I'd just like a more thorough explanation of each LSD setting, in regards to this game.
All most of us care about in this game, are cause and effect. Results rule realism.
By my testing, if I step on the gas out of a corner in my ZR1, and the wheels spin, and the car goes for a loop, I've found that lowering the Initial AND the Acceleration LSD, equate to more traction.
You've got it right.
Why, how, rhyme or reason have no bearings to me. I'd like them to, but unfortunately they don't. I just know it works.
Explaining LSD can be very long especially for 4x4. On top of that, there's a few ways and effects that melts to explain this and both "schools" gives results. That's why people don't understand it fully because they see effect from school 1 with school 2 and don't understand.
School A (like ABC school) aka the most important things to understand :
Initial : the strength of your differential, the power they give to your wheel when they locks. Controls the understeer/oversteer (few = understeer, lots = oversteer)
Accel : the moment, or the minimal strength before LSD locks at acceleration (few = not often, a lot = often)
Deccel : the moment, or the minimal strength before LSD locks at deceleration and engine braking (neutral gaz/brake) (few = not often, a lot = often).
In school A when the LSD locks depending on accel/decel values, it gives you an oversteer/understeer regulated by the initial torque value.
This explanation covers 90% of LSD uses and functionalities.
But...
School D (aka the Downside school, or
The Dark School Of Despair, Don't You Dare ?) :
Initial : the strength of the constraint given to your wheel. The more of it, the more impetuous your car will be and the less your LSD locks (because you have to apply a bigger strength to your wheels at one moment to lock it)
(few = LSD lock very often but very soft (given School A), lots = LSD locks not often, but very strongly (again, given School A)).
Accel : the spinning probability when your LSD locks when you accelerates (few = few spins so your two wheel turn the same = less grip in turn, more in straigth lines and quick corner (school A main effect), lots = lots of free spin so your two wheels don't turn at the same speed = less grip in long curves or quick corners (quick corners is school A main effect), more in mid curves, less in straigth lines)
Deccel : same as accel but when decelerating.
Quick cornering will always be School A main effect, because the strength differences you put to your wheel is very strong anyway (strong braking, strong turning, strong accel), enough to lock any LSD unless you're in Citroen 2CV (not enough power to activate it).
This is something you want to avoid in your setups, unless you know exactly what you are doing. If you come that way, it's the Dark Side. It's very specific to some cars. And sometime no LSD (stock's 0/0/0) is good. Like on the 2CV for example, because the true good setting would be something like 2/1/20 and 5/5/20 is too far from that.
(btw yeah I tuned a 2CV, yeah I used it online, yeah I won the race
). I love the "dodoche" like we nickname her in France. When I was a boy people around gave them because they couldn't sell them. Now this car became a collection model, it cost 15/20k euros
Scenario 1: As I accelerate out of the corner, the car gets VERY loose, and want's to spin out.
Lowering the initial -5 will make this condition better or worse?
Raising the Accel +5 will make this condition better or worse?
I use he school A to answer to this. Unless you reached extremes values (7 or 55), you won't see much of school D results and effects.
- Lowering the initial will make the car less willing to quit the surface of the earth so it's "better" for the pilot. But initial also plays a role when decelerating so if you're happy with how your car decelerate, don't touch it.
- Raising the accel +5 will say to the car "clear for orbit". In this scenario, it's "worse" for the pilot.
Now please note I used comas, because in hand of a talented pilot, they will know how much pressure they can give to the gaz at corner exit, and still profit from a high LSD while in mid-curves. High initial will be profiting driving wheel players, low LSD will be profiting DS3 players. I use DS3 so my LSD perception is biased by my lack of precision on the accel/decel control. And I have no clutch which could mess a lot with your LSD too.
Also high initial/accel LSD helps a lot in quicks corners. If you tune this perfectly, with a good suspension, your car will "autoturn" : the weigth of your car will suffice to lock te LSD and turn your car at your place. Check my
Abt Audi TT a few laps around the Nür to see what I mean, or
the Scuderia around this or that track, or, if you have big underwears, my
Yellowbird in drift challenges without driving aids around Suzuka or Eiger.
Scenario 2: On corner entry, I coast in and through the apex, but the car has an understeer issue.
Raising the initial +5 will make this condition better, or worse?
Lowering the De-Accell -5 will make this condition better or worse?
- Initial +5 will be better under school A. But it will touch your out-of-curve results.
- Lowering -5 will be worse under school A. Low decell = high stability in corner entries but low while in curve.
By the way : these scenario are very good to tune the basics of a fine-tuned LSD. But it won't cover everything. You have a quick cornering scenario too, and a corner entry scenario to cover aswell. That's why I use Rome and Highspeed ring (invert High speed ring, at the end of the tunnel, have an excellent corner entry scenario and the whole track is an excellent scenario 2).
For extreme mid-curve scenario, you have Daytona.
For extreme quick cornering scenarios and scenario 1, you have London I guess.
Suzuka is very interesting. It's a all-scenario track, you should use it a lot for LSD setups.
Is 5 at a time, to large of a jump? What increments do you recommended to begin with?
I use +/-4, not +/-5, because you can divide 4 by 2 to fine tune it ! Then move to +/-2 then +/-1
If I want to move fast, I use +/- 10.
I should give a LSD pattern, but re-discovering school D last week messed it.
Depending on the traction of your car, LSD can be set at mean values.
The basic pattern of 10/40/20 isn't good to work with.
Most FF will be around 30/10/40
Most MR will be around 8-15/40/20
Most FR will be around 10-20/20/40. Some FR acts like MR because the transmission to the rear is very heavy (my Jaguar have even a RR LSD since her engine weigth nothing compared to the transmission unit).
