T.J.
T.J. MegaDork
11/24/19 10:29 p.m.

So, I understand how RWD limited slip differentials affect a car's handing.

Let's say you have a clutch type limited slip where you can adjust preload, number of springs and accel and coast ramp angles. I am trying to wrap my head around how these settings will change a car's handling. I get the basic idea that with a freer diff, the inside wheel will end up spinning, so it seems that more locking on acceleration would help with getting out of corners. What I can't quite grasp is how the coast setting affects corner entry. With a RWD car, less coast locking effect tends to help the car rotate at corner entry, but on a FWD car the wheels turn, so it seems that more locking effect would not necessarily make the car understeer as much as a RWD car. Seems like too much locking effect will lead to understeer since the front wheels still have to rotate at different speeds when cornering.

The internet seems to have lots of info on RWD differential setup/tuning, but not much for FWD cars.

Learn me.

 

spacecadet
spacecadet Dork
11/24/19 11:05 p.m.

The simple solution is don't use a locking diff. Helical Diffs are the simple answer because they're not locking and make no noise.

The major thing to determine with a FWD diff is if you'll ever unload a tire or put a tire in a zero grip situation. If so.. You need a locking clutch type diff. Helical diffs are torque sensing and always need both wheels on the ground and they both need some grip to work. Ice under one tire renders a helical diff effectively useless for example.

If you do go locking, 1 way is the next answer because it's then not affecting handling under braking. But Cuusco also says you could use a 1.5 way diff. But doesn't recommend a 2 way in any circumstances.

 

Link to their document.

 

https://www.cusco.co.jp/en/pdf/LSD%2520Guide%2520Final.pdf&sa=U&ved=2ahUKEwis95nwxITmAhUKWK0KHUzPBAsQFjACegQICRAB&usg=AOvVaw3wV2ZJFpIKW6NWqWTaz57A

 

 

 

My experience is all with Helical Honda LSD's and I love them. They're super predictable and they make zero noise. A clutch type diff will make some noise.

 

Some Honda tech guys had a debate on this and their consensus seems to be that 1.5's stabilize a very loose FWD car under braking.

honda tech discussion.

Stubbs
Stubbs New Reader
11/24/19 11:52 p.m.

No direct experience, but when I was doing research on FWD clutch LSD there were two schools of thought.  First was only have lockup on accel.  It seemed that was the general recommendation, and I think it was more geared towards a one size fits all mindset.  Accell only (or one-way) eliminates the wheel spin on corner exit, and doesn't induce corner entry push.  It would likely be easier to adjust to and a less experienced driver may prefer it.    

2 way (lock-up on accel, and decel) was recommended to me if you had a good grasp of FWD handling dynamics ie. left foot braking and other tricks to get the car to rotate.  Thought being is you have better control under breaking with a 2-way, but you need to be understand how  to utilize that control to make it worthwhile.

There are 1.5 way lsds allow a weaker lock-up on decel which give you some of the benefit of the 2 way without as much of the corner entry push. 

Now if you have an LSD that allows you to manipulate the cross pin and ramp angles, and pre-load you can alter it to suite your driving style.  That's a  whole nother ball of wax.

Preload is set by either springs or belleville washers, the higher the spring force the higher the preload.  This sets how quickly the lsd reacts to torque.  A lower preload in a FWD car would reduce the tendency to push on low throttle turns.

The cross pin and ramp angle define how the clutch pack is compressed when subjected to torque.  The cross pin wedges against the ramp and spreads the pressure ring compressing the clutch pack and reducing slip.  The angle of the cross pin and ramp changes how much torque it takes to lock the diff, the lower the angle the earlier it locks up.  Additionally the shape of the cross pin plays a role with a rounder profile increasing the torque required.  1.5 way usually have a rounded cross pin and steeper ramp on the decel side.

Clutch plate order plays a role in how much torque is needed to lock up as well.  Normal order is FP(friction plate) DP (Drive plate):  FP,DP,FP,DP,FP,DP  You can reduce torque transmitted by pairing up Drive and Friction plates ie FP,FP,DP,DP,FP,DP the reduces the friction surface area.  Adjustability is determined by number of drive plates.

Here's a couple resources on how the different components are used for tuning.

https://www.cusco.co.jp/en/pdf/LSD%20Guide%20Final.pdf

https://motoiq.com/project-dc2-integra-getting-more-grip-with-os-giken-and-better-response-with-centerforce/

http://gplmotorworks.gplworld.de/GPL-Differential-Info.html

As far as how to set it up for best performance for FWD, that's not an easy question to answer.

A lot of it comes down to driver preference.  I think the catch all would probably be a 1.5 way with a medium preload.  That gets you some of the benefit of a 2 way with less entry and mid corner push.  Mid to lower preload would reduce push in on throttle turns.

