The netBike video looks very interesting but I don't see how a motorbike sim would work given the typical controllers on a PC. In addition to the normal brake, accelerator, clutch and steering axes, you would also need a controller that would allow you to shift body weight in 3D space and that would require at least 3 degrees of freedom and preferably 6, 12 or even 18 dof depending on how many mass centres are used to model the body and how each of them moves in 3D space. I just don't see how you could do that in a user-friendly manner using existing controllers.
They're not classic chicanes, which are intended to slow cars down. Instead, they funnel traffic. It's easy to go through most of the Fern Bay chicanes without braking or lifting when you're by yourself. It's not so easy when you're racing side by side with someone.
Excellent, thanks for those pics, Gentlefoot. I'll have to see if I can get my hands on that book too.
Nothing secret, just pottering around with a little hobby project to keep my C++ and OpenGL skills current. I'm taking the opportunity to learn something new at the same time and, given my interest in motorsports, introductory vehicle dynamics seemed a logical choice.
Thanks Speedy Pro, that makes a lot of sense. As the control arms tend near parallel, the instant centre tends towards infinity and the line projected from the tyre contact patch to the instant centre tends to flatten along the ground. This builds on what Shotglass said above. In practice, I think I'll treat parallel control arms as a special case and either perturb one slightly so that they're no longer parallel and process as usual or just set the roll centre automatically in those cases.
Last edited by BuddhaBing, .
Well, that approach is taken straight from the book 'Race Car Vehicle Dynamics' by Milliken and Milliken. It's considered to be one of the canonical books on vehicle dynamic theory so I'm reluctant to say that the approach they've taken is wrong.
The attached figure is taken from the book and demonstrates how a roll centre is derived using this approach. The instant centre for each wheel is calculated and a line is then projected from the middle of each wheel's tyre-ground contact patch through its instant centre. The intersection of these two projected lines is the roll centre.
As the chassis moves, the suspension geometry will change and the instant centres and roll center will move.
The attached figure is taken from the book and demonstrates how a roll centre is derived using this approach. The instant centre for each wheel is calculated and a line is then projected from the middle of each wheel's tyre-ground contact patch through its instant centre. The intersection of these two projected lines is the roll centre.
As the chassis moves, the suspension geometry will change and the instant centres and roll center will move.
Thanks, guys.
The kinematic approach seems a bit questionable, particularly in the way it decouples the lateral and the logitudinal instant centres, so I may explore a force-based approach to deriving roll centres instead.
The kinematic approach seems a bit questionable, particularly in the way it decouples the lateral and the logitudinal instant centres, so I may explore a force-based approach to deriving roll centres instead.
Depends very much where you live. The average sale price for a house in Seattle a few months ago was $520,000 USD. It's even higher in some areas. The average for all of the US is $220,000 USD.
It's a similar picture here in Canada. In Vancouver, the average sale price is $560,000 CAD while the average for all of Canada is $335,000 CAD.
By contrast, the average house price in the UK is £200,000 GBP with London being the most expensive market, averaging £310,000 GBP.
And for comparison, here's a table of the average individual income level for the US, Canada and the UK.
Average income (individual):
US: $23,535USD (2004)
UK: £23,600 GBP (2006)
Canada: $25,400 CAD (2005)
Sources::
Housing :-
Canada: http://www.cbc.ca/money/story/2006/12/14/homesales.html
US: http://www.realtor.org/press_r ... using_prices_recover.html
UK: http://www.hbosplc.com/economy ... usePriceIndexJune2007.pdf
Income :-
US: http://pubdb3.census.gov/macro/032005/hhinc/new02_001.htm
UK: http://www.statistics.gov.uk/pdfdir/ashe1006.pdf
Canada: http://www40.statcan.ca/l01/cst01/famil105a.htm?sdi=income
Taxation rates alone are misleading since they don't account for other factors such as user fees, insurance premiums and other costs of living. I've lived and worked in Canada, the UK and the US and, in my experience, when you roll all the taxes, user fees, insurance premiums and other costs of living together, you spend roughly the same percentage of your income for an equivalent standard of living with equivalent levels of service.
Vehicle Dynamics: Kinematic Roll Centre
I have a question about deriving kinematic roll centres and instant centres that I hope someone here can help me with.
When calculating kinematic roll centres, you first derive the instant centre. The instant centre is the intersection of a line extending through the suspension's upper control arm (UCA) and another line extending through its lower control arm (LCA). This is shown in the attached diagrams. In the first two cases, calculating the instant centre is straightforward. However, in the third case, the UCA and LCA are parallel to one another and the two lines never intersect. How is the instant centre - and by extension the roll centre - derived in this situation?
