The ultimate differential thread

to be fair it doesn't seem the factory provides this as an individual p/n but you can certainly get them custom made...whether this can be considerered "serviceable" is debatable i guess

 

Ok so it seems that I can't edit the original post but maybe a mod can help me do it later.

I wanted to go over how Helical LSDs work (known commonly as Torsen but that's just a brand name, like Kleenex is a brand name of tissue paper, Quaife ATB is another brand/model of helical LSD but they operate in the same way). I'm going to start off right away and say worm gears/wheels are NOT involved in any way. Worm gear/wheel pairs are general ratio'd in excess of 20 to 1, you will never find any gear pair in a differential approaching anywhere near this, it's also fundamentally not how helical diffs operate (which you'll learn after the giant wall of text below).

Below is a crude depiction of a sample helical LSD

Image Unavailable, Please Login
What's not shown is how the differential carrier (the thing that connects to the drive side of the differential assembly that first "sees" the torque/power from the engine/transmission) which connects to the green and blue gears simply via the channels the blue/green gears sit in (imagine the differential carrier is the hotdog bun and the green and blue gears are the sausage, if you hold the bun with the sausage pointing directly left/right and you push the bun out away from you, the sausage goes with it, but the sausage is free to spin inside the bun).

The yellow and blue gears are not connected to one another directly
The red and green gears are not connected to one another directly
The yellow and green gears ARE connected directly
the red and blue gears ARE connected directly

Gears that are connected directly must rotate in opposite directions (one clockwise, one counter clock wise, though in this perspective you can think of it as "rolling forward" or "rolling backwards")

So from this, we know the following

1)Yellow and Green are opposite

2)Green and Blue are Opposite

3)Blue and Red are opposite

4)Therefore Yellow and Red are opposite (Y->G->B->R, 1-0-1-0, proof that differential action is achieved via this gear arrangement)

When the differential carrier is spun when both left/right outputs have equal traction (or when the driveline torque is sufficiently low that it does not overwhelm the tractive force of the tires), the green and blue gears do not rotate on their own axis, the yellow/red gears do not rotate relative to one another and the entire arrangement rotates at the same rate as the differential carrier. Power/torque is distributed to the axles more or less evenly just like in an open differential.

When differential action is required (navigating a curve) let's say the yellow side needs to be slower, red side faster, the gear arrangement allows for this.

What makes this setup "interesting" is that all the gears here are helical cut, they are not straight spur gears (if they were, the differential would still work, but would provide no limited slip functionality).

When force is applied to a spur gear (or it is applying force to another spur gear) it generates no significant thrust/side load. For a helical gear it's quite obvious that there IS side load axially due to the angle of the helically cut teeth (how much side load depends on this angle). This side load pushes the red/yellow/green/blue gears against the differential carrier itself which creates friction. The more load is applied, the harder these side gears want to "drill into" the side of the differential carrier, the more friction there is. The friction between these surfaces "tries" to resist the relative movement between the yellow/green/blue/red gears (limiting relative motion between the red/yellow outputs, limiting slip).

Some helical LSDs rely solely on this metal on metal contact between the faces of the red/yellow/blue/green gears against the differential case to generate the friction for limiting slip. Metal on metal does create friction but these are well machined surfaces so while it's got more internal friction than an open diff (which has basically very little to none) it's not going to generate as much internal friction as a multi-plate clutch type lsd. What some manufacturers do for some applications is introduce some friction material/plates (tiny clutches basically) between the end faces of the gears and the differential casing, giving the side load forces a bit more "bite" and helping the differential bind together when needed. Clutches are wear items and OEMs don't like wear items when they can avoid it (even at the expense of performance) so you generally will not see these types in OEM applications though it's not unheard of. See the youtube videos below detailing the disassembly of an helical LSD from an Integra R and NA/NB MX5, the integra R helical diff doesn't have additional friction material, the MX5 does (which is odd because the Integra R made a lot more power).

Integra R helical Diff Video from "EricTheCarGuy"


MX5 T-2 (Torsen Type 2) Diff from "Cars on The Road"


Ultimately ALL limited slip differentials rely on friction to bind the two output shafts together. In a traditional clutch pack type LSD it relies on one or more clutch disks connecting the output shafts to the housing, in "active" diffs they rely on the same thing but the actuation is handled electrically or hydraulically as opposed to mechanically, in a helical diff it relies on friction due to the side load generated when helical gears are asked to transmit load.

What seperates the performance between the "dumb" LSDs (active diffs are going to introduce another element of how the software is programmed, well outside the scope of this giant wall-o-text) are the ultimate limits of how much friction the internals can generate and the "profile" of engagement between the preload/base level of friction and the maximum (as well as any/if-any difference in profiles between positive vs negative engine torque).

