• Welcome to PiBoSo Official Forum. Please login or sign up.
 
July 09, 2020, 09:29:48 PM

News:

World Racing Series beta14 available! :)


GP Bikes beta18c

Started by PiBoSo, June 17, 2020, 08:46:09 PM

Previous topic - Next topic

Vini


h106frp

Its just that the clutch in MaxHud is the controller/lever position so its influence would be controller/autoclutch, the bar in debug is the clutch pressure plate position so it shows the slipper activation as well as the rider input.

Vini

Just going by the feel I get from riding the bikes, I have to say that at max. lean the slipper clutch is disengaging too easily in general. While the engine brake feels fine at upright braking and quickly makes the rear step out, it is not pulling the bike inwards enough while cornering.
Clips like those above are extreme examples of this behaviour, where the clutch seems to suddenly let go completely. To be fair, this only happens at a few select corners.

h106frp

June 28, 2020, 12:47:09 AM #18 Last Edit: June 28, 2020, 12:51:54 AM by h106frp
I have been trying to update some bikes and testing (or at least trying to) an updated Nordy and to be honest I am finding the current bike physics a total disaster - the 'crabbing' effect is very severe, camber problems are severe and the VR appears to be unable to control a turn with the ammount of lean to negotiate a turn excessive.

I have also noticed the back stepping out in a very regular manner on long turns, the clutch appears to be engaged OK but the bike ratchets around the turn and the longer you have to hold a turn the more lean the VR feeds in until the front folds, tried pretty much everthing (VR, tyres, geom) without a solution or even a clue as to what is occuring.

Pretty much at the point of giving up trying until the various problems highlighted in recent posts are properly fixed as its just not much fun having to ride around the problems.

In the garage you can alter the ramp and free travel for the slipper and it does appear to funtion but I can only guess that you need to set the clutch max torque carefully in .cfg to emulate the spring rate/slip

Vini

I have to say that most of these negative developments only came with 18c.

Stout Johnson

June 28, 2020, 06:13:30 AM #20 Last Edit: June 28, 2020, 06:29:10 AM by Stout Johnson
Quote from: Vini on June 27, 2020, 10:45:20 PMThere seems to be something very wrong with the slipper clutch. I'm not 100% sure but I think this is only a problem since 18c.

@PiBoSo: During long off-throttle phases in high gears and at max. lean, the clutch sometimes seems to disengage completely, even when giving it throttle.
Mimo recorded this clip demonstrating the issue: https://streamable.com/fmlgfr
I have had this issue before beta18c. Just to make sure we are not following wrong leads here.

Quote from: Vini on June 28, 2020, 12:11:18 AMWhile the engine brake feels fine at upright braking and quickly makes the rear step out, it is not pulling the bike inwards enough while cornering.
I totally agree with your second observation - the bike not being pulled inwards enough. Concerning the engine brake while upright, I have to disagree. You can get the rear to lock heavily just by braking moderately hard and no downshifts. This is totally exaggerated. My observation only relate to the Schwaben M2 bike, I have not tried any mod bikes. Maybe our differing observations are bike dependant. We agree on the general problems though.

Quote from: h106frp on June 28, 2020, 12:47:09 AM[...]VR appears to be unable to control a turn with the ammount of lean to negotiate a turn excessive.

[...]the longer you have to hold a turn the more lean the VR feeds in until the front folds
I agree, I have made similar observations. In some corners, you end up with significantly higher lean angles than it would be realistic for that kind of corner. I have the feeling that whichever bug is causing the camber corner problems is also causing this strange lean behaviour. But I am not totally sold on the VR being the cause, at least not the only cause. Sometimes I think that the suspensions simulation might be involved in that?

@Piboso: GPB has so much potential, there are so many things that are beautiful in GPB. If you can tune the grip/tire temps relation, tune the engine braking and slipper clutch simulation then it would probably be a good start. If you then would be able to detect and correct what is happening in camber corners, look into the tire issues in wet then GPB might be there physics wise. I politely ask you to please make these physics issues your number one priorities. They are on the table and many of us here are willing to help and test. But please concentrate on physics - this is where GPB used to have its strength compared to other titles. 

