Quote from: Myst1cPrun3 link=msg=106032After some testing, I believe I have come up with a theory and a possible solution to the seemingly random front end crashes, as well as the issues with camber changes mid corner.
(Thinking MG hairpin at Victoria)
I did touch on this in a previous topic before it was moved, as it got way off the original conversation point. (As much my fault as any that was so I apologize)
Now:
Until I started fiddling with tyres, I thought the front end crashes on camber changes was because of the suspension, and a lacking part in the simulation or mod..
However, in the tyre files, max lean can be set using the DX1 and DY1 values.
Now this differs per front and rear tyre so you can have a tyre have a larger max lean angle than the other.
The bike will only lean to the angle of whichever tyre has the smallest max angle.
So for instance if you let the front have a higher max lean angle than the rear, it is the rear that'll determine the total bike lean.
This becomes important in camber changes.
It's no secret the front tyre is narrower than the rear, so in terms of lean angle, it has a slightly different steering rate it seems. (Not sure on Reasons or physics behind this just how it seems)
If you are on the max lean angle of a tyre, and then the camber changes, the bike does not get stood up, the lean angle changes to beyond the maximum set value in the tyre file, and a crash occurs with seemingly no warning. Typically these crashes are front end crashes as that's the narrower tyre and it's also what it's the camber change first.
As a result I found giving the front tyre 0.1 extra lean on the DX1 and DY1 values when compared to the rear tyre helped a lot.
This extra angle doesn't affect the bikes handling on the flat, as the rear tyre will be the limiting factor, but when the camber changes, it means the front doesn't simply 'fall off the edge' like is known to happen in GP Bikes.
This isn't to say you can't crash, you can, it just adds a bit of control of the bike over camber changes.
Here is an example of the M2v2 at the MG hairpin, where I was being quite aggressive on approach, and could make the apex confidently, despite the known issues.
It is a marked improvement, but I am going to do more tuning when the 'big tyre update' drops for the v2
You are leaving many values in the inkwell. Trees won't let you see the forest
I don't understand what that means.
Gonna guess lost in translation
I Also don't understand the point of this thread?
@Mystic: I think Manu is trying to tell you that when tuning the tyres, you can't just alter a couple of values without those alterations affecting other factors in the tyre file that will need adjusting too to compensate? Correct me if I'm wrong Manu. ;)
Quote from: Hawk on June 17, 2020, 12:58:51 PM@Mystic: I think Manu is trying to tell you that when tuning the tyres, you can't just alter a couple of values without those alterations affecting other factors in the tyre file that will need adjusting too to compensate? Correct me if I'm wrong Manu. ;)
Ah.
Well that's a very roundabout way of saying that ;D
From my testing of fiddling with these values there actually wasn't anything else that needed adjusting, the tyres seemed to function as intended.
Was just a way of making behaviour over crests/through dips while on lean more controllable and predictable.
Not a debate or anything, or even a definite solution. Just a mere observation that allowing the front tyre to lean more using that value 'helped' the front-end washes over uneven surfaces/camber changes.
I fail to see why people are fixating on it :o
Dont want to become overly involved in this but its worth loading MaxTyre to visualise whats happening. A tyre has no such thing as a max lean angle, its just the point where the slip becomes greater than the friction holding the tyre against the tarmac (until you roll of the edge) and is mainly down to how well the contact patch deforms around the profile and circumference (no idea if this is modelled), curiously you have maximum contact patch when leaned at around 45-50 degrees with modern designs.
DD is correct, a narrow tyre of given torus will arc around a turn faster (smaller radius) than the wide tyre as it rolls around its profile faster for a given lean, narrow tyres need less lean than wider ones for a given turning arc, hence the front washing out first and the perceived nimble handling of small bikes - you only run a tyre thats just wide enough for the power and weight of the bike - oversize tyre are a silly styling thing and ruin handling. Unfortunately narrow tyres do not appear to model well at the moment and the performance appears to drop on a cliff edge below a certain width.
Quote from: Hawk on June 17, 2020, 12:58:51 PM@Mystic: I think Manu is trying to tell you that when tuning the tyres, you can't just alter a couple of values without those alterations affecting other factors in the tyre file that will need adjusting too to compensate? Correct me if I'm wrong Manu. ;)
Quite so! Sorry for my hieroglyphs ;RE
When I can I will try to give some details.
Quote from: h106frp on June 17, 2020, 01:22:30 PMDont want to become overly involved in this but its worth loading MaxTyre to visualise whats happening. A tyre has no such thing as a max lean angle, its just the point where the slip becomes greater than the friction holding the tyre against the tarmac (until you roll of the edge) and is mainly down to how well the contact patch deforms around the profile and circumference (no idea if this is modelled), curiously you have maximum contact patch when leaned at around 45-50 degrees with modern designs.