It's quite easy to see that there's basically no limit to how many clutch disks you can add in a multiplate LSD setup compared to a helical LSD, so in this sense maybe you can make the argument that a multiplate LSD is "superior" but ultimately for most people for most applications, either design can achieve the purpose and it's simply a case of what's available for your platform.






Additional explanation courtesy of an Acura Integra 98-01 service manual.
Image Unavailable, Please Login
Image Unavailable, Please Login

Image Unavailable, Please Login

 

So our boy Dan decided to triple down on the diff misinformation on his new vid (Mods - I can allude to this right ? please don't take down this thread, I spent a lot of time on gathering the technical info)

let's see how fast he (Dan) takes down this comment

Image Unavailable, Please Login

 

Lol, that was fast

 

my open diff 5 series will spin both tires in a straight line too

open diffs still have enough internal friction to bias something like 5 percent torque from one side to the other (~55/45 source is from here https://www.fsae.com/forums/archive/index.php/t-3424.html?s=5ce42bd67e1ff12ad0ef134603742939 and https://www.musclecardiy.com/performance/selecting-aftermarket-differentials-to-improve-performance/ and pretty much common knowledge in engineering circles) meaning if you don't have a bigger grip difference between the left/right wheel in excess of this amount, itll light up both tires just fine

the ediff disable (just purely software) will still leave the clutch packs in the unit with the proload provided by the thrustwashers/springs so it will probably give a bit more bias than a basic open diff (the thrustwashers/springs is there to takeup backlash but won't provide progressive lockup)

There's two ways to look at "evaluate" the amount of "slip limiting" an LSD has.

One is torque bias ratio which is (how much more torque can the differential send to one side vs the other in a relative sense).

The other is how much more torque in actual torque units (ft-lbs for example) can it send to one side than the other.

Neither are wholly appropriate because they operate on different mechanisms.

The amount of binding force internally (which is ultimately what governs the ability to bias torque) operate on two different mechanisms.

You have preload, which you almost always have a bit of, so if the internal clutches already have some pressure applied to it, then even at "idle" the differential has the ability to bias torque by the amount of the preload.

Then you have the angles of the ramps on the pressure ring which provides progressive lock

The more torque the differential receives from the engine, the more those spider gear pinions push on the carrier which push on the clutches, the angle here determines how much TBR the diff will have, it can be weak (in conical type single plate LSDs like the tochigi fuji diff in the MX5 which are barely better than an open diff it's like somewhere between 1.5 and 2.0 to 1) or it can be very strong (a multiplate LSD on a drift car will easily lock up both axles under load)

what you can see in the real world experiments is one thing, but when there is actual physical proof to the contrary (the 430 ediff teardown photos/video here https://www.carpassion.com/forum/thema/60887-funktion-einer-steuerbaren-differentialsperre/ also posted below as attachments) and you're still relying on dumb things like (lift the car and spin the tire or whether it can do a burnout) then you know you're just being stubborn

Image Unavailable, Please Login



Image Unavailable, Please Login Image Unavailable, Please Login Image Unavailable, Please Login Image Unavailable, Please Login Image Unavailable, Please Login Image Unavailable, Please Login

 

I gave up trying to respond via comments to his videos, he's just deleting anything I post

Such is the world nowadays I guess.

If I'm actually wrong, I'm super welcome to have someone point out to me exactly what the mechanism the E-diff sans hydrualics uses to enable progressive binding of the axles (which is actually binding either or both axles to the differential housing). Preload on the clutch pack doesn't count, if preload was sufficient for proper driving, what the hell is everything else for ?



Again, if you want to delete the E-diff because it's a finnicky system and it's locking up TOO much such that it's detrimental because of messed up sensors or it being overly sensitive to tire sizing, etc. .... Go ahead, I'm not judging. But don't spread lies about how it's better now than it was when it was fully functional (which is what Dan is saying). It can be better than your busted diff, that's great, but that's not the same thing as saying it is better than what the factory gave you. I'm sure a junkyard LS is more reliable than a Ferrari v8 with a busted rod ... but it's not better than it would be if it was working in tip top shape.

