Jump to content


- - - - -

Test Mod Version B00.0


  • Please log in to reply
181 replies to this topic

#16 M Needforspeed

M Needforspeed

    Denny Hulme

  • GPLLinks Team
  • PipPipPipPipPipPipPipPipPipPip
  • 2,453 posts
  • Gender:Male
  • Interests:GPL - start a collection of GPL era die cast racing cars -Skiing
  • Sim interest:GPL

Posted Sep 01 2010 - 05:08 PM

..let 's go with the keyboard affair.Thank you for the others tests, and your impressions and data collection.

I read your graphs. What is the tyre wear percentage ?

  I don 't want to interfere between you and Brr, as you are the only that posted results so far and thank for that..

But one thing I am sure as well as those who worked on carsets physics, is the real Tyres had a very hard compound in the sixties. I have many different countries period mags reports of F1 and F2 races from 1967 .If I cross all the infos on Tyres,the global rule was:

- One or two compound for a team in one given race.Many teams sharing the same compound.

- No Tyres changes for all the event was frequent.Tyres used both for practice and race was seen on many rounds.

-Drivers got their best laps time toward the end of the race.Tyres kept their efficiency all along the race, except on some tracks.Tyre wear was linked more to the driver style, when  noticable, than from track shapes or chassis. Rindt, for instance, was very demanding on tyres on street courses, with his countersteering driving style.

in fact, racing tyres  where still in their infancy, compared with today standards.


not a critic, I ' ll like to know what Brr think can be a reasonable scaling for putting values on tyre wear

#17 Bob Simpson

Bob Simpson

    The answer man

  • Moderator
  • 2,242 posts
  • Gender:Male
  • Location:Ottawa
  • Sim interest:GPL

Posted Sep 01 2010 - 05:25 PM

Since keyboard driving means that you must use brake help, or essentially anti-lock brakes, you can't give them a real torture test.  I don't have time to test this mod myself, but it would be interesting to see what happens with the brakes engaged constantly or at least when the throttle is still applying power in corners as it is with some GPL alien drivers.

#18 John Woods

John Woods

    Be Somebody

  • GPLLinks Team
  • PipPipPipPipPipPipPipPipPipPip
  • 2,175 posts
  • Gender:Male
  • Interests:Too Much Fun
  • Sim interest:GPL

Posted Sep 01 2010 - 05:55 PM

identiti_crisis,

Could you please offer a detailed comphrehensive methodology so others can do exactly what you are doing and generate comparable results?

There have always been (at least) two schools of thought with Grand Prix Legends enthusiasts: "traditionalists" who prefer replicating a particular season as exactly as possible, and "futurist realists" who want to pursue ultimate development of simulation experience. I think neither can survive without the other, together they have assured longevity of Grand Prix Legends, and both groups should learn to live with it or maybe start another thread, on which I will happily support both views.

#19 identiti_crisis

identiti_crisis

    Mike Spence

  • Members
  • PipPipPipPip
  • 31 posts
  • Sim interest:GPL

Posted Sep 01 2010 - 06:22 PM

View PostM Needforspeed, on Sep 01 2010 - 05:08 PM, said:

..let 's go with the keyboard affair.Thank you for the others tests, and your impressions and data collection.

I read your graphs. What is the tyre wear percentage ?

  I don 't want to interfere between you and Brr, as you are the only that posted results so far and thank for that..

But one thing I am sure as well as those who worked on carsets physics, is the real Tyres had a very hard compound in the sixties. I have many different countries period mags reports of F1 and F2 races from 1967 .If I cross all the infos on Tyres,the global rule was:

- One or two compound for a team in one given race.Many teams sharing the same compound.

- No Tyres changes for all the event was frequent.Tyres used both for practice and race was seen on many rounds.

-Drivers got their best laps time toward the end of the race.Tyres kept their efficiency all along the race, except on some tracks.Tyre wear was linked more to the driver style, when  noticable, than from track shapes or chassis. Rindt, for instance, was very demanding on tyres on street courses, with his countersteering driving style.

in fact, racing tyres  where still in their infancy, compared with today standards.


not a critic, I ' ll like to know what Brr think can be a reasonable scaling for putting values on tyre wear

The tyre wear is actually tyre "condition", scaled to 90%: (condition - 90) * 10
The readme states that it's a bit simplistic at this stage (probably evident from the linear wear rates).

The tyres, as they stand in this mod, are definitely too soft to be comparable to the real life '67 cars.  But I'm not sure what the ultimate goal is - this is just a test mod that just happens to be based on the '67s, for simplicity's sake :)

View PostBob Simpson, on Sep 01 2010 - 05:25 PM, said:

Since keyboard driving means that you must use brake help, or essentially anti-lock brakes, you can't give them a real torture test.  I don't have time to test this mod myself, but it would be interesting to see what happens with the brakes engaged constantly or at least when the throttle is still applying power in corners as it is with some GPL alien drivers.