DD is correct, a narrow tyre of given torus will arc around a turn faster (smaller radius) than the wide tyre as it rolls around its profile faster for a given lean, narrow tyres need less lean than wider ones for a given turning arc, hence the front washing out first and the perceived nimble handling of small bikes - you only run a tyre thats just wide enough for the power and weight of the bike - oversize tyre are a silly styling thing and ruin handling. Unfortunately narrow tyres do not appear to model well at the moment and the performance appears to drop on a cliff edge below a certain width.
While my point was itself in the tyre files, and it may not be 100% max lean directly, most motorcycle games have a 'wall' that prevents the rider from leaning beyond a certain angle
My point is editing this value seems to limit how far over the bike leans. Regardless of the rest of the tyre values. It seems to ovverule them.
On my m2v2 for instance I. Use it to limit the lean, feel free to edit those values in the tyres there and see the difference in handling. (Most Noticeable in the wet)
But if I removed it the bike could lean further without crashing due to the grip of the tyres.
Now while it may not be known in the formulas as Max lean, that does seem to be the effect it has in game which is what I'm trying to convey
I can only suggest trying Extreme alterations of the values suggested to understand what's meant.
In any case, allowing the front to lean further than the rear allows for a more stable bike over elevation/camber changes mid turn.
As h106frp tried to explain, there's no such a thing as "max lean angle" in GPB. Even on a real bike, it depends on plenty of factors: MotGP bikes can maybe lean 60deg, but in some cases you can lose the fron when leaning 45deg or even less ...
In GPB, on top of these, you also have the virtual rider in the loop, so drawing conclusions is very hard.
As h106frp suggested, I'd advise to have a look at how GPB tyre model works, what DX1 and DY1 are and what their effect is (in a nutshell, they define the max longitudinal and lateral "grip" a tyre has at optimal longitudianl slip / sideslip angle).
Long story short, increasing them you've just given your tyre more "grip" so yeah, it handles better. No big surprise.
There's little chance that anything good can be obtained reasoning on things like "front and rear tyre have different max lean angles".
P.S.
If I recall correctly, PiBoSo long ago stated he's aware of the issue of front losses on turns with camber: if the true solution was simply to give the tyre a bit of extra grip, that would probably be already done. It must be way more complex than that.
The original 990 had a wide rear tyre compared to the front but handles well compared to many. Their have been many mods of it but maybe looking at the 990 might help
DD
Quote from: HornetMaX on June 18, 2020, 02:42:26 PMAs h106frp tried to explain, there's no such a thing as "max lean angle" in GPB. Even on a real bike, it depends on plenty of factors: MotGP bikes can maybe lean 60deg, but in some cases you can lose the fron when leaning 45deg or even less ...
In GPB, on top of these, you also have the virtual rider in the loop, so drawing conclusions is very hard.
As h106frp suggested, I'd advise to have a look at how GPB tyre model works, what DX1 and DY1 are and what their effect is (in a nutshell, they define the max longitudinal and lateral "grip" a tyre has at optimal longitudianl slip / sideslip angle).
Long story short, increasing them you've just given your tyre more "grip" so yeah, it handles better. No big surprise.
There's little chance that anything good can be obtained reasoning on things like "front and rear tyre have different max lean angles".
P.S.
If I recall correctly, PiBoSo long ago stated he's aware of the issue of front losses on turns with camber: if the true solution was simply to give the tyre a bit of extra grip, that would probably be already done. It must be way more complex than that.
I'm not stating that it is a 'concrete' Max Lean value... I'm stating that the effect on the bike, which is perhaps where my explanation falls short. (I'd be a crap development rider ;D )
Quite what else happens in the background, I have no clue, nor do I know the actual definition of the value.
All I do know, is that having a higher value on the front, (indirectly it appears) for whatever reason, makes the front end behave better over camber changes, as it can use more of the circumference of the tyre without leaning off the edge of it.
Quite how this works, again, I cannot say, those are just my findings. Whether its a permanent solution, or a bandaid fix, I cannot say either.
Just thought it would be helpful for some people to use should they have issues with camber changes :P
I have been going through the usual re-tuning of a bike using Victoria and I have developed a bit of a new theory on what occurs at the bottom of Lukey Height/MG. After noticing the problem only ever occurs when entering the turn with high lean (you can always blast through it any old how when you pick it up after a drop) I wondered about chassis flex and damping, theory being that as the bike has a high lean angle we lose damping and travel from the suspension units and become dependant on the mechanical flex and damping - much in line witrh Vinis comments about the bikes being overly 'stiff' but maybe with a slightly different perspective.
