-1. > Which result in SR = (0*.25/-5)-1 = -1. (my tires have a radius of > .25m for example)

TIRF Slip Ratio may not be the correct choice. My understanding of the history: (I know something about this machine...I was over at TIRF yesterday) TIRF SR was developed because the TIRF test machine has the capability to vary slip ratio in real time (variable-speed hydraulic-motor-drive to the belt-speed and also to the wheel-rpm). The designers used loaded radius because it was easy and quick to measure for real-time control. The TIRF machine is never used to perform SR tests with the roadway in reverse, so the definition was never required to work in reverse. The roadway can reverse, and this is used to measure at tire assymmetry (conicity and plysteer), but only at very low speeds. High speed in reverse would probably cause trouble with the tracking of the steel-belt roadway. More on TIRF at <http://www.tireresearch.com>. -- Doug Douglas Milliken <[email hidden]> Milliken Research Associates Inc. ps. Thanks to all of you that bought a copy of RCVD it has now gone into a 5th printing, a new record for SAE text books.

Doug, I have a question, why would tires slip be different from moving from one direction to the reverse. I could understand due to some items like castor adjustment being negitive instead of positive and the toe out now being toe in, but it that what would effect the car, but I was just wondering why it would be that way. I would not think its due to tire construction, but maybe I am wrong here. I do know the Hoosier slicks I am running do not care to be remounted and ran in a different direction, but I always thought that was due to the heat cycling of the tire and well the rubber migration you would get when they are up to temp when racing. You just have me wondering on this. Dave Robinson Just a little Solo ll Racer Doug Milliken <[email hidden]> wrote in article <[email hidden]>... > > > TIRF Slip Ratio may not be the correct choice. My understanding of the > history: (I know something about this machine...I was over at TIRFyesterday)

Quoted text (21 lines)

> > TIRF SR was developed because the TIRF test machine has the capability to > vary slip ratio in real time (variable-speed hydraulic-motor-drive to the > belt-speed and also to the wheel-rpm). The designers used loaded radius > because it was easy and quick to measure for real-time control. The TIRF > machine is never used to perform SR tests with the roadway in reverse, so > the definition was never required to work in reverse. > > The roadway can reverse, and this is used to measure at tire assymmetry > (conicity and plysteer), but only at very low speeds. High speed in > reverse would probably cause trouble with the tracking of the steel-belt > roadway. More on TIRF at <http://www.tireresearch.com>. > > -- Doug > Douglas Milliken <[email hidden]> > Milliken Research Associates Inc. > > ps. Thanks to all of you that bought a copy of RCVD it has now > gone into a 5th printing, a new record for SAE text books. > >

On 30 Sep 2000 00:13:56 GMT, [email hidden] (J. Todd Wasson) wrote: > I had to use a few IF statement to really get the slip ratios to calculate >right. I'll read your post again and get back to you later. It's party time >at First Avenue, Minneapolis, MN tonight!! Have fun partying :) Ruud van Gaal, GPL Rank +53.25 MarketGraph : http://www.marketgraph.nl Pencil art : http://www.marketgraph.nl/gallery/ Car simulation: http://www.marketgraph.nl/gallery/racer/

>I assume btw that you missed out the other slip ratio calculation, >when TIRE_RMP(TIRE)>0 and TIRE_FREE_RPM>0 (when slip ratio is like the >regular TIRF/SAE formula's). Yes, the regular equation is processed before these IF's are done. >Strange though that you do ABS(T/T-1) instead of >ABS(T/T)-1. But that's just a gut feeling that's probably wrong. EGAD!!!!! You're right! Thanks for catching that bug! Late nights have a way of making you do that kind of thing :-) Will have to fix immediately! Cheers, Todd Wasson --- Performance Simulations Drag Racing and Top Speed Prediction Software http://PerformanceSimulations.Com

Wow, it never ceases to amaze me the high caliber of people that are lurking out there reading newsgroups! Thanks for writing RCVD Doug! I'll have to dust off my copy and do some more reading. -NKN Todd www.toddlamb.com

Quoted text (7 lines)

