Initial Torque: What do you do?

[oppositelock]

Alright, I think I've found a way to end this debate once and for all.

Now, after checking dozens of stock LSD settings the most common I see are 5/20/10 and 7/30/15. There are some oddballs like 50/80/0, but these are the most common. Now, for the "initial is inverted!" crowd, riddle me this:

Why would the 5/20/10, which is supposed to emulate a rather limp factory LSD, have a fully locked initial that would look excessive even on a race car?

Why would the 7/30/15, with it's higher accel and decel, have a looser initial than the 5/20/10?

Why would the fully adjustable LSD, with it's 10/40/20 default, have a looser initial than the 7/30/15?

Why would the entire adjustable range of the full custom LSD initial setting be dedicated to loosening it from the stock 5 or 7 value, rather than increasing the amount of lock? Do stock diffs really have so much lock that they need to be loosened for racing?

I know PD does weird stuff, but this doesn't make sense on any level whatsoever. Lower value = less lock. Case closed. And, once again, the LSD is locked when static and the initial value indicates how much force is required to unlock it, not vice versa.

 

[Slurm]

I agree with everything except that last statement. I believe initial is the strength of the diff when at rest. This is what it means for real diffs, so I would expect it to work in a similar way on GT5.

 

[oppositelock]

I agree with everything except that last statement. I believe initial is the strength of the diff when at rest. This is what it means for real diffs, so I would expect it to work in a similar way on GT5.

I think you're agreeing with me and don't know it. Both your statement and my statement are totally compatible.

 

[Slurm]

lol, I re-read and yes, I agree.

 

[itstheDRE4M]

Description from GT4.

Initial Torque ----

Apart from cornering where limited-slip is fully applied, it also slightly limits slip in the differential at other times. The force applied to the limited-slip in advanced is called the initial torque. It is possible to change the handling of the car by changing it's initial torque. Increasing the initial torque increases the response when initial-slip is active, but makes the car harder to turn. On the other hand, a weak initial torque makes the car easier to turn but reduces the response of the limited-slip.

GT5 Description ---

Initial Torque --

Initial Torque is the amount of torque necessary to activate the LSD, and prevents sudden changes in performance and handling when the LSD is activated during acceleration and deceleration. The stronger this setting, the less likely unexpected changes are to occur, but the more likely you are to experience understeer.

 

[Will27]

This site has a good read on how real LSDs work.

http://www.ppi-ats.com/LSD/LSD_basics.html

You'll find info on initial torque at the bottom of the page.

Thanks for finding that Filled the gaps in my knowledge and explains everything im feeling in GT5

 

[Poppins]

Alright, I think I've found a way to end this debate once and for all.

Now, after checking dozens of stock LSD settings the most common I see are 5/20/10 and 7/30/15. There are some oddballs like 50/80/0, but these are the most common. Now, for the "initial is inverted!" crowd, riddle me this:

Why would the 5/20/10, which is supposed to emulate a rather limp factory LSD, have a fully locked initial that would look excessive even on a race car?

Why would the 7/30/15, with it's higher accel and decel, have a looser initial than the 5/20/10?

Why would the fully adjustable LSD, with it's 10/40/20 default, have a looser initial than the 7/30/15?

Why would the entire adjustable range of the full custom LSD initial setting be dedicated to loosening it from the stock 5 or 7 value, rather than increasing the amount of lock? Do stock diffs really have so much lock that they need to be loosened for racing?

I know PD does weird stuff, but this doesn't make sense on any level whatsoever. Lower value = less lock. Case closed. And, once again, the LSD is locked when static and the initial value indicates how much force is required to unlock it, not vice versa.

So by your logic, explain to me why when you have one wheel on grass with 5/5/20 and 60/5/20 they act like they are locked the same amount. Because from what youre saying that shouldnt happen, and it does, because like I said initial torque doesnt affect how hard it is to lock it up.

 

[Poppins]

Description from GT4.

Initial Torque ----

Apart from cornering where limited-slip is fully applied, it also slightly limits slip in the differential at other times. The force applied to the limited-slip in advanced is called the initial torque. It is possible to change the handling of the car by changing it's initial torque. Increasing the initial torque increases the response when initial-slip is active, but makes the car harder to turn. On the other hand, a weak initial torque makes the car easier to turn but reduces the response of the limited-slip.

GT5 Description ---

Initial Torque --

Initial Torque is the amount of torque necessary to activate the LSD, and prevents sudden changes in performance and handling when the LSD is activated during acceleration and deceleration. The stronger this setting, the less likely unexpected changes are to occur, but the more likely you are to experience understeer.