Yes, but to complicate things further, it's like no ABS system on earth, in that the applied brake power is seemingly progressively scaled according to available grip, so that the tyre never begins to lock up at all (unlike normal ABS, or the ideal case of some percentage slip for optimal traction.)
On the other hand, assuming there's no dodgy left-foot braking going on, the absence of braking help allows for marginally shorter braking times (owing to the higher average braking force) and hence more heat rejected to the air, rather than soaked into the disc.  Look at the difference between the surface spikes and the internal spikes.

We really need somebody to test this with an analogue brake, since the (numerical) balance cannot simply be inferred!

EDIT:

View PostJohn Woods, on Sep 01 2010 - 05:55 PM, said:

identiti_crisis,

Could you please offer a detailed comphrehensive methodology so others can do exactly what you are doing and generate comparable results?

There have always been (at least) two schools of thought with Grand Prix Legends enthusiasts: "traditionalists" who prefer replicating a particular season as exactly as possible, and "futurist realists" who want to pursue ultimate development of simulation experience. I think neither can survive without the other, together they have assured longevity of Grand Prix Legends, and both groups should learn to live with it or maybe start another thread, on which I will happily support both views.

I boot the game, choose a car, then the circuit, drive for as many laps as I can without breaking the car on the default (Papy) setup using the keyboard.  I'm slow as it is, but even more so on the 'board - about 1' 28" around Kyalami, today.  I have the odd excursion, usually minor since the car is uncontrollable on the grass and usually ends up in broken parts - there's also the odd half-spin from over-exuberant "flicking" due to the inherent inaccuracy of the keyboard - both of these are visible in the data I've posted, for the eagle-eyed.  I brake mostly in a straight line and early-ish, too, since the braking "help" won't let you brake into the corner, not that you could catch the oversteer easily anyway...

I don't know that copying my style exactly is what we need at this point.  Variety is probably more important; that everyone can drive according to their usual habits is most important. Let's try to "break" it :P
Of course, brr knows what's best for his productivity.


Here's an example of braking into the first corner; blue is internal temperature, green is surface - and there is the (almost imperceptible) expected lag.

Attached Files


Edited by identiti_crisis, Sep 01 2010 - 06:33 PM.


#20 John Woods

John Woods

    Be Somebody

  • GPLLinks Team
  • PipPipPipPipPipPipPipPipPipPip
  • 2,175 posts
  • Gender:Male
  • Interests:Too Much Fun
  • Sim interest:GPL

Posted Sep 01 2010 - 08:36 PM

identiti_crisis,

I meant, how do you go from data generation, as you explained above, (and I agree we all best have our own way of doing that), to presenting results with all these nifty graphs? Is there a utility that reads selected data and makes them? Is this part of the test mod? That would truly blow me away, but I'd want to use it with all of Grand Prix Legends OEM-version and mods.

#21 brr

brr

    a GPL editor

  • Members
  • PipPipPipPipPipPipPipPipPipPip
  • 1,244 posts
  • Gender:Male
  • Sim interest:GPL

Posted Sep 02 2010 - 12:00 AM

View PostJohn Woods, on Sep 01 2010 - 08:36 PM, said:

identiti_crisis,

I meant, how do you go from data generation, as you explained above, (and I agree we all best have our own way of doing that), to presenting results with all these nifty graphs? Is there a utility that reads selected data and makes them? Is this part of the test mod? That would truly blow me away, but I'd want to use it with all of Grand Prix Legends OEM-version and mods.

The test mod optionally generates the data into a log file. It is simply a few integer values per line, separated by spaces. Cut and paste into a separate file and then read it with a suitable program (spreadsheet software, scientific computing such as Matlab or R, etc...)

#22 brr

brr

    a GPL editor

  • Members
  • PipPipPipPipPipPipPipPipPipPip
  • 1,244 posts
  • Gender:Male
  • Sim interest:GPL

Posted Sep 02 2010 - 12:15 AM

View PostM Needforspeed, on Sep 01 2010 - 05:08 PM, said:

not a critic, I ' ll like to know what Brr think can be a reasonable scaling for putting values on tyre wear

I'm trying to get the tyre wear to a level where changing the tyres during a race might be a good idea. 15 laps at Kyalami is a bit over 20 minutes, which might be too short.

#23 brr

brr

    a GPL editor

  • Members
  • PipPipPipPipPipPipPipPipPipPip
  • 1,244 posts
  • Gender:Male
  • Sim interest:GPL

Posted Sep 02 2010 - 01:48 AM

Here's another test, Ferrari/Kyalami with default setup. I drove as long as it seemed reasonable to stay on track considering the tyre wear, and it turned out to be 18 laps (best lap 1:23.4, at the end laptimes were 2.5-3 seconds slower).

Some observations: brake fade can be noticed in the first couple of corners. Tyre wear comes first, noticeable at about 10 laps. Later heavy braking starts to feel a bit difficult, causing a couple of spins in T1. Laptimes get slower, and after 18 laps it seems driving is difficult enough that it would make sense to change tyres.

The brake fade model as such seems to be quite good. Tyre wear is currently based on total slip velocity (i.e. how fast the contact patch is moving relative to track surface) but in a linear way. I have not yet run into a proper explanation how tyres actually wear.