Thanks.
When calculating kinematic roll centres, you first derive the instant centre. The instant centre is the intersection of a line extending through the suspension's upper control arm (UCA) and another line extending through its lower control arm (LCA). This is shown in the attached diagrams. In the first two cases, calculating the instant centre is straightforward. However, in the third case, the UCA and LCA are parallel to one another and the two lines never intersect. How is the instant centre - and by extension the roll centre - derived in this situation?
Thanks.
Niels, the C6 Corvette is easily my favourite mod for rFactor and I'm glad to see that you and some1 are still tweaking and refining it. Can't wait to try out the new version when it's released!
He doesn't post very often but when he does it's usually in the GTR2 forums. The only time I've seen him post in the rFactor forums were in a thread cross-posted to several places where he asked painters to submit designs for his team's 2007 paint scheme.
re: real (racing) drivers and sims. Here's what Sean Edwards (2006 FIA GT3 champion, current FIA GT2 driver) has to say on the matter:
"... my dad (ex F1 driver and Group C driver) can't drive sims to save his life!
But I have been playing sims since I was young, just like I have been racing in real life since I was young, so I have learnt both ways of feeling the car and therefor do well in both real life and on sims
I have won countless leagues and most races I do online. I am also 2nd overall in the FIA GT3 championship, so I guess im doing well in both!
Most racing drivers can't drive sims, but ones who have been doing it for a long time can and are normally very good at it. Another example is Stuart Moseley (LMS Works Radical driver), he plas sims a lot and we battle it out for league victories, so he is just as good as me at sim racing.
But you are right, in most cases, real drivers can't drives sims without a lot of praccy because of the "feel" aspect."
- From http://forum.rscnet.org/showthread.php?p=3167202#post3167202
"... my dad (ex F1 driver and Group C driver) can't drive sims to save his life!
But I have been playing sims since I was young, just like I have been racing in real life since I was young, so I have learnt both ways of feeling the car and therefor do well in both real life and on sims
I have won countless leagues and most races I do online. I am also 2nd overall in the FIA GT3 championship, so I guess im doing well in both!Most racing drivers can't drive sims, but ones who have been doing it for a long time can and are normally very good at it. Another example is Stuart Moseley (LMS Works Radical driver), he plas sims a lot and we battle it out for league victories, so he is just as good as me at sim racing.
But you are right, in most cases, real drivers can't drives sims without a lot of praccy because of the "feel" aspect."
- From http://forum.rscnet.org/showthread.php?p=3167202#post3167202
I wouldn't put much faith in a comparison which just lists the feature sets of various sims. Just because a given feature is implemented does not mean that it has been implemented well. Also, look at rFactor: all mods use the same core engine yet there is a huge range of quality among them, from very poor to excellent.
I have or used to have most of the recent PC racing sims. They all have things that they do well and, on the flip side, they all have things that they do not so well. None of them are perfect and trying to come up with some definitive objective measure of which is best or whether one sim is better than another is a pointless exercise. It depends on what you're looking for.
Finally, take opinions about what is good or bad with a grain of salt. A quick look at the number of rFactor mods which are acclaimed as having excellent physics/handling yet which have glaring errors (e.g. in their suspension geometry, pushrods on one side pointing in the wrong direction is a relatively common error) should be enough to make it clear that opinions can't trusted. You know what you like, that's all that really matters.
I have or used to have most of the recent PC racing sims. They all have things that they do well and, on the flip side, they all have things that they do not so well. None of them are perfect and trying to come up with some definitive objective measure of which is best or whether one sim is better than another is a pointless exercise. It depends on what you're looking for.
Finally, take opinions about what is good or bad with a grain of salt. A quick look at the number of rFactor mods which are acclaimed as having excellent physics/handling yet which have glaring errors (e.g. in their suspension geometry, pushrods on one side pointing in the wrong direction is a relatively common error) should be enough to make it clear that opinions can't trusted. You know what you like, that's all that really matters.
Last edited by BuddhaBing, .
Speaking of which, Niels, how confident are you about the moments of inertia you've derived for the C6? The reason I ask is that I've been looking at the NHTSA (National Highway Traffic Safety Administration) inertia database (http://www-nrd.nhtsa.dot.gov/v ... ertia_database_metric.pdf) and the MOIs that you're using for the C6 look to be on the low side. They're even lower than the MOIs for a 1998 Dodge Neon, which is a much smaller and lighter car. The Neon weighs 12199N (1244kg) and has MOIs of (1748, 441, 1945) kg-m^2 in the pitch, roll and yaw axes respectively. Your C6, according to the HDV, weighs 14926N (1522kg) and has MOIs of (1650,395,1942) kg-m^2.