The core physics need to feel believable and realistic. That's what brought me to GPB in the first place. I want to be able to a realistic feel for what the bike and the tires are doing. I want to be able to ride on a wet and drying track and try hit the drying spots on the racing line and get a feel for how much I can push on the brakes and on the throttle. I want to experience the slides and skids. That's what GPB is all about for me and what used to be fun.  But at the moment it does not feel believable anymore.
    -----------   WarStout Kawasaki Team   -----------

h106frp

Well, had a look back and the weird physics started when the off track handling went all auto-pilot and that was around b13, PB stated this is due to tyre modelling and required an urgent fix.... 5 betas on and waiting  :(

Vini

June 28, 2020, 01:22:46 PM #22 Last Edit: June 28, 2020, 01:46:43 PM by Vini
Quote from: h106frp on June 28, 2020, 12:14:49 PMWell, had a look back and the weird physics started when the off track handling went all auto-pilot and that was around b13, PB stated this is due to tyre modelling and required an urgent fix.... 5 betas on and waiting  :(
What "weird physics" are you talking about? IMO, beta18b was the best version so far, given the modders got the bike tuning right.

Quote from: Stout Johnson on June 28, 2020, 06:13:30 AMConcerning the engine brake while upright, I have to disagree. You can get the rear to lock heavily just by braking moderately hard and no downshifts. This is totally exaggerated. My observation only relate to the Schwaben M2 bike, I have not tried any mod bikes. Maybe our differing observations are bike dependant. We agree on the general problems though.
Yes, it is heavily bike dependent. I don't think many modders have properly analyzed the new slipper clutch behaviour and adjusted the bike accordingly. On the MotoGP bikes, the engine brake feels much better and predictable than in previous betas and full lock-ups never happen.
You can even lift the rear while sliding sideways Marquez-style: https://streamable.com/z2m9ts


I think in general, beta18(b) was a great step forward, it just needs some additional slipper clutch (engine brake) tuning for extreme situations.
18c went in the wrong direction, though. Only thing it improved was giving a slightly better feel for the rear under acceleration.
Would be good to know if the VR issues of 18c and the change in rear feel are directly connected but I somehow doubt that.

h106frp

The loss of player control off-track (almost impossible to crash until you re-join tarmac) was marked as the point where the tyre were 'broken'. Maybe fixing the tyres will improve other aspects.

We have had several versions of bike inertia just for b18 and the 18a and 18b were certainly incorrect so most bikes wil need modifying moving forward.

Personally I think the slipper is working but the engine back torques (.engn) now need adjusting to suit and its not clear how maxtorque comes into the mix.

Certainly pre 18c you could get the bikes to work, now its almost impossible with the rear feeling like its being lifted(levered) off the track during turns.

Vini

Example of the virtual rider just randomly letting go mid-corner: https://streamable.com/irxjpe
No camber change, no trailbraking, no bump and no extreme lean angle.

Myst1cPrun3

Quote from: Vini on June 28, 2020, 11:28:54 PMExample of the virtual rider just randomly letting go mid-corner: https://streamable.com/irxjpe
No camber change, no trailbraking, no bump and no extreme lean angle.

Camber change = corner goes from flat to up hill so yes there is

Extreme Lean angle = Your brake lever is literally scraping the floor, and the wing through it...

You bottomed the bike on the floor

Vini

That's not what camber means but either way that corner is pretty much flat. Maybe 1° or something.

The bike only bottomed out after crashing. The perspective is deceiving you. The elbow was just about touching the ground and not a single part of the bike was scraping. Screenshot one frame before the front started folding: https://i.imgur.com/s3lutsO.jpg

But if you want to start some nonsense debate with me again, let's continue in a separate thread and keep this one on-topic, ok?

h106frp

Pretty much the sort of behaviour I am observing, tried different bikes with different suspensions, collisions, VR etc but the crashes are very predictable for any given corner.

PiBoSo

June 29, 2020, 08:18:48 AM #28 Last Edit: June 29, 2020, 09:57:44 AM by PiBoSo
Quote from: Vini on June 28, 2020, 12:57:24 AMI have to say that most of these negative developments only came with 18c.

Please note that nothing has been changed in the Beta18c bike physics, except for stability at very low speeds.
"Obviously your ambition outweighs your talent".

Vini

June 29, 2020, 12:03:16 PM #29 Last Edit: June 29, 2020, 12:54:23 PM by Vini
How exactly did you change the stability at very low speeds? I'm guessing by changing the virtual rider algorithm?

There is definitely something wrong with the front (the VR) since 18c and this has been observed by a few people now. You can see it in the example vids I posted. Those things did not happen in 18b.