This description from GT4 fits what I think is right completely. The higher the Initial the faster the response and the "harder" it kicks in but it doesnt affect how easy the diff actualy locks, it just affects how hard it kick in when it does end up doing so.

And Accel Sens is how easy it kicks in.

 

[itstheDRE4M]

This description from GT4 fits what I think is right completely. The higher the Initial the faster the response and the "harder" it kicks in but it doesnt affect how easy the diff actualy locks, it just affects how hard it kick in when it does end up doing so.

And Accel Sens is how easy it kicks in.

Initial has nothing to do with lock, but how responsive the LSD is. 5 = slower --- 60 = faster. 60 gives you a quicker lock.

Accel / Decel is how sensitive the lock is. 5 = open (slip) --- 60 = locked (no-slip)

Why couldn't you think of this last night dannnng dude.

Still doesn't answer DECEL. because a setting of 20/10/5 gives you mad understeer under acceleration, but 20/10/60 has a substantially less amount of understeer.

 

[Poppins]

Initial has nothing to do with lock, but how responsive the LSD is. 5 = slower --- 60 = faster. 60 gives you a quicker lock.

Accel / Decel is how sensitive the lock is. 5 = open (slip) --- 60 = locked (no-slip)

Why couldn't you think of this last night dannnng dude.

Still doesn't answer DECEL. because a setting of 20/10/5 gives you mad understeer under acceleration, but 20/10/60 has a substantially less amount of understeer.

yeah that decel thing you posted yesterday just ****ed me completely.

 

[itstheDRE4M]

yeah that decel thing you posted yesterday just ****ed me completely.

You're not going to say those results are flawed??

 

[oppositelock]

Description from GT4.

Initial Torque ----

Apart from cornering where limited-slip is fully applied, it also slightly limits slip in the differential at other times. The force applied to the limited-slip in advanced is called the initial torque. It is possible to change the handling of the car by changing it's initial torque. Increasing the initial torque increases the response when initial-slip is active, but makes the car harder to turn. On the other hand, a weak initial torque makes the car easier to turn but reduces the response of the limited-slip.

GT5 Description ---

Initial Torque --

Initial Torque is the amount of torque necessary to activate the LSD, and prevents sudden changes in performance and handling when the LSD is activated during acceleration and deceleration. The stronger this setting, the less likely unexpected changes are to occur, but the more likely you are to experience understeer.

The GT help is the absolute last place I'd look for reliable info. Every one of them reads like they were translated by someone who didn't understand what they were reading from a text written by a guy in Japan who didn't even work on the game code. I remember a description about damping saying that the compression should be half as strong as the extension. Well yeah, if you look at a real shock dyno that's usually true, but if you set all your cars with something like 4/4 compression and 8/8 extension you end up with a bunch of lousy handling cars. I mean, are you really supposed to never use more than 5/5 compression, and then only with 10/10 extension? Of course not. GT likely takes this into account and makes each compression value softer than it's equivalent extension value, just as it should be. I mean, look at the car descriptions that were likely written by the same people. You would take those at face value? "Holy hell, this says my Corvette is actually a truck! It must be true, then..." In this case I'd say the GT4 description is slightly better translated and that nothing has changed in the LSD code.

 

[itstheDRE4M]

The GT help is the absolute last place I'd look for reliable info. Every one of them reads like they were translated by someone who didn't understand what they were reading from a text written by a guy in Japan who didn't even work on the game code. I remember a description about damping saying that the compression should be half as strong as the extension. Well yeah, if you look at a real shock dyno that's usually true, but if you set all your cars with something like 4/4 compression and 8/8 extension you end up with a bunch of lousy handling cars. I mean, are you really supposed to never use more than 5/5 compression, and then only with 10/10 extension? Of course not. GT likely takes this into account and makes each compression value softer than it's equivalent extension value, just as it should be. I mean, look at the car descriptions that were likely written by the same people. You would take those at face value? "Holy hell, this says my Corvette is actually a truck! It must be true, then..." In this case I'd say the GT4 description is slightly better translated and that nothing has changed in the LSD code.

I'm only comparing descriptions here and I only believe GT4.

 

[oppositelock]

I'm only comparing descriptions here and I only believe GT4.

Ah, sorry then. I misunderstood your intention.

 

[oopssorryy]

. I mean, look at the car descriptions that were likely written by the same people. You would take those at face value? "Holy hell, this says my Corvette is actually a truck!

You sir have just made my day!