With the newer inertia calcs it appears we have a bit more range to play with in the damping and chassis stiffness ranges and it does improve the situation very noticeably if you can add a bit of flex and damping to the chassis. It might be worth you trying and comparing it to what you obtain from tyre adjustments.
On a side note, you guys really need to stop testing bikes at Victoria. The simple, sweeping layout masks almost all handling issues of bikes.
Similarily, aids (including TC/AW) should be turned off when testing physics and high direct lean values should be used, as this too avoids "fake stability".
Quote from: h106frp on June 18, 2020, 10:16:08 PMI have been going through the usual re-tuning of a bike using Victoria and I have developed a bit of a new theory on what occurs at the bottom of Lukey Height/MG. After noticing the problem only ever occurs when entering the turn with high lean (you can always blast through it any old how when you pick it up after a drop) I wondered about chassis flex and damping, theory being that as the bike has a high lean angle we lose damping and travel from the suspension units and become dependant on the mechanical flex and damping - much in line witrh Vinis comments about the bikes being overly 'stiff' but maybe with a slightly different perspective.
With the newer inertia calcs it appears we have a bit more range to play with in the damping and chassis stiffness ranges and it does improve the situation very noticeably if you can add a bit of flex and damping to the chassis. It might be worth you trying and comparing it to what you obtain from tyre adjustments.
I'm a bit surprised by that. Not saying it's not true of course, but I'm not convinced ... I mean, I'm not sure it should happen even with a totally rigid chassis (as it was in older GPB betas).
Side note: I think Victoria is a good test track. It has a bit of everything. Not sure why Vini crusades against it :)
After a bit of reading that is apparently the perceived wisdom on chassis flex and why super stiff chassis is a bad idea for bikes, as you lean the bike over the 'bumps' still occur vertically but the suspension can only translate the vector of travel so the displacement is less than it would be for upright, as the displacement in the damper rod movement is less the damping contributed is equally less effective so we become dependant on the lost motion in the chassis construction to contribute some of the damping.
Obviously anybodys guess if this is how its applied in GPB!
But an evening testing and it does appear to improve matters with high lean/camber - maybe @PiBoSo could explain how chassis flex is used in the physical model?
Its not 'something for nothing' though higher chassis flex appears to make it much more difficult to get the VR right with the rider struggling to lift the bike back up after the turn :(
Chassis flex has always influenced bike handling in GPB very positively. Sadly, a lot of mods are way too stiff.
Quote from: HornetMaX on June 19, 2020, 11:56:49 AMSide note: I think Victoria is a good test track. It has a bit of everything. Not sure why Vini crusades against it :)
Because I have ridden plenty of bikes there that seemed to be good but then revealed themselves to be absolute shite on more technical tracks. Victoria lacks slow corners, bumps, quick direction changes and hard braking. You only have all those long corners where you basically have enough time to get any bike to turn nicely.
The track is simply not bike dependent at all, that's partially why real MotoGP races are always so exciting there.
Quote from: Vini on June 19, 2020, 01:47:30 PMChassis flex has always influenced bike handling in GPB very positively. Sadly, a lot of mods are way too stiff.
Quote from: HornetMaX on June 19, 2020, 11:56:49 AMSide note: I think Victoria is a good test track. It has a bit of everything. Not sure why Vini crusades against it :)
Because I have ridden plenty of bikes there that seemed to be good but then revealed themselves to be absolute shite on more technical tracks. Victoria lacks slow corners, bumps, quick direction changes and hard braking. You only have all those long corners where you basically have enough time to get any bike to turn nicely.
The track is simply not bike dependent at all, that's partially why real MotoGP races are always so exciting there.
+1 Vini
Totally agree with your deductions about testing and track testing for bikeMOD devs, at least to get the bike handling well initially. ;)
Quote from: h106frp on June 18, 2020, 10:16:08 PMI have been going through the usual re-tuning of a bike using Victoria and I have developed a bit of a new theory on what occurs at the bottom of Lukey Height/MG. After noticing the problem only ever occurs when entering the turn with high lean (you can always blast through it any old how when you pick it up after a drop) I wondered about chassis flex and damping, theory being that as the bike has a high lean angle we lose damping and travel from the suspension units and become dependant on the mechanical flex and damping - much in line witrh Vinis comments about the bikes being overly 'stiff' but maybe with a slightly different perspective.
With the newer inertia calcs it appears we have a bit more range to play with in the damping and chassis stiffness ranges and it does improve the situation very noticeably if you can add a bit of flex and damping to the chassis. It might be worth you trying and comparing it to what you obtain from tyre adjustments.