> -- Doug > Douglas Milliken <[email hidden]> > Milliken Research Associates Inc. > > ps. Thanks to all of you that bought a copy of RCVD it has now > gone into a 5th printing, a new record for SAE text books. >

On 30 Sep 2000 23:46:22 GMT, [email hidden] (J. Todd Wasson) wrote: > I'm back. Thanks, I had a fun time, and woke up with a nice hangover :0P !! Well that's what weekends are for. I have been creating a small leaflet for a wedding tomorrow. So ouch, no work done on the carsim. Although I did buy a book about DirectX/3D math/BSP trees/Portals/Network UDP etc. Very interesting for future stuff, although BSP trees seem a bit unsuited for open area racing tracks, for example. ...

Quoted text (7 lines)

> IF TIRE_RPM(TIRE) < 0 AND TIRE_FREE_RPM < 0 AND TIRE_RPM(TIRE) < >TIRE_FREE_RPM THEN > SLIP_RATIO(TIRE) = -ABS(TIRE_RPM(TIRE)/TIRE_FREE_RPM - 1) > END IF > IF TIRE_RPM(TIRE) > 0 AND TIRE_FREE_RPM < 0 THEN > SLIP_RATIO(TIRE) = ABS(TIRE_RPM(TIRE)/TIRE_FREE_RPM - 1) > END IF

Thanks, I'll try it out. Strange that I haven't seen a formula containing both directions, and strange that I didn't see the restraint in RCVD. Probably with a SGN() here & there. I (think I) see that you're using the slip ratio in such a way that if you drive backwards, you effectively reverse the tire (mounting) and the slip ratio for my situation (throwing the car backwards while the wheels are not moving) becomes a tractive force. In other words, a wheel rotating backwards is just flipped around in the slip ratio numbers. Even though Doug Milliken (honored btw to have Doug with us as an expert commentator, amongst others like Doug (what's in a name) Arnao!), even though Doug Milliken stated that mounting a tire in the wrong direction can be bad, I will look at the tire as being symmetrical in the longitudinal direction. For one, I don't want to have nearly equal slip ratio graphs for *reverse* traction and braking, and second, you are not supposed to drive backwards for extensive periods of time in racing, now do you! ;-) I assume btw that you missed out the other slip ratio calculation, when TIRE_RMP(TIRE)>0 and TIRE_FREE_RPM>0 (when slip ratio is like the regular TIRF/SAE formula's). But not to worry, I should be able to figure it out now, now that I know the regular formula works only for specific ranges. Strange though that you do ABS(T/T-1) instead of ABS(T/T)-1. But that's just a gut feeling that's probably wrong. Ruud van Gaal, GPL Rank +53.25 MarketGraph : http://www.marketgraph.nl Pencil art : http://www.marketgraph.nl/gallery/ Car simulation: http://www.marketgraph.nl/gallery/racer/

On Sat, 30 Sep 2000 04:41:37 GMT, Doug Milliken <[email hidden]> wrote: ... >TIRF Slip Ratio may not be the correct choice. My understanding of the >history: (I know something about this machine...I was over at TIRF yesterday) Even though I know the SAE version is better than the TIRF one, I use it because I have no idea how to calculate the *effective* radius of a loaded wheel. My intuition says the *effective* radius (SAE) is always a little more than the *loaded* radius (TIRF), but I wouldn't know a ratio between these two. Probably has a 2nd order effect anyway on the slip ratio. Perhaps doing effectiveRadius=(radius+radiusLoaded)/2 will be largely enough for a car sim. Ruud van Gaal, GPL Rank +53.25 MarketGraph : http://www.marketgraph.nl Pencil art : http://www.marketgraph.nl/gallery/ Car simulation: http://www.marketgraph.nl/gallery/racer/

Quoted text (33 lines)