 

[Poppins]

You're not going to say those results are flawed??

nope because i have no idea how that is possible in any way so i dont see how it can be flawed

i agree with you that 60 IT makes it more sensitive to locking though.

if you have 60/5/20 and take a corner it will lock.
if you have 5/5/20 and take a corner it will not lock.

i think it might be a %age of how much torque gets transmitted to the other wheel when it locks

 

[oppositelock]

So by your logic, explain to me why when you have one wheel on grass with 5/5/20 and 60/5/20 they act like they are locked the same amount. Because from what youre saying that shouldnt happen, and it does, because like I said initial torque doesnt affect how hard it is to lock it up.

They have the same lock on acceleration because they have the same accel value. This doesn't disprove anything I said.

 

[Poppins]

They have the same lock on acceleration because they have the same accel value. This doesn't disprove anything I said.

Im just saying that Initial Torque doesnt affect how easy it is to lock but rather the response of it and speed.

 

[oppositelock]

Im just saying that Initial Torque doesnt affect how easy it is to lock but rather the response of it and speed.

It doesn't effect how it locks. It effects how it unlocks. Once you get your head around that all will become clear.

 

[BlueShift]

Now try and take a corner on cement with it set to 5/5/20 and 60/5/20 and you will see that you spin your inside tire with 5/5/20 and not with 60/5/20 because its getting 60% of torque transferred to it (..)

The Initial Torque is just how much power is transferred.

60% more torque.

The Initial Torque is just how much more power is transferred.

60 initial give 60% more to the not slipping wheel, and the rest is halfed between the two wheels.

Like I said, it can lock the other way :

60/60/x : a high response/very sensitive accel/don't care decel

If left wheel slips enough to lock the LSD, say left roll 2 wheel turns for 1 rigth :
20% is given to the left wheel, 80% to the rigth (0+40% left/2 /// 60 of initial + 40 left/2).

That make spin the rigth wheel more and slow down the left wheel.

Result, the rigth wheel tend to spin more and should make the lock lock the other way : left gets 80% (was already spinning, so it will spin more) and rigth get 20% (it was spinning, maybe it'll stop or keep spinning).

This gives the left/right/left/right moves at corner exit, and brutal heads-to-tails in the middle of S turns on powerful cars.
That's also why strong initial are slow to stabilize (+ them being slow to unlock, too high initial tend to be bad unless you know exactly what you do : controling the unlock time to use more steering in the apex, but without making your car too hard to control in S turns or brutal changes of direction).

 

[budious]

You sir have just made my day!

Most of the descriptions have to be translated from Japanese text from PD's home offices over here in the states, I'm guessing somebody runs it through Google or Yahoo translator and then decodes the rough garble to proper sounding English, even if it's technically inaccurate translation, to put together the English builds for the game. Otherwise, if there was a full time translator employed for the sole purpose of this job, he or she needs to be given a pink slip.

 

[Slurm]

If you want to test the way this stuff works: Standing starts with one wheel on the grass - grip level off track set to real.

Find a car and a start-type that just about works cleanly with both wheels on tarmac - it's not much use of you'd be smoking both tyres on tarmac anyway.

You could try holding it on the brakes or pulling away from low revs. Once you've tried your test on tarmac, go to the edge of the track. I really can't be bothered anymore but if someone wants to test this, it seems to me the most scientific way of doing it and the most likely way of getting conclusive results.

 

[budious]

All you need to do is make a course using the Mt Aso template with lots of uphill switchbacks to navigate through, it's as punishing an LSD test as you will find on any of the included hard courses.

 

[budious]

I been playing around with LSD tuning tonight on a test car, not really going by any certain philosophy or method anyone has suggested, more a trial and error run... but I can say this is like trying to hit the lottery when the winning configuration might be 3 chances out of the 175,616 variable configurations.

I find a combination where initial and acceleration settings comply really well with each other and produce good lap times, I think a minor change - an increment of 1 - to deceleration should not really affect the current initial/acceleration balance, but boy am I wrong. Goodbye consistent laps... move it back the one increment and they return.

 

[Poppins]

Very interesting BlueShift, thats kind of how I thought of it but I dont realize how you mean 60% MORE torque to the slower wheel, I thought it was that meant when it switched over to the slower wheel, it got 60% of total (so 40% left / 60 right) but Id like to know where you got this 20/80% thing?

If left wheel slips enough to lock the LSD, say left roll 2 wheel turns for 1 rigth :
20% is given to the left wheel, 80% to the rigth (0+40% left/2 /// 60 of initial + 40 left/2).

Because If the Left Wheel is turning twice for 1 right turn before it locks that is probably 95% slip no? (minimum acceleration sens?)

Also your math would make me think that having it set at 30 makes it very stable under back and forth, because it has enough torque to transfer to the outer wheel and provide good forward power and less enough to not cause insane violent snap's

 

[Poppins]

It would also make sense to me that Initial Torque is "inverted" just like acceleration/braking sensitivity.

