Quote from AeroMechanical :

I've found in the FBMW that you can up-shift as described by the real engineer as his 'preferred' method (in another thread somewhere) by just keeping the throttle floored and giving the clutch a tap as you shift. Pretty cool. It probably doesn't gain you anything, but seems to work fine- doesn't raise the clutch temp any more than lifting normally (meaning not at all really). Still, in that car, I find it easier to just not use the clutch at all except accelerating from a standstill and corners requiring heavy braking for the sake of simplicity.

Love,
Aero

I've read some "user's manual" from a sequential gearbox manufacturer. They recommend, in upshifts, to "gently" pull on the stick while flooring throttle, then lift it for a split second when you want to shift. When the gears get unloaded the stick should move and change into next gear automatically. Else just pull on the stick and rev up to the limiter...

I find it works fine in the FBM, although it feels a bit weird to have stick movement out-of-time with the actual change.

Actually, the only variable you should consider is aiming for 0 clutch slip when you release the clutch. It is very obvious when you're in a real car, because when the clutch is lifted with slip you will feel a sudden acceleration/deceleration.

Generally, the factors that make you blip (lift) more:
a) you're shifting at a higher rpm
b) the two gear ratios are farther apart
c) it takes you longer (shorter) to move an h-shifter

To atledreier
If you rev the engine to way over redline, it should explode. Some rods will simply break and the piston might fly out punching through the hood. You can't take care of that with water and oil temperatures, ever. There's a Youtube video where a guy revved his Eclipse to 10k, bam! And you could hear metal stuff dropping to the ground. There's another one in a motorbike drag race, who has the left hand blown off by a piston that flew out.

Quote from Glenn67 :

meh, that's why suspension tunning is often seen as a black art. Which doesn't mean it is hard, it's just that there is a hell of a lot of information out there to assimilate. And often a little information is just as dangerous as no information but hey it's fun, like a game within a game

For me, I find two things helped. First is an understanding of the spring-mass-damper system, so at least things like frequencies and damping constants mean something. The other is some idea in linear algebra about choosing a basis to get decoupled variables. For example, ARB's really shouldn't be viewed in the variables (front, rear) for the stiffness, it should be viewed in something like (front+rear, front-rear), where the first "variable" tells you how much overall body roll you get, and the second tells you about over/understeer balancing.

Because of the narrower Blackwood?

Quote from Mako. :

So, I logged on, got patch y, got a drift setup from a friend, and logged onto a server. it was I think, KY GP. I tried my best, but I had a lot of trouble, even after I practiced for 4 hrs, I didnt FEEL like there was much improvement. Here's the hook, through. Yesterday, I had a lot of trouble with the DFP, I kept trying out different FF settings, and I just could not wrestle the wheel, I always reacted too slow, and couldn't turn it fast enough to do much of anything!

I had it at 50 at first, and then went all the way down to 35... still sucked at it. Today, however, I went on an auto-x setup for drifting, on some server, and had a ton of fun, WITH MY WHEEL SET TO 100ffb! And I kind of did it subconsciously, not even noticing it untill I pause for a sec, and went.. "wait.... holy shit, WTF?"

Then, I progressively got worse and worse for some reason... maybe it was just the wheel overheated or something and didnt react as fast, maybe it was me, dunno. But, I had quite a bit of trouble with it. I dunno how I even really did it, but all I know, is that it felt EASIER than 50! Does anyone else get this?

When I still had a DFP, I grew very tired trying to drift and got progressively worse in a while. But isn't 100% in LFS just too much? Have you tried setting the DFP in "hardware 100% FFB" mode, by pressing "right paddle+select+L3" (or something similar, can't remember)? I think it helped a bit.

Quote from tristancliffe :

In bike gearboxes there is a neutral between 1st and 2nd (which is why bike 'boxes are usually written as 1 down, 5 up), not between each gear.

Well, in bike gearboxes, some have "neutral" between every two gears, some don't. But that "netural" isn't really neutral. It's just some void that you fall into if you mis-shfited. I've ridden bikes with and without it.

How a sequential can mechanically jump from any gear to neutral is beyond me... Or is it just a pre-programmed sequential shifting back to neutral? All of F1's shifting goes through the computer anyway.

Good job! This is what I've always wanted to know.

May I suggest you plot two 2-D plots: magnitude of lateral force, and magnitude of longitudinal force. Both plotted versus long. and lat. slip % (hence 2-d plot).

I have seen this kind of plot from the Forza physics Dr.'s webpage, and it's very communicative about force combining.

Don't the usual GTR cars have minimum weight or kg/hp per regulations? I think they could, IRL, just make the car over light then ballast to the heart's content.

Quote from Bob Smith :

Well, yes, if you are cornering and then apply throttle, the outside wheel will be turning faster to turn a larger radius, but once the grip of the inner tyre is reached, it will begin to spin, so the outer wheel will then get the majority of the torque until the inner wheel stops spinning. So, depending on the situation, either wheel could get the majority of the torque, and that's before bumps are taken into account.