>Thanks, I'll try it out. Strange that I haven't seen a formula >containing both directions, and strange that I didn't see the >restraint in RCVD. Probably with a SGN() here & there. >I (think I) see that you're using the slip ratio in such a way that if >you drive backwards, you effectively reverse the tire (mounting) and >the slip ratio for my situation (throwing the car backwards while the >wheels are not moving) becomes a tractive force. >In other words, a wheel rotating backwards is just flipped around in >the slip ratio numbers. Even though Doug Milliken (honored btw to have >Doug with us as an expert commentator, amongst others like Doug >(what's in a name) Arnao!), even though Doug Milliken stated that >mounting a tire in the wrong direction can be bad, I will look at the >tire as being symmetrical in the longitudinal direction. >For one, I don't want to have nearly equal slip ratio graphs for >*reverse* traction and braking, and second, you are not supposed to >drive backwards for extensive periods of time in racing, now do you! >Thanks, I'll try it out. Strange that I haven't seen a formula >containing both directions, and strange that I didn't see the >restraint in RCVD. Probably with a SGN() here & there. >I (think I) see that you're using the slip ratio in such a way that if >you drive backwards, you effectively reverse the tire (mounting) and >the slip ratio for my situation (throwing the car backwards while the >wheels are not moving) becomes a tractive force. >In other words, a wheel rotating backwards is just flipped around in >the slip ratio numbers. Even though Doug Milliken (honored btw to have >Doug with us as an expert commentator, amongst others like Doug >(what's in a name) Arnao!), even though Doug Milliken stated that >mounting a tire in the wrong direction can be bad, I will look at the >tire as being symmetrical in the longitudinal direction. >For one, I don't want to have nearly equal slip ratio graphs for >*reverse* traction and braking, and second, you are not supposed to >drive backwards for extensive periods of time in racing, now do you!

I'm also honored to have Doug Arnao and Doug Milliken popping in like this. Thanks to both of you :-) Ruud, I haven't yet defined seperate curves for tractive vs. braking force due to slip ratio, so can't help you there. >I assume btw that you missed out the other slip ratio calculation, >when TIRE_RMP(TIRE)>0 and TIRE_FREE_RPM>0 (when slip ratio is like the >regular TIRF/SAE formula's). Yes, I left out the other calculation :-) > Strange though that you do ABS(T/T-1) instead of >ABS(T/T)-1. But that's just a gut feeling that's probably wrong. My last post appears to have been lost in the depths of cyberspace. If it turns up, you'll see me exclaiming "EGAD!!" and praising you for spotting a bug. However, on second thought, the ABS(t/t-1) is right, as the slip ratio needed to be calculated in that particular situation the same way as normal, but with any potential negative sign ignored. Todd Wasson --- Performance Simulations Drag Racing and Top Speed Prediction Software http://PerformanceSimulations.Com

On 02 Oct 2000 02:48:27 GMT, [email hidden] (J. Todd Wasson) wrote: ...

Quoted text (9 lines)

>>For one, I don't want to have nearly equal slip ratio graphs for >>*reverse* traction and braking, and second, you are not supposed to >>drive backwards for extensive periods of time in racing, now do you! > > I'm also honored to have Doug Arnao and Doug Milliken popping in like this. >Thanks to both of you :-) > > Ruud, I haven't yet defined seperate curves for tractive vs. braking force >due to slip ratio, so can't help you there.

That's ok, it complicates the traction vs. braking ratio a bit, but I'm drawing lots of pictures. It's just I thought I was doing something wrong when suddenly the formula turned out bad for certain situations.

Quoted text (8 lines)

>> Strange though that you do ABS(T/T-1) instead of >>ABS(T/T)-1. But that's just a gut feeling that's probably wrong. > > My last post appears to have been lost in the depths of cyberspace. If it >turns up, you'll see me exclaiming "EGAD!!" and praising you for spotting a >bug. However, on second thought, the ABS(t/t-1) is right, as the slip ratio >needed to be calculated in that particular situation the same way as normal, >but with any potential negative sign ignored.

Yes, it could have been either way, depending. I still have to draw some pictures to see if it relates the same way in my program, with the IF's. Ruud van Gaal, GPL Rank +53.25 MarketGraph : http://www.marketgraph.nl Pencil art : http://www.marketgraph.nl/gallery/ Car simulation: http://www.marketgraph.nl/gallery/racer/