If you remember the problems bikes used to have on Silverstone entering that right hand rising bend as you exit from under the bridge, the front end always used to wash-out? Well that issue was fixed with correct chassis-flex settings if you remember mate? ;)
Quote from: h106frp on June 19, 2020, 01:22:51 PMAfter a bit of reading that is apparently the perceived wisdom on chassis flex and why super stiff chassis is a bad idea for bikes, as you lean the bike over the 'bumps' still occur vertically but the suspension can only translate the vector of travel so the displacement is less than it would be for upright, as the displacement in the damper rod movement is less the damping contributed is equally less effective so we become dependant on the lost motion in the chassis construction to contribute some of the damping.
This is absoluely true, even if it is probably worth to remember that the travel you get from a non-rigid chassis is way smaller than what you get from suspensions (so yeah, it helps a bit, but it's not something that can absorbe bumps). The thing I'm not convinced about is that it helps/solves the problem we have in GPB on turns with camber and slope changes. Or at, best, why it does it.
Quote from: Vini on June 19, 2020, 01:47:30 PMQuote from: HornetMaX on June 19, 2020, 11:56:49 AMSide note: I think Victoria is a good test track. It has a bit of everything. Not sure why Vini crusades against it :)
Because I have ridden plenty of bikes there that seemed to be good but then revealed themselves to be absolute shite on more technical tracks. Victoria lacks slow corners, bumps, quick direction changes and hard braking. You only have all those long corners where you basically have enough time to get any bike to turn nicely.
The track is simply not bike dependent at all, that's partially why real MotoGP races are always so exciting there.
Well, Victoria has:
- A fast straight
- Plenty of fast turns, including a high-speed direction change
- Sort of a hard braking
- Some not-so-fast turns
- Some bit of slope changes
And, of course, it's the track GPB comes with.
So yeah, suggesting modders to stop testing their mods on Victoria is not a good idea in my book.
If you're suggesting for modders to
also test on other tracks, then yes, that could be OK.
You should probably name the track you thinks are relevant.
Quote from: HornetMaX on June 20, 2020, 01:15:17 PMSo yeah, suggesting modders to stop testing their mods on Victoria is not a good idea in my book.
If you're suggesting for modders to also test on other tracks, then yes, that could be OK.
You should probably name the track you thinks are relevant.
Sure, they can
also keep testing at Victoria, it's just that those results are pretty unimportant compared to more technical tracks.
Almost any track I can think of is better for testing than Victoria.
Some examples: Valencia, COTA, Mugello, Sepang, Aragon
If we're going for test tracks, then the real world test tracks are probably gonna be the best choices.
After all they test there for a reason... ;)
Catalunya/Montmelo has pretty much everything, and Jerez is good to. Sepang has a wide variation, and cheste/Valencia is good for low speed handling and testing traction.
Victoria and Assen 2005 should be the best tracks to use, as they are (in theory) the most optimised for the sim being official content.
With the latest updates is anyone else noticed the return of the 'old' problem of the rider 'just giving up' and dropping the bike in turns?
Yeah, while the rear feels nicer the front has lost stability.
@Mystic: I agree. There is a reason PI is not a real MotoGP test track...
Quote from: h106frp on June 18, 2020, 10:16:08 PM[...] After a bit of reading that is apparently the perceived wisdom on chassis flex and why super stiff chassis is a bad idea for bikes, as you lean the bike over the 'bumps' still occur vertically but the suspension can only translate the vector of travel so the displacement is less than it would be for upright, as the displacement in the damper rod movement is less the damping contributed is equally less effective so we become dependant on the lost motion in the chassis construction to contribute some of the damping.
Obviously anybodys guess if this is how its applied in GPB!
I totally agree. Chassis flex is very important, especially during high lean angles, as the suspension is almost non-effective anymore.
https://motodna.net/flex-or-not-flex/
https://www.cycleworld.com/2015/10/16/motogp-racing-chassis-flex-and-stability-are-key-to-winning-races/
If the suspension is acting as irl, then chassis flex should be similarly important as it is in real life. Piboso could probably bring light into this matter.
Another important factor is the tires acting as a sort of suspension.
So essentially, you have front and rear suspension, chassis and also front and rear tire acting as suspension-like entities and therefore possibly affecting grip level. Not sure how that is reflected in GPB?
Quote from: Vini on June 19, 2020, 01:04:04 AMOn a side note, you guys really need to stop testing bikes at Victoria. The simple, sweeping layout masks almost all handling issues of bikes.
Similarily, aids (including TC/AW) should be turned off when testing physics and high direct lean values should be used, as this too avoids "fake stability".
Totally agree with all here, escpecially the part on direct lean.
Quote from: Vini on June 19, 2020, 01:04:04 AMThe track is simply not bike dependent at all, that's partially why real MotoGP races are always so exciting there.
On Victoria the only turns worth testing are lucky heights (to some degree) and especially the right tigh turn afterwards (turn 10). Some info can also be obtained by accelerating through "Stoner Corner" (turn 3). All the other parts are really not giving much information.