I had bought a transmission that ended up having a mugen 2 way in it, that's what led to all the research.  The recommendation I got was the 2 way is the better diff for racing, but it requires more driver finesse.   

I ended up not running it and got a torsen instead since the clutch type chatter a lot, and my car is driven mostly on the street. 

  

Knurled.
Knurled. MegaDork
11/25/19 5:12 a.m.

More preload will definitely hurt turn in.  I've only ever driven the extreme end of that example (welded diff) and the turn in is horrible unless you are on the power so you can get some helpful tire slip on the inside front to allow the outside front to steer you through.

 

The more preload you have, the more aggressively you have to drive, in other words.

therieldeal
therieldeal Reader
11/25/19 7:21 a.m.

Helical/Quaife has never let me down on pavement.  No noise, no parts to wear out, it just works.

MrFancypants
MrFancypants New Reader
11/25/19 9:13 a.m.
spacecadet said:

The major thing to determine with a FWD diff is if you'll ever unload a tire or put a tire in a zero grip situation. If so.. You need a locking clutch type diff. Helical diffs are torque sensing and always need both wheels on the ground and they both need some grip to work. Ice under one tire renders a helical diff effectively useless for example.

 

Look into Wavetrac differentials. I'm not exactly sure what magic is going on inside of them (I'm not an engineer), but I believe that they offer limited slip torque transfer in no/low load conditions without the use of clutches. 

Nugi
Nugi Reader
11/25/19 9:26 a.m.

I have driven 1.5, 2.0 clutch, viscious, phantom,and torsen lsd's in FWD hondas.

Clutch are amazing, but require maint and apparently racecar noises bother some people (blasphemy!) I would strongly reccomend the 1.5 if going this route. The 2.0 felt like a spool, but could have been his setup. 

Torsen are THE choice for a daily. Better than nothing on the track, but as mentioned, you still have to keep your front springs and bar soft enough not to lift much. You will feel it, but its not in-your-face like a clutch. 

Viscous are hardly seen anymore. They tended to 'overheat' (maybe froth?) and get weird when pushed a long time. Never in the way, but never quite as much effect as you want. Avoid.

Fake LSD. AKA phantom grip, universal, lunchbox lockers, and other sub $50 ebay preload devices all suck. Better off welded or open diff at this point. 

Be sure to use the correct fluid suggested by the LSD mfg. Seen a few prematurely fail because the owner 'ran what they had at the store'.

spacecadet
spacecadet Dork
11/25/19 9:34 a.m.
MrFancypants said:
spacecadet said:

The major thing to determine with a FWD diff is if you'll ever unload a tire or put a tire in a zero grip situation. If so.. You need a locking clutch type diff. Helical diffs are torque sensing and always need both wheels on the ground and they both need some grip to work. Ice under one tire renders a helical diff effectively useless for example.

 

Look into Wavetrac differentials. I'm not exactly sure what magic is going on inside of them (I'm not an engineer), but I believe that they offer limited slip torque transfer in no/low load conditions without the use of clutches. 

I'm familiar with them and while they do appear to be a bit better than ATB or helical diff. They're still not as good as a locking diff. Second half of this video is damming. 



http://www.wavetrac.net/different/

 

 

Paul_VR6
Paul_VR6 Dork
11/25/19 9:54 a.m.

We've run high preload torsen in the drag cars for a long time and have tried them in the street and road race cars. Turn in is quite tricky but I think they could be a lot better than a traditional torsen with a car with a stiff front. Gary Peloquin makes them for the VW but it seems there are similar options for some of the other manufacturers. 

T.J.
T.J. MegaDork
11/25/19 9:54 a.m.

Thanks for the replies and the links.

I am wondering about race cars here, so clutch noises are not a concern.

It seems a bunch of journalists got to drive a Civic TCR. This article has a quote that says getting the diff setting just right is 75% of the setup.

Here is my take away so far:

Pre-load: The lower this is the more smoothy the acceleration and coast locking effect take place, so set this to the lower side. If too high it could lead to understeer at corner entry.

Set acceleration ramp angle aggressively (low) enough to avoid inside wheel spinning when getting on the accelerator coming out of a corner.

Coast ramp angle if too low promotes understeer, but too high may make the car unstable under braking.

Trail braking to get the car to rotate is a good idea.

 

Paul_VR6
Paul_VR6 Dork
11/25/19 10:17 a.m.

In reply to spacecadet :

I think the Wavetrac is working about how it should, otherwise it wouldn't move at all. A Torsen would be stuck not moving at all.

MrFancypants
MrFancypants New Reader
11/25/19 10:34 a.m.
Paul_VR6 said:

In reply to spacecadet :

I think the Wavetrac is working about how it should, otherwise it wouldn't move at all. A Torsen would be stuck not moving at all.