Without knowing how the various curve modifiers, grip multipliers and all the rest are applied in real-time, no, I've no idea. It could be anything.
ISI added a pneumatic trail parameter to the tyre files in the most recent rFactor patch so clearly they were still working on tyre modeling until as recently as a few months ago. Pneumatic trail is a pretty fundamental characteristic and its late inclusion - in explicit form, at least, it may have been a derived value before - does make me scratch my chin a bit.
Add to that the fact that several rFactor tyre files use longitudinal traction and braking slip curves that look suspiciously like lateral force curves also makes me wonder how well tyres are actually modeled in-game.
That's the point really. The curves go out that far - for whatever reason - but that doesn't mean that such extremely high slip angles are ever going to be encountered in-game. Given that, characterizing the curves by their shape/behaviour at such extremes isn't very useful.
Expanding on that point, let's compare an rFactor slip angle v lateral force curve against the real-world data shown in the following figure taken from Dr Kasprzak's paper:
http://i9.tinypic.com/4m94e1c.jpg
Note that the data has been non-dimensionalized and that the black plot represents a characteristic curve that has been derived from the measured data. Eyeballing the curve, we see that the peak occurs at a (non-dimensional) slip angle of around 1.8 and that it falls to around 90% of the peak value at a slip angle of about 6. In other words, at a slip angle roughly 3-4 times higher than the slip angle at which the peak occurs, the lateral force falls to around 90% of peak.
Now, let's compare that to a slip angle v lateral force curve from rFactor. In this case, I'm using the 1997_GT1_Goodyear_Tires.tbc file from ISI's Panoz car set. Here's that curve:
http://i11.tinypic.com/6azk36t.jpg
The actual data is shown in columns on the left side of the image while the right side shows the curve plot.
At first blush, the curve looks quite a bit different. In particular, it falls off all the way down to about 62% of peak at the far right. 62%? That's a huge drop off, right? The curve can't be right!
Well, this initial impression is misleading, since the curve extends way out to a slip angle of around 120°. Let's instead look at the behaviour of the curve in the same range of slip angles as the real-world data we just examined. Looking at the data table, we see that the peak occurs at step 11. At step 33 (3 times the slip angle at which the peak occurs), the lateral force is still 98% of peak and at step 44 (4 times the slip angle at which the peak occurs), the lateral force is still 95% of peak.
In other words, the rFactor curve actually shows a smaller drop off than the real-world tyre curve within the same range of slip angles.
Note that I'm not making any value judgements about the curves nor arguing that the rFactor curves are correct/accurate - for example, real-world data often tends to flatten off more quickly than the rFactor curves, which generally continue to drop off. I'm only trying to show that dismissing curves out of hand based on their behaviour at extremely high slip angles might be a bit hasty.
Last edited by BuddhaBing, .
I'd like to clarify that point somewhat.
While there is data that shows little if any post-peak drop off (e.g. data obtained by engineers at the Yokohama R&D department in this paper: http://scitation.aip.org/journ ... t/vol_34/iss_2/100_1.html), it's not strictly true that significant drop off is very unusual nor that it occurs only at relatively high inclinations. For example, here are some graphs that show substantial drop off even at low inclination angles:
http://i14.tinypic.com/4mmzceh.jpg
http://i14.tinypic.com/6b00wvn.jpg
And here are some graphs showing similar drop off across a range of load and inclination angles for another type of tyre (Formula SAE tyre):
http://i9.tinypic.com/4m94e1c.jpg
http://i11.tinypic.com/668z7g5.jpg
These graphs were taken from Edward Kasprzak's Ph.D. thesis entitled 'Extension of the Nondimensional Tire Theory to General Operating Conditions'. Dr Kasprzak now works for Milliken Research Associates.
I contacted Dr Kasprzak a couple of weeks ago to ask some questions about the data presented in these graphs as well as some general questions about post-peak drop offs. He responded, in part, that "[a]ll tires exhibit falloff after the peak. Depending on the construction and compound it is more or less severe." and "[t]here's always a falloff, but the amount depends on the specific tire. Even things like inflation pressure and inclination angle affect how much falloff there is."
While many of the ISI tyre curves show eventual drop offs down to as low as 60-70% of peak, bear in mind that these curves go all the way out to extremely high slip angles i.e. 120°. In the range of slip angles likely to be encountered in-game, the drop off isn't as severe. Also worth bearing in mind is that the ISI tyre curves - including the location of the peak and the rate and extent of post peak drop off - are modified by load, camber and other factors which means that the raw curve data in the files isn't necessarily indicative of how the tyres will behave in-game after all the modifiers are applied.