Testing physics at this stage does not need to be that extensive, since future changes to tyre wear and tyre grip probably affect other things so its an iterative process. There are also non-physics things to be tested: online, laptime security (can these laptimes somehow end up in GPLRank as 67 laptimes?) etc..


http://dl.dropbox.co..._def_18laps.jpg
http://dl.dropbox.co..._def_18laps.txt

#24 brr

brr

    a GPL editor

  • Members
  • PipPipPipPipPipPipPipPipPipPip
  • 1,244 posts
  • Gender:Male
  • Sim interest:GPL

Posted Sep 02 2010 - 08:48 AM

The tyre wear will be changed significantly in the next version, so no reason to test the current tyre wear. I found a shortcut to implementing it which should be fairly accurate (within limits of the tyre model of course). Its coming in the next major version. Here's a few laps around Kyalami, ending with tyre-destroying doughnuts just to check that tyre abuse really shows in the wear levels. The wear rate is too high, but fixing it is just a matter of tweaking one parameter.

Attached File  tyrewear.jpg   48.35K   50 downloads

Edited by brr, Sep 02 2010 - 08:49 AM.


#25 brr

brr

    a GPL editor

  • Members
  • PipPipPipPipPipPipPipPipPipPip
  • 1,244 posts
  • Gender:Male
  • Sim interest:GPL

Posted Sep 03 2010 - 03:06 AM

Does a tyre puncture actually affect driving in GPL? It happens so rarely that I don't remember anything else than noticing it as the listed damage after retiring the race. Tyre wear needs some sort of a graceful exit, otherwise the car becomes a hovercraft after too much tyre wear.

#26 John Woods

John Woods

    Be Somebody

  • GPLLinks Team
  • PipPipPipPipPipPipPipPipPipPip
  • 2,175 posts
  • Gender:Male
  • Interests:Too Much Fun
  • Sim interest:GPL

Posted Sep 03 2010 - 12:42 PM

Would not a race host admin perhaps prefer a configurable wear variable enabling rough specification of reasonable number of pit stops over given race? Or maybe a range of user choices, like hard, soft, inters. Posted Image

Edited by John Woods, Sep 03 2010 - 12:45 PM.


#27 brr

brr

    a GPL editor

  • Members
  • PipPipPipPipPipPipPipPipPipPip
  • 1,244 posts
  • Gender:Male
  • Sim interest:GPL

Posted Sep 03 2010 - 01:05 PM

View PostJohn Woods, on Sep 03 2010 - 12:42 PM, said:

Would not a race host admin perhaps prefer a configurable wear variable enabling rough specification of reasonable number of pit stops over given race? Or maybe a range of user choices, like hard, soft, inters. Posted Image

Different tyres are easy to do so that will probably be implemented. Anything that involves passing messages between server and client is probably too complicated, at least at the moment.

#28 John Woods

John Woods

    Be Somebody

  • GPLLinks Team
  • PipPipPipPipPipPipPipPipPipPip
  • 2,175 posts
  • Gender:Male
  • Interests:Too Much Fun
  • Sim interest:GPL

Posted Sep 03 2010 - 03:45 PM

Yeah, wasn't thinking again. Except that a fixed not infinite tire wear mode might be less appropriate at shorter or longer tracks. How to deal with Targa where a driver might wear out tires half way round? And, some racing organizations have mandatory green flag pits stops, (even if unnecessary otherwise). So working with both of those possible circumstances prompted curiousity. For instance, a mandatory pit stop rule could offer an advantage of refreshing tires or choosing not to, depending on user selection of "compound" and race strategy. But as I think may have been noted previously, advantages of more grip with fresh tires may not necessarily recover time lost from stopping?

Edited by John Woods, Sep 03 2010 - 05:16 PM.


#29 Burnsy865

Burnsy865

    Denny Hulme

  • Supporter
  • PipPipPipPipPipPipPipPipPipPip
  • 1,835 posts
  • Gender:Male
  • Location:Brisbane, Australia
  • Sim interest:GPL and P&G

Posted Sep 03 2010 - 04:56 PM

View PostJohn Woods, on Sep 03 2010 - 03:45 PM, said:

How to deal with Targa where a driver might wear out tires half way round?

I often think while lapping at Targa how good it would be if we could call into one of the various service areas and fix things, I doubt this could be implemented and it is more of a track thing than a mod thing, just a thought, maybe a silly one but you never know with this community that never ceases to amaze.

#30 John Woods

John Woods

    Be Somebody

  • GPLLinks Team
  • PipPipPipPipPipPipPipPipPipPip
  • 2,175 posts
  • Gender:Male
  • Interests:Too Much Fun
  • Sim interest:GPL

Posted Sep 03 2010 - 05:19 PM

Maybe tire wear could be based on fixed by marque coefficient of selected race length/compound selected/track length? Or something like that...then only required to factor car, what track its on, and race length to set tire wear variable.

Edited by John Woods, Sep 03 2010 - 05:27 PM.





0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users

Sim Racing Links