4WD Centre LSDs Refuse to Work When the Gearbox is in Neutral
1. Use viscous centre LSD on a 4WD.
2. Flip the car over.
3. Throttle up to "top speed" and shift to neutral.
4. Use the handbrake to lock rear wheels. Front wheels won't slow down significantly. The centre LSD is not working as it should !
5. Keep the clutch full disengaged and shift into any gear. Front wheels will slow down. The LSD is working now !
6. Back to neutral. The LSD will refuse to work again.
All cars are known to have some issues when they are upside down, but it shouldn't matter too much cause you can't drive them anyway when they are upside down. Does it behave like this when the car is not upside down?
:rolleyes: This has nothing to do with the car being flipped, obviously.
Seems like the LSD code isn't active when the car is in neutral. Of course not technically correct, it's probably part of an optimization or a simple oversight. In any way, this has hardly any impact on actual racing or any form of driving for that matter, so no one should lose any sleep over that bug not getting fixed.
Don't be such a smartass android, you're not familiar with LFS's code. Obviously it shouldn't have anything to do with the car being flipped, but how would you know? I've seen quite a few physics issues posted here which were caused simply by the fact the car was upside down.
im pretty sure the h geared cars are in neutral for a short moment while shifting which will cause much the same problems as the non preloaded diffs did for mouse users when trying to modulate the throttle by tapping
Nothing to do with car being upside down, as shown in the picture ( reversing, rear wheels locked, no foot brake ).
Yes.
And when you pull the handbrake to enter tight corners while using H-stick to shift the car may suddenly oversteer ( That's what made me aware of this bug. ).
This wil definetly upset the awd cars when shifting mid corner. You are happily loading up the center diff, then flit into neutral and it's fully open, then fully loaded a short while after. Good catch.
Don't be such an LFS apologist then, going for the cheap excuse "oh it was because the car was flipped." BS. What "quite a few" issues are you talking about, btw? I don't remember anything that was caused by the car being on it's roof. It's just a handy technique used for easy demonstration of what forces are applied on the wheels since there's no road in the way.
I'd be greatly surprised if the car behaviour in LFS would have any dependency on its orientation, since that would be a pretty crappy hardcoded canned thing to code in. The only thing that'd change is the direction gravity is acting on the movable sub-components, i.e., the tyres.
True, didn't think of H-gates doing that, though I don't understand what effect this has on racing? Sure, left and right wheels do need a constantly active diff or there would be a difference whenever you corner and put it in neutral, but the central diff is only active when the front and rear wheels somehow have a speed disparity, which certainly won't happen in any great fashion in neutral.
The only place where I could see it have an effect is under braking if you somehow rely on the front wheels preventing the rear wheels to lock up via help of the LSD, but that would be quite a shitty setup anyway. Correct me if I'm wrong
well the centre diff is in operation every time you corner, that is its reason for being there and if you were to shift with a hshifter during cornering it would have an effect of unbalancing of the car due to the suden change in state of the centre diff, the only place were it wouldnt be a problem would be on the straight
A centre differential tries to equalise any speed difference between the front and rear wheels. Its main use is in low grip four wheel drift situations (rally), to make sure you can spin all wheels instead of one end of the car (front or rear) spinning the wheels getting all power and not allowing any torque to go to the other end.
If you are in a throttle neutral situation (which is definitely the case when the car is in neutral) then the centre diff should be hardly active at all, since just driving/rolling around a corner does not generate a significant speed differential between front and rear tyres.
Good point, but what about downshifting under braking? If you're locking up or on the verge of locking up under heavy braking the sudden release of the center diff can be enough to break traction and lock up a wheel...
this bug might only be possible to reproduce reliably with the center diff (due to being able to slow down one side off the diff manually) but that doesnt mean it doesnt apply to all viscous diffs or maybe all diffs in general
most if not all setups in lfs use heavy engine braking with the brake balance set to take engine braking into account
id be surprised if the sudden shift in bake balance while downshifting during heavy trail braking wont unsettle the car
additionally you have to take into account that the fully open center diff will drastically change the torques going into the torque sensing front and rear diffs
also as far as i can tell the drivetrain in lfs is pretty much weightless and the diffs also react instantly so there doesnt seem to be any inherent low pass behaviour in the drive train model and introducing wide band spikes while shifting may cause some pretty curious transient effects (although the spongy tyre model might filter them all away instantly)
Actually it's rather easy to verify it only happens on the centre LSD. I just did in 5 minutes of testing with the RB4 on the autocross track, chase-cam and forces view on.
Normal diffs:
Accelerate to a high speed, clutch in and start cornering with a high locking diff (clutch pack or viscous at maximum). Flip the gearstick between in-gear and neutral, notice no difference in cornering behaviour and forces view. Would neutral open any of the diffs, the change in cornering would be drastic. Also note that the central diff (which does open up and was set to max locking in the test) has no noticeable effect whatsoever.
Central diff:
Set central diff to 20Nms/rad, brake balance full rear at 1200Nm brake force, disable ABS. Accelerate to high speed, clutch in and hit the brakes. Note on the forces view how the front tyres assist in braking due to the centre LSD. Put the gear stick in neutral and see how the front braking disappears at an instant as the centre LSD opens up. This is the bug, and also the way to show the greatest difference in behaviour it can cause (= not all that much even in this extreme case).
So? When you clutch in / shift to neutral there is no engine braking.
Remember, what we're looking at is the difference between clutched-in and neutral-gear situations.
(Also "heavy engine braking"? Come on, the engine brake effect is rather small. Sure it's enough to shift the balance or lock the wheels when you're already on the limit, but the braking effect itself isn't all that great.)
What torques? The only torque coming through when braking is that of a front-rear wheelspeed difference, i.e. when one set of tyres has locked up.
The more I think about it, the more I believe that the actual effect on racing, even during braking is minimal. The unsettling of the car from the sudden disconnection of the engine (by clutching in) seems to be far greater than that of the centre diff opening up while already having disconnected the engine from the drivetrain.
Even ignoring that and concentrating just on the braking difference between clutched-in and neutral-gear state, keep in mind that the centre diff is still a viscous one - only a big speed difference will have a noticeable effect, and this speed difference is only given if the wheels have already locked up. It also becomes smaller and smaller the more the non-locked wheels slow down due to braking. And who the puts and lets the gear in neutral during braking anyway?