I thought the pre-load only acted as a minimum torque difference, so shouldn't have any effect most of the time (assuming low pre-load anyway).

Gah, another thread been has taken over trying to understand diffs.

I guess you're right about inside wheel spinning!

Isn't the preload the force "pre-loaded" onto the clutch pack before engine torque gets loaded through the ramp? So preload makes sure there's still some locking when you're at neutral throttle or when you clutch. Low preload gets drowned by the ramp, perhaps, but still the force on the clutch pack should resemble (locking*torque+preload).

Sorry for the hijack... Is it possible to pull this into a new thread?

To tristancliffe:
Yes they do.
http://auto.howstuffworks.com/differential3.htm

Quote from Bob Smith :

Yes, pitching (be it squat or dive) is affected by both the front and rear springs. But to get understeer you have to be turning, not braking, by definition. I'm unsure how this would affect trail braking.

Yes I was thinking about understeering at trail braking, or the very initial part of the turn-in.

Quote from Bob Smith :

So the outside wheel could get 130% of the engine torque while the inside wheel is given -30%?

First, the outside is likely spinning faster than the inside, so the torque to the outside should be lower than the torque to the inside. LSD's, whichever kind, always takes torque from the faster side to feed the slower side.

Second, I'd say "the torque at the outside wheel is reduced by the amount (30% of engine torque + preload), while the torque at the inside wheel is increased by the same amount." This is on top of the 50:50 split of engine torque by the (open) diff.

Quote from Bob Smith :

Friction having an impact might make more sense, although a tan function will let you go past 100% lock with low ramp angles. I don't totally believe the formula I found either, though. It was just mentioned in some document I found while googling just the other week, and doesn't tie up with the values posted by KartRacer either.

You acutally can get higher than 100% force scaling out of a ramp, like how you can get higher than 100% of input torque out of a gear ratio. Ever see people splitting big rocks? Just insert a wedge into a crack and hammer that wedge (like a ramp anyway), and the hammering force gets amplified beyond 100%, proportional to cot(wedge angle), too.

Quote from tristancliffe :

Not really, no. You can do the maths on spring rates quite easily (as a starting point), starting with either wheel rate (N/m or lb/in etc) or wheel frequency (Hz or CPM), but you still need to know what you want in those cases.

There is no magic value for spring selection.

After some experiments...

In LFS, the "stiffness" for springs and the damper values, I think, are converted to equivalent wheel rates. Because in the FZ5, the front spring is close to perpendicular, but the rear spring is close to 45 degrees, BUT I've set all the stiffness and damping rates' ratios equal to the weight distribution and the car obviously has the same frequency front vs rear. If the "stiffness" meant the spring rate, but not the wheel rate, the rear will be under-springed/over-damped compared to the front due to the large inclination (?), but it's not the case.

So yes, I should have said "stiffness" instead of spring rate, because in LFS we seem to get the equivalent wheel rate directly...

To Bob:

It seems that the front has to hit bumps first could mean that softer front is better. So, did your "neutral" mean that "wheel" rates' ratio equals weight distribution? (Which should mean same frequency for both ends.)

As for adding understeer by stiffening the front... I don't think it's the best method, or can achieve what the ARB can't, because:

a) In a straight line, braking with the same effort, the steady state load increase in the front is only affected by the car's pitch change (on top of the so-called "weight transfer"). In this case, stiffening both front and rear have the same effect of reducing pitch change, which means to add understeer through springs, you should stiffen both ends, not only front.

b) When cornering, stiffer front springs is exactly the same as stiffer front ARB.

I'll have to think about transient effects...

Last edited by yoyoML, Fri 21 Dec 2007, 15:30.

Quote from Bob Smith :

I thought the relationship between ramp angles and locking factors was a simple cosine, so that:

60° = 50% (rather than 40%)
80° = 17.5% (rather than 20%)

???

I think it'd be tangent, for using a ramp to scale force in one direction to another:

(locking %) = A * 1/tan(ramp angle) = A * cot(ramp angle)

where A is a constant determined by ramp placement and friction coefficients of the clutch pack.

Quote from tristancliffe :

It's partly because it's wrong (the Race_S setups aren't very well researched, and aren't meant to represent a real setup necessarily), and partly because weight distribution isn't thw whole story - wheel frequency, unsprung masses, the wheel rate (and rising rate if applicable) etc.

Probably I asked this because the RACE_S setup is too strange. Still, is spring ratio = weight distribution a good starting point?

AndroidXP:

I think the amount of "dive", or pitch change, is just related to the overall spring stiffness instead of front vs rear ratio. For example, if you take X N/m of spring off the front and increase the rear by X N/m, the pitch change will be exactly the same in straight-line braking. The only difference is that the car (CoG) will sink a bit due to the front compressing more and the rear extending less.