That said, I do agree that the way the tyres behave in most of the cars in rFactor just doesn't feel right.
Last edited by BuddhaBing, .
Reason : Wrong URL for Yokohama data
Yes, the carFactory developer (Kangaloosh) has stated that he had to kludge a few parameters to get the desired behaviour in-game - which he checks by comparing Motec telemetry generated by the car in-game against his own calculations as to how the car would behave given a set of inputs. This certainly doesn't speak well for the way that rFactor interprets/handles physical input parameters.
A handful of newly released mods for rFactor have made me change my mind about the quality of the underlying physics modeling in rFactor. After trying most of the mods that were considered to be the cream of the crop and finding the handling of all of them to be, at best, unconvincing and, at worst, outright strange, I was convinced that the root problem was the physics modeling in rFactor itself and dismissed it as deeply flawed.
However, the newly released Caterhams (by Major Parts, cubits and Kangaloosh) and the C6 Corvette (by some1 and Niels), together with the RealFeel force-feedback plugin by TechAde, have changed my mind. The suspension and some other physical parameters in these mods were created using a new tool called carFactory, which takes the real physical properties of the car being modeled and generates the appropriate suspension and physical parameters for the mod. The end result, as shown by the handful of mods that have been designed using carFactory, is a car that handles much more believably. Further, the RealFeel plugin generates force feedback from the forces acting on the steering arm, similar to the way force feedback is generated in LFS. With well-designed suspension, the force feedback is excellent and the cars actually feel like real cars. Throwing a Caterbusa or the C6 around a well-designed track is an absolute delight and the driving experience is, in my opinion, at least as good as LFS.
Unfortunately, the RealFeel plugin highlights how poorly designed most of the mods for rFactor are, and that includes both cars and tracks. Put a well-designed car mod on a poorly designed track and the flaws in the track become very apparent. At this point, there are only a small handful of car and track combinations that really showcase how good rFactor can be. carFactory and RealFeel have raised the bar. I hope that they quickly become accepted standards and that more and more quality mods are released.
It's a bit sad that it has taken over 2 years for good quality, believable mods to be released for rFactor. Much of the blame for this can be laid squarely at ISI's feet since they didn't release proper design tools along with the game. In fact, considering the relatively poor quality of the cars designed by ISI themselves, it's questionable whether their own in-house content developers have access to proper design tools either.
However, the newly released Caterhams (by Major Parts, cubits and Kangaloosh) and the C6 Corvette (by some1 and Niels), together with the RealFeel force-feedback plugin by TechAde, have changed my mind. The suspension and some other physical parameters in these mods were created using a new tool called carFactory, which takes the real physical properties of the car being modeled and generates the appropriate suspension and physical parameters for the mod. The end result, as shown by the handful of mods that have been designed using carFactory, is a car that handles much more believably. Further, the RealFeel plugin generates force feedback from the forces acting on the steering arm, similar to the way force feedback is generated in LFS. With well-designed suspension, the force feedback is excellent and the cars actually feel like real cars. Throwing a Caterbusa or the C6 around a well-designed track is an absolute delight and the driving experience is, in my opinion, at least as good as LFS.
Unfortunately, the RealFeel plugin highlights how poorly designed most of the mods for rFactor are, and that includes both cars and tracks. Put a well-designed car mod on a poorly designed track and the flaws in the track become very apparent. At this point, there are only a small handful of car and track combinations that really showcase how good rFactor can be. carFactory and RealFeel have raised the bar. I hope that they quickly become accepted standards and that more and more quality mods are released.
It's a bit sad that it has taken over 2 years for good quality, believable mods to be released for rFactor. Much of the blame for this can be laid squarely at ISI's feet since they didn't release proper design tools along with the game. In fact, considering the relatively poor quality of the cars designed by ISI themselves, it's questionable whether their own in-house content developers have access to proper design tools either.
Last edited by BuddhaBing, .
There is but as you say there is also a lot of data which shows no or very little drop off after the peak. For instance, here's a paper by some engineers in the R&D department at Yokohama which shows measured data for two tyres: a 195/65R15 and a 205/55R16. Neither shows a significant drop off after the peak. Note: the paper is in two parts; the second part shows the slip angle v lateral force curves.
http://scitation.aip.org/journ ... l_34/iss_2/84_1.html#div8
http://scitation.aip.org/journ ... ol_34/iss_2/100_1.html#F6
I've also seen a paper which included some data from Bridgestone which showed little to no drop off after peak but I can't lay my hands on that paper now.