This...  and in the context of a race car I'm not of the impression that a Wavetrac would be the right choice over a more aggressive and configurable clutch type differential, but in a 2wd car that it transfers enough torque to get the car moving is good. I think if combined with the brake actuated electronic torque transfer systems some cars are equipped with (Volkswagen XDS, for example), a Wavetrac might be all you need outside of a competition setting.

spacecadet
spacecadet Dork
11/25/19 1:19 p.m.

you could also left foot brake a ATB diff to get the same effect as the wavetrack, I understand where it's different. But it's not enough of a departure that I'd consider installing one over a standard ATB or Torsen. they also claim lifetime warranty and I don't like the way it works enough to believe that. the meshing of the spring plate and the other inner half makes me ask questions on how that's going to wear. 

Knurled.
Knurled. MegaDork
11/25/19 1:33 p.m.

In reply to spacecadet :

The Wavetrac always seemed to me to be a regular Torsen type diff, with some preload so that it acted less like an open diff.

spacecadet
spacecadet Dork
11/25/19 2:16 p.m.

In reply to Knurled. :

It's an ATB with preload, not a Torsen. Well not the old school torsen. I see now that Torsen is selling a helical ATB  as well. where the older version is considered a Type A and the helical is a Type B

 

 

T.J.
T.J. MegaDork
11/25/19 5:04 p.m.

So, in general too much pre-load and coast locking act similarly to RWD. More pre-load and lower angles (more locking) lead to understeer for FWD and RWD. 

Acceleration ramp angle is opposite in that too much locking while accelerating leads to FWD understeer but RWD oversteer. 

Number of clutches has a similar effect in that more clutches give more locking effect. 

Ideally it could be setup so a bit of throttle lift or a little touch of the brakes in the mid corner will rotate the car if you are a little wide and going to miss the apex. If you are heading to the apex, get on the gas before it and feed in throttle as you unwind the wheel. 

Knurled.
Knurled. MegaDork
11/25/19 9:54 p.m.
T.J. said:

So, in general too much pre-load and coast locking act similarly to RWD. More pre-load and lower angles (more locking) lead to understeer for FWD and RWD. 

Acceleration ramp angle is opposite in that too much locking while accelerating leads to FWD understeer but RWD oversteer.

RWD understeer.  Nothing understeers like a rear driver with a spool.

 

If you pitch it sideways and are spinning the tires, sure, but at that point, you are beyond the scope of differential function vs. chassis dynamics and well into "what E36 M3 will the car put up with if I throw it into X dynamic".

devina
devina Reader
11/26/19 6:29 p.m.

I have been researching this for my road racing Integra and most road racing fwd cars run a clutch type diff.  There are some torsens but as mentioned above, a torsen loses its effect if you are hopping curbs.  My old crx racer had a Mugen clutch type that worked really well.  

 

I am leaning towards an M Factory 1.0 way clutch style over the winter.

therieldeal
therieldeal Reader
11/27/19 8:38 a.m.

If it helps sway your decision at all, my friends locally have an STU prepped Acura… RSX? which has broken over and over again due to clutch diff failures.  I think they have tried all of the popular brands and broke them all.  They finally gave up and put a Quaife in it… got it back on track for the last race day of the season… And wound up setting the STU track record.

Vigo
Vigo MegaDork
11/27/19 12:50 p.m.

A clutch diff on a race car has to be looked at as a maintenance item. There's no hydraulic pressure involved so the clamping force is not great, the clutch diameters (surface area) are small, and they HAVE to slip some to not be a locker. All that means that using it = wearing it out.  A helical or torsen is either fine (99% of the time) or its broken. Introducing a little preload into the middle probably doesn't change the wear factor by much at all. 

So basically, if taking out and apart a differential doesn't sound like fun to you, don't run a clutch type in your racecar. My  .02.

spacecadet
spacecadet Dork
11/27/19 1:21 p.m.
Vigo said:

So basically, if taking out and apart a differential doesn't sound like fun to you, don't run a clutch type in your racecar. My  .02.

ME ME!! THIS IS ME! 

I'll take my flawed helical to never have to touch it. 

GameboyRMH
GameboyRMH MegaDork
11/28/19 8:52 a.m.

In general I like 1-way operation for most uses on pavement. If your diff is reducing slip on coast or decel, that's going to make the car rotate less in a turn. There are some potential positives to this action but to me, the negatives outweigh it, and sometimes these positive uses amount to fixing a suspension problem with a drivetrain solution, which I don't like. 1.5-way or 2-way action might make sense for drifting, rally or offroading where the car spends a lot of time sideways or on slick surfaces.

As others have mentioned, a clutch-type diff is a PITA in terms of street comfort and maintenance, so unless you're really serious about taking trophies home I'd recommend a helical/Wavetrac.

Our Preferred Partners
hfnXkR4BAA88Yv6KhARZJ2uk2yslPYiYHiUMyR77ysHujKBd2cdHnDJ83HdpmFQt