Hmm... I'll think about it more.

Last edited by yoyoML, Fri 21 Dec 2007, 15:28. Reason : It's "rear extending [B]less[/B]."

I remember the first time I started some pre S1 version and it BSODed Windows, then I avoided LFS like plague for a while. Besides, the older versions ran really badly on my Radeon 8500 (like mentioned), the versions that had solid black windows.

Some time later hardware shader supprt (??? not sure) was added, which boosted performance big time.

Oh, I remember XRG+3 pro AI+3 laps Blackwood for quick earning of credits. And XRT was called GTT with the 6 gears. Engines made popping sounds if over-revved. The second-last corner on Blackwood (downhill right) had a much larger tarmac run-off area on the left. Someone tweaked track width to zero to make a motorcycle.

I only bought S2 this year, but had been playing demo for quite a long time before...

I was looking at the RACE_S setup for the FZ5, which has some 60:40 weight distribution, but the spring rates are like 125:105! If you "drop" the car it is quite obvious that the nose bounces up way quicker than the tail.

So what is the reason for this very strong front springs compared to the weight it bears? If it's to reduce some power oversteer, then why not just use the anti-roll bar?

I think one of the drum's problem is that the braking force saturates, and is not quite linear to the input. It may be due to the "self actuation" (if you read Howstuffworks), so that braking force rises quickly for the first part of pedal pressure but rises slowly after. Disc brakes (fluid+piston) has no such non-linearity, only deviating from true linear response due to the load sensitivity in the pad+disc.

Quote from bbman :

That shouldn't stop it getting weaker when you're really overdoing it...

Drop-off is there, but more subtle than in ISI-sims (due to their canned effects), but it definitely is there (or I'm imagining things)... 101% with no center spring in profiler, 60% in LfS here...

ISI's FFB overshoots the drop-off, due to the fact they take only the lateral force and modify it according to % of contact patch slipping. In short, their FFB
is closer to what you get from the self-aligning torque with 0 caster. That thing drops to 0 when you're totally sliding the front tyres, and is reflected in some hands-off stable understeering in all ISI games.

LFS instead takes the real forces through the steering column, and the effect of caster is implied. Positive caster should exert a torque that somewhat drowns out the simple drop-off in self-aligning torque, so
a) the drop-off point, if any, should be delayed
b) ultimately the FFB never goes to 0

Hope it helps.

Quote from deggis :

Mafia, even that it's not a driving game it has clutch support.

Totally, completely agree.

You get to shoot people up and drive their cars, too.

I would prefer flight-sim style panning views: one click of view right and release = view stays at 45 degrees right; another right click = view stays at 90 degrees right. I think even with the head turned only 90 degrees each side, we still turn our eyeballs to see some total 135 degrees each side, so probably from left 135 degrees you click view right 3 times to center and another 3 times to look 135 degrees right. Considering in flight-sims you also have view up/down panning and the scheme works, I'd say in race-sims it would work even better.

A lock to nearest car could work. It could be activated by: "press view left, press view right, release both" to lock to the nearest car on the left. It could be deactivated by single clicking either view left or view right to snap to the nearest set (0, 45, 90, 135 degree) view point. But an instant "snap" to the targeted car would be disorienting, while a "quick smooth pan" to the target should be better.

Last edited by yoyoML, Fri 14 Dec 2007, 05:45.

Quote from P1lot :

People have moved it laterally with the addition of an extra plate. Check here http://forum.rscnet.org/showthread.php?t=277657

Moving the brake pedal somehow makes me feel uncomfortable. It means when you brake you not only exert a force on the pedal, but also a torque that twists it toward the gas pedal side. The original hinge probably wasn't designed to take twisting torques like that.

And considering the brake springs are the hardest... I would move the gas pedal toward the brake instead. Or only the lower part of it rotated to the left a bit. Combined with some spacers under the brake pedal it should be fine.

1. Clutch (whether braking or not)
2. Blip + Shift
3. Release clutch

When the clutch is in, blipping and shifting do not interfere with each other anyway, so they can be done in any order or simultaneously. To save shifting time, of course you'd like to blip and shift simultaneously. The only important thing is that the moment you release the clutch, the revs are matched between the two sides of the clutch.

Quote from Peptis :

The shape of the curve is determined from the physical model.

I would guess that the curve would have to be monotonically decreasing and that its second derivative would be positive over the range. With these constraints you can get a pretty good approximation of the curve with only three points. If we had a theoretical model for the tyre curve then we could do even better.

And that the curve should converge to some positive value at very high loads, representing the shear strength limit. At very high loads, even higher load should just make the rubber break away from the surface faster instead of adding any lateral force.

mu = a*exp(-load/b)+c <-that should do the trick LOL!