At some point, I'm sure the materials and physics will be sufficiently well understood that physical models can be developed which accurately predict tyre behaviour throughout the full range of operating conditions.
My wife and I live on 45 acres of mostly forested land which has parkland on 3 sides yet is only a 20 minute drive from downtown Victoria, BC. These photos are pretty representative of the property. At the back of the property is a fairly high hill which has excellent panoramic views of the Gulf and San Juan Islands and the Olympic mountain range in Washington state. It's our own little piece of paradise.
Todd, it's available -- for a fee -- from the link at the bottom of this page: http://www.mae.buffalo.edu/research/phd-kasprzak.shtml If you want to avoid the fee, you could probably ask him if he has a copy he can forward to you. It's an interesting read and well worth a look for those interested in the subject matter.
I too would be very interested to hear Doug Milliken's take on the data shown in those figures and, more generally, on the various approaches used to model tyres in these consumer level simulations.
I haven't heard back from Kasprzak yet but, out of curiousity, I had a look as his PhD dissertation which is titled 'Extension of the Nondimensional Tire Theory to General Operating Conditions'. It deals extensively with tyre measurement and tyre data modeling and includes a lot of slip angle vs lateral force curves, some showing normalized data others showing raw data. Attached are just a handful of samples. Note that each shows a significant drop off after the peak.
http://i11.tinypic.com/4l96wyh.jpg
http://i15.tinypic.com/4u2oyzm.jpg
http://i8.tinypic.com/52naku9.jpg
http://i14.tinypic.com/6bu2iyt.jpg
Curious about where the data came from and how it was obtained, I checked the dissertation and found the following:
"Throughout this chapter and the remainder of this dissertation the development takes place with validation against two different sources of measured tire data. The first is data collected by the Formula SAE Tire Test Consortium, an all-volunteer organization co-founded and directed by the
dissertation author for the purpose of providing high quality tire data to students for use with their Formula SAE designs, student projects and academic research. The consortium is discussed extensively in Chapter 6. Formula SAE tires are small (less than 20 inch diameter) tires, normally bias ply (not radial), for use in lightweight (under 800 lb) single seat race cars. Typical tire loads are always below 300 lb.
The second set of tire data is provided by Milliken Research Associates, Inc. It is a radial racing tire used in a professional racing series from a few years ago. At the request of Milliken Research’s customer, for whom the data was initially collected, the specific identity of the tire will remain proprietary. It is a full-size racing tire with typical tire loads ranging from a few hundred pounds to over 3000 lb. Both data sets were produced at the Calspan Tire Research Facility." - Kasprzak, pp79
Given that some of the data shown in the graphs came from Milliken Research Associates and clearly shows a significant drop off after the peak, I am very curious about Todd's assertion that Doug Milliken told him several times "that [post-peak drop off] is just not observed with car tires in the lateral direction". The data clearly shows that it does.
http://i11.tinypic.com/4l96wyh.jpg
http://i15.tinypic.com/4u2oyzm.jpg
http://i8.tinypic.com/52naku9.jpg
http://i14.tinypic.com/6bu2iyt.jpg
Curious about where the data came from and how it was obtained, I checked the dissertation and found the following:
"Throughout this chapter and the remainder of this dissertation the development takes place with validation against two different sources of measured tire data. The first is data collected by the Formula SAE Tire Test Consortium, an all-volunteer organization co-founded and directed by the
dissertation author for the purpose of providing high quality tire data to students for use with their Formula SAE designs, student projects and academic research. The consortium is discussed extensively in Chapter 6. Formula SAE tires are small (less than 20 inch diameter) tires, normally bias ply (not radial), for use in lightweight (under 800 lb) single seat race cars. Typical tire loads are always below 300 lb.
The second set of tire data is provided by Milliken Research Associates, Inc. It is a radial racing tire used in a professional racing series from a few years ago. At the request of Milliken Research’s customer, for whom the data was initially collected, the specific identity of the tire will remain proprietary. It is a full-size racing tire with typical tire loads ranging from a few hundred pounds to over 3000 lb. Both data sets were produced at the Calspan Tire Research Facility." - Kasprzak, pp79
Given that some of the data shown in the graphs came from Milliken Research Associates and clearly shows a significant drop off after the peak, I am very curious about Todd's assertion that Doug Milliken told him several times "that [post-peak drop off] is just not observed with car tires in the lateral direction". The data clearly shows that it does.
FGED GREDG RDFGDR GSFDG