Most RR will be with a school D 5/5/10 or even a opened LSD like 5/5/5 if you have big spinning problems, or alternatively on some very stable RR (does it exist ?), you could go up to 10/20/20.
4RM depends where is the engine (well, the weight in fact) is and what is your torque distribution. Their integral transmission parts canb be very heavy, so the engine is not really a true indicator.
I came with a formula, that is "almost" right but forgot school D, and I gave it some days ago, but it's not yet polished with school D. I used a refined version I will post now for my
CTR2 and it's pretty working. I have to remove a few points of accel front/rear though.
This is only for road setups. I forgot how to tune a rally LSD for 4RM (I was one of the few to try to tune rally cars in GT4 in another forum). But the good thing is my PS2 with GT4 and my 700 cars save is back online and my old darkside highly experimental Speed 12 setup still "works" somehow in GT5
... So nothing is lost.
Basically, you have to see your torque repartition. Let's say it's 30%/70%.
You start to use a pattern.
Front : it should be something like x/30/y
Rear : something like 25/30/z
x = 70% (coming from your rear torque repartition) * your front initial 25 + your front initial 25. In this exemple, it gives 0.7*25+25 = 42.5. Try 42 with rear weigth 4RM, try 43 on balanced or front weigth 4RM.
y and z depends a lot on your car.
y * your rear drivetrain = z.
In this exemple, with 3/4 brake settings, 0.7y = z.
With 4/3 brake settings, y is more like equal to z (and fun is out, I think).
On some 4RM, the car is very stable so you want a lower y and z, like 7/10. On other you can go until "extreme" decel values like 21/30. And on a few special car, you will go to 42/60 (
Citroen GT) but don't slow down with a wheel in the grass.
The accel 30/30 can be moved also depending on first school D problems you'll get or where your weight is (Rear like 20/20, front like 40/40). The more you'll have without school D problem the more it'll be good.
Now
all of these are only pattern, not final results. Some car will have a very special chassi that could derogate to all these rules.
You start from there and move sligthly values until you have it "rigth". I recommend only touching the front or rear LSD, not both unless you know exactly what your are doing (it require a few succesfull 4RM setup
).
Don't forget tuning an LSD can't be made under 1 hour of driving. Any player that would say the opposite are either newbies or very talented and gifted people. And people that's would use tuning generator, esp for LSD, are complete morons thinking it would give good results. these pattern are to be refined. Slowly, and by feeling and driving experience. No computer program can give you that.
Also,
my ZZII LSD is quite bull**** as it is now. I reworked it to an ok version but still unhappy with an understeer problem at the high speed turn in Nür's long straigth line (=> grass turn). The others are ok.
And don't forget that LSD should be fine tuned only
after your suspension is definitive. You can use pattern before.
LSD also comes after definitive brakes (decel) and aero (initial/decel, sorry I just can't see accel dependance, it's too fine for me).
Also, the gearbox have
a very huge effect on some initial/accel setting (short gearbox must use low initial/accel setting, wide gearbox = more values of intial/accel to play with). I'd say 4 clicks of difference.
For aero, I'd say 3 clicks of differences, brakes 4 up to 8 clicks, and suspension is too fundamental to find something correct when it's not definitive. Camber play also a big role in LSD setting (4-8 clicks at least).
Toe dependancies... Toe like suspension is fundamental too. Toe also completly destroys your LSD readability and perception of oversteer/understeer. You should always tune toe last getting back to LSD is something is wrong, then toe then LSD etc. The very first thing you should do before removing driving aids is putting 0/0 in toe on any setup. Toe use makes abrupt grip/don't grip things in your tires that can force you to fundamental changes in your LSD, and you'll never know if it's the LSD or the toe.
I don't like toe, plus I'm a **** at tuning it.
For Toe/Camber dependency on Initial/Accel/Decel, I'd suggest playing with the
E-Type Jaguar setup. More camber on the rear and you should lowering your initial/accel but there you can't, LSD is open. More camber at the front and you can put more accel/decell, but try that and you're spining around at corner exit. Like for the Yellowbird, I learned a lot on LSD there (and on Ride heigth/suspension but that's another story).
Currently, I'm working on perfectly understanding school D, and discovered a school S (aka Secret school) last few day, but it may be only guesses/superstitions/ignoring stuff. I put effects in it like engine hiccups, LSD, "counterturning retaliation" (try a stock maxed out Yellowbird to see what I mean : not all of this is given by the suspension.) and LSD loosing effect with too low/high values (not only because of School D annihilating what School A gives : there still things I can't explain.) Maybe all of this are bugs when the game calculates things and found round values (esp for engine hiccups)... It may be School A from/to School D transition effects... But I'd like to be at least 50% sure.
(edit - engine hiccups is turbo lag, nothing to do with LSD).
I also understood the school M (Midcurve school) but I just can't really explain it, because I use "feelings" for that school. The patterns used are to give good results for that school too. It mean accel or decel won't do anything to your car while in highspeed ring's curves for exemple.
My Veyron is a perfect LSD of this but trying to acheive this result will gives you understeering cars (the Veyron can't turn **** anyway
and this setup's ride heigth is to be re-worked anyway). At last it became a good car to put on Highspeed ring.
Just to say, there's a lot of effects I don't explained here. And I may have put BlueShift's bull****s too
Also, the more you read and tune, the more you'll understand. It may be the second most difficult setting to understand (the most difficult to theorize, for me, is and always will be the dampers, esp damper's school S
). I'd recommend as usual the part two of Scaf's tuning guide. It explain school A very good and he gives hints for school D.