In the fact that an low Accel Sens of (5) is like having a high 95% slip and up high at 60 is like having a low 40% slip

While for Initial Torque it acts the same, set low (5) and it wont kick in if accel sens is set too low(because the amount of torque necessary to activate it is 95% to one wheel) and set it high (60) and it will kick in under light throttle (only needs 40% to go to one wheel before locking)

Which would make Accel and IT pretty much the same thing. This is how it seems to me in game.

Initial Torque is the amount of torque needed to lock the lsd (lower means more torque)
Accel Sens is the amount of slip/speed difference needed to lock (lower means more slip)
Decel Sens is the amount of slip/speed difference needed to lock as well (higher means less slip)

and so I dont seem like im wrong about the "inverted thing" all im saying is that LOWER means MORE is needed.


which would make sense as

5/5/20 would need 95% torque to the faster wheel to lock as well as letting tire speed differ by up to 95%
60/60/20 would need 40% torque to the faster wheel to lock it and tire speed difference by up to 40% (which makes it very tight like it is in game)
5/60/20 needs 95% torque as well but the tire speed can only differ by 40% which causes lock up anyways (in high power especially)
60/5/20 needs only 40% torque to lock but the tire speed can differ by up to 95% so it behaves like 5/60/20 with the disadvantages of high IT

30/50/20 on the other hand which is seeming very good and stable for me would need 70% torque to lock but only lets wheel speed differ by up to 50%.

I really think this is how it works now.

CAN SOMEONE AGREE ON THIS?

 

[nomis3613]

So many posts, so quickly! It's great to see people excited about picking apart GT5 tuning.

edit - for people willing to have clearer and better explanations (and official, how things worked really for GT4 LSD)
http://books.google.fr/books?id=WFF...t4 initial torque&pg=PA82#v=onepage&q&f=false
http://books.google.fr/books?id=WFF...t4 initial torque&pg=PA83#v=onepage&q&f=false

For tuning stuff, I trust that Piggyback GT4 book as much as I trust the in-game text (eg very little)
example 1: "the racing brakes kit enables you to reduce your braking distaces considerably". Wrong, the brake upgrade does nothing unless you've got super-sticky tyres.
example 2: "while your driving force will increase (by increasing LSD accel), your ability to turn will diminish". Misleading, this only affects handling while accelerating, turn-in is not affected by LSD accel.
example 3: "the higher the value of the auto setting, the more your transmission will exploit the full rev range of the engine". Totally misleading, taller ratios mean less multiplication of torque, so less acceleration. They should be saying "higher values use a larger range of revs so the engine is often not operating near peak power, also higher values give less multiplication of engine torque at the wheels. Both of these reduce acceleration.
example 4: it recommends use of TCS to compensate for a rwd car's tenancy to oversteer with high rear stabiliser values. This is a terrible band-aid solution!
Sorry for the off-topic rant, just wanted to clear things up since it was mentioned.

Haha its all good Im not a lost sheep I just know that initial torque doesnt affect how easy it locks but moreso how instant it kicks in.

According to real-life theory, you're sort of correct about how it kicks in, but that's really just a side effect. It's happening in your test because:
- accel=60 means the axle is locked when accelerating
- at each gearchange there is a brief period of no acceleration so for an instant LSD accel doesn't apply
- in this time if initial=60, the diff is always locked anyway so there is no change at all. So nothing strange happens
- but if initial=0, the axles very briefly goes from fully locked to fully open and then back to fully locked. This is causing the feeling of it "kicking in".

However, this shouldn't affect your laptimes, so IMHO the biggest effect of LSD initial is turn-in vs resistance to spinning out.

 

[Poppins]

If you look at my #'s it makes sense. (everything /100)

Low IT (5) Causes it to need 95% torque to one wheel before it kicks the other wheel in (5+95=100)
High IT (60) Causes it to need 40% torque to one wheel before it kicks the other wheel in (60+40=100)
Low Accel Sens (5) Causes the faster wheel to be able to slip up to 95% (Spin 95% faster) (5+95=100)
High Accel Sens (60) Causes the faster wheel to only be able to slip up to 40% (60+40=100)

again my %'s arent proven but they make sense, as you dont want more then 95 or less then 40 in any situation

so having 5/50/20 isnt going to lock as easy as 30/50/20, but wont really be that noticable except in corners and such

 

[Adrenaline]

so many posts, so quickly! It's great to see people excited about picking apart each other.

ftfy

 

[nomis3613]

^ yeah, that too!