physx Warrior FV510 PhysX issues

[J. Point 44]

Hey,

 

I'm working on the config of a model I created but I'm having trouble with the config of the Physx, more precisely the speed.

 

I am aware of the normal reason for this, being that the memory points for the wheels aren't all the same diameter but that isn't the case, the warrior was stuck at about 12-13km/h when this was the case with the sprockets at a different size but after fixing this the new max speed is a whopping 8 km/h!

 

picture of the memory LOD for the axis & bounds: 

 

 

physX.hpp link: https://drive.google.com/open?id=1_clmWh1yXm2z_BAag3cgdt9eqgLZv-DU

 

model.cfg  link: https://drive.google.com/open?id=1YTLLd0_N6zpp9vNuSeWGRRM9dN2k8ABo

 

The only thing I can think of might be the mass?

 

The mass for the geometry is 25400.0 and the mass for the Geometry PhysX is 6.45.

Pictures of both:

 

 

 

Any help is very very appreciated! 

[x3kj 1247]

First of all, the mass of physX LOD has no effect at all.

To find out what causes this issue you should use arma3diag.exe (only available when using devbranch). When using diag.exe, activate EPEVehicle dialog for debugging help https://community.bistudio.com/wiki/Arma_3_Diagnostics_Exe

 

Based on experience there could be two core reasons:  a) the gearbox is simply not switching properly and is stuck in first gear  b) some weird hiccup with the wheels/physx bodies, where parts are counted as draggin on the ground and therefore creating a lot of resistance forces

 

Best way for us to help you would be if you could make a video of your tank (attempt to) drive, with EPEVehicle dialog enabled.

[J. Point 44]

 

13 hours ago, x3kj said:

First of all, the mass of physX LOD has no effect at all.

To find out what causes this issue you should use arma3diag.exe (only available when using devbranch). When using diag.exe, activate EPEVehicle dialog for debugging help https://community.bistudio.com/wiki/Arma_3_Diagnostics_Exe

 

Based on experience there could be two core reasons:  a) the gearbox is simply not switching properly and is stuck in first gear  b) some weird hiccup with the wheels/physx bodies, where parts are counted as draggin on the ground and therefore creating a lot of resistance forces

 

Best way for us to help you would be if you could make a video of your tank (attempt to) drive, with EPEVehicle dialog enabled.

Thanks for the reply.

I'll try and get the diagnostics run tomorrow and update here.

 

Am I ok to change the phsyx lod mass to 1 then? I don't know if it will help you to find the problem but some of it behaviour while driving:

 

• When I stop accelerating or as soon as the vehicle speed drops itself to 10km/h or below (when it was still able to go that speed) the vehicle hits the brakes, hard and stops. Same on reverse. Changing the brakeIdleSpeed has no effect on this behavior. So right now with 8km/h being the max as soon as I stop accelerating it stops. The gearbox switching might be the cause but I've tried it on full-auto and still no luck.
• When I tweak the dampingRateInAir and MOI the vehicle goes totally out of control, accelerating on itself as soon as it gets above 10km/h (again, before the axis and bounds being corrected) and would go to about 90km/h and then start flipping like crazy.

[J. Point 44]

 

@x3kj

[x3kj 1247]

10 hours ago, J. Point said:

• When I stop accelerating or as soon as the vehicle speed drops itself to 10km/h or below (when it was still able to go that speed) the vehicle hits the brakes, hard and stops. Same on reverse. Changing the brakeIdleSpeed has no effect on this behavior. So right now with 8km/h being the max as soon as I stop accelerating it stops. The gearbox switching might be the cause but I've tried it on full-auto and still no luck.
• When I tweak the dampingRateInAir and MOI the vehicle goes totally out of control, accelerating on itself as soon as it gets above 10km/h (again, before the axis and bounds being corrected) and would go to about 90km/h and then start flipping like crazy.

1) there is only one functional mode for the gearbox in A3 currently and that is automatic mode "auto"  (i guess full-auto does simply the same).

2) dampingrateinair and MOI are the main parameters that dictate how fast/slow your vehicles accelerates, decelerates without engine power and it's topspeeds.  Tweaking them carefully, one at a time is necessary. That you stop as soon as you release throttle indicates that you have fairly high movement resistant currently. If the vehicle accelerates on it's own then that is a result of dampingrateinAir beeing too low for the chosen MOI value. It's unphysical behaviour (energy conservation law thrown out the window) due to BIS's tank physx integration as far as i can tell, but thats what we have to deal with.

3) in the dialog you see that at current top speed you have an rpm of ~1860 at the engine. If you check your engine settings, you see that at this point the engine is losing torque, would it turn faster. Your vehicle not accelerating further means that the MOI/dampingrateinair setting at this speed is in balance with the torque generated by the engine. JFYI - you also lose torque and rpm due to the physx clutch - if you check the rpm value behind Gear 2, it says 1589 rpm as Gearbox input - so you are losing 270 rpm due to the clutch at this mode of operation.

 

For playing with dampingrate and MOI it is efficient to use diag.exe's config merge function. You can patch the config while the game is running that way. You only need to restart the mission (or respawn the vehicle) to get the change, instead of restarting the game.

 

[J. Point 44]

46 minutes ago, x3kj said:

2) dampingrateinair and MOI are the main parameters that dictate how fast/slow your vehicles accelerates, decelerates without engine power and it's topspeeds.  Tweaking them carefully, one at a time is necessary. That you stop as soon as you release throttle indicates that you have fairly high movement resistant currently. If the vehicle accelerates on it's own then that is a result of dampingrateinAir beeing too low for the chosen MOI value. It's unphysical behaviour (energy conservation law thrown out the window) due to BIS's tank physx integration as far as i can tell, but thats what we have to deal with.

3) in the dialog you see that at current top speed you have an rpm of ~1860 at the engine. If you check your engine settings, you see that at this point the engine is losing torque, would it turn faster. Your vehicle not accelerating further means that the MOI/dampingrateinair setting at this speed is in balance with the torque generated by the engine. JFYI - you also lose torque and rpm due to the physx clutch - if you check the rpm value behind Gear 2, it says 1589 rpm as Gearbox input - so you are losing 270 rpm due to the clutch at this mode of operation.

 

 

So if I understand correctly you are saying there are no issues with friction, wheel size, physx geometry hitting, or stuff like that? Why is my wheel rotation speed on the right higher than on the left? they are exact copies in the p3d model.

 

The current settings for the enginePower, maxOmega and peakTorque are taken from the Real life engine that is in the Warrior and I tweaked to torqueCurve to something that makes sense to me. 

 

 

Are the above settings the cause or part of the problem? Or is the only think I need to tweak the MOI & dampingrateinair?

 

 

Finally; I read about the diag.exe's config merge function in the link you sent before but I don't understand how I can get that to work? These settings aren't in a single config file, this is the current PBO structure:

 

 

Thanks.

[x3kj 1247]

1 hour ago, J. Point said:

Why is my wheel rotation speed on the right higher than on the left? they are exact copies in the p3d model.

At slow speeds physx is inaccurate and friction varies, even for identical sides.

https://feedback.bistudio.com/T77522

Try finetuning MOI and dampingrateinair first and see what happens. That the tank accelerates on its own in previous trials is a good indicator that there is no issue with stuck wheels etc.

When using RL values, remember that there are many losses on a vehicle, so even if the tank has 1500HP on paper, a big fraction can be taken up by engine cooling and all sorts of auxiliary pumps and things. That said, i can't recommend obsessing over "real values", because there are a few unphysical things (especially how MOI and dampingrateinair behave) that just void all the accuracy put into other parts.