 

That's literally what the E-diff delete is, completely blocking off the hydraulics from ever actuating (you can do it via software OR you can completely eliminate the F1 pump and all associated lines/accesories which obviously eliminates hydraulic actuation)

The E-diff is literally an open diff with a clutch pack on one side that's entirely hydraulically controlled (hydraulically actuated, electronically controlled if you want to be pedantic)

You can see it from the teardown and from this parts/schematic (https://www.eurospares.co.uk/Ferrari/430/F430_Spider_(RHD)/PartDiagrams/0034/DIFFERENTIAL_AND_AXLE_SHAFT)

Image Unavailable, Please Login

If you squeeze the clutch pack together, the left hand side axle and differential housing will bind (thus binding the right hand side axle as well)


This is the diff from a 360 from @67mer's post here https://www.ferrarichat.com/forum/threads/what-to-do-about-360-lsd-limited-slip-diff-low-slip-torque.646304/#post-148046230






If you can find the pressure ring in these teardown photos (https://www.carpassion.com/forum/thema/60887-funktion-einer-steuerbaren-differentialsperre/) I'll give you a million dollars.

Clutch type mechanical LSDs aren't rocket science, they're actually incredibly simple but effective devices. That's what makes this BS with Dan so annoying. Like fk man, either point out how limited slip bias is achieved or STFU. All this talk about putting a car up in the air and spinning wheels and doing a pseudo burnout (a launch isn't a burnout, put one wheel on grass and the other on asphault and try it again) is like so pre-internet when no one knew dick and everyone was an idiot. We actually HAVE the actual knowledge/data/technical diagrams now available at our fingertips, don't rely on rules of thumb which have ALWAYS been problematic.

I'm going to repeat again, ALL limited slip differentials operate based on friction. That friction between the output (two of em, either front/rear or left/right) and input (either from the center diff or engine depending on whether you're 2wd or 4wd) is what limits slip, that's literally by definition. Differential action is slip, limiting differential action (binding the output with the input) is limiting slip.

*You can get that friction from preload (downside is this never changes, thrust washers will ALWAYS apply the same amount of force whether you're free-wheeling or doing a launch control start)
*By mechanical design (angles in the pressure plate/angles of the helical gears in a helical LSD, angles of the bevel gears in a connical LSD/open diff)
*Via some form of active control (electronically/hydraulically actuated clutch packs i.e E-diff)
*Viscous coupling (whether you think viscosity is "friction" is in my eyes just an exercise of how pedantic you want to be)

The ONLY reason I can think of why Dan and NGS are pushing this narative so hard is because they don't want to tell their potential clientel "yeah if you take your car to our shop to do the manual conversion, we're strongly going to push you to do an E-Diff delete because it makes our lives easier but I don't want to tell you straight up itll make your car worse"

Again, I welcome anyone to point out exactly where in the E-diff sans hydraulics is the component/linkage/etc. which achieves progressive loading of the clutch pack. If ANYONE points this out, I'll eat my fkin hat.
Image Unavailable, Please Login View attachment 3614246

 

I still think a mechanical LSD is a good option over the E-diff

The thing about it being overly sensitive to tire diameter is a bit extreme. It's not hard to have mismatch in excess of 3% between the axles once you deviate from OEM sizing

 

Good to know, I'm thinking of getting a 430 as a project car, good to know this is an option

 

Yes and it's easy to get a mismatch which exceeds the threshold in software once you start to deviate from OEM sizing, if you keep this in mind it's also easy to stay close to this too so make of it what you will

 

I'm the OP of the thread but not the member you cited here

I largely agree with you with respect to th E-diff delete, I think there are legit reasons to software delete this functionality AND also bad reasons too. While there might be SOME probative value in correlating E-diff delete to "lack of maintenance", it's certainly a lot easier to just simply look at the rest of the car and see if it's commensurate with a vehicle is which is poorly/immaculately maintained.

If you want a mint condition F430 that would be concours ready in 20 years ... then sure, you really have to have OEM everything even if it is objectively worse but that simply isn't something I personally care too much about.

 

Also worth noting that the perceived "improvement" in handling of the delete IS real

Open diffs DO handle better than LSDs .... because the purpose of an LSD isn't to inherently (just by being there) improve nimbleness, their purpose is to improve the ability to put power down earlier out of corners.

If you're NOT applying throttle, you want the diff to be open and both axles to rotate as freely as they wish. If you ARE applying throttle, then potentially you get some improvement in differential torque SOMETIEMS (if one wheel has grip and the other is on ice or you're turning hard enough to get significant load transfer, you'll get a yaw moment from the thrust being off center vs the CoG of the vehicle) .... but this is certainly the exception and not the rule. You need to be on throttle, near grip limit, significant load transfer side to side, and measuring your performance with a stop watch to really "regularly" see the benefit of an LSD. In short, you need to be pushing hard enough for an LSD to be "better" than an open, and it isn't simply a case of pulling enough lateral G, it's a case of putting enough lateral G AND you getting on the go-fast-pedal earlier than you would otherwise be able to in an open-diff'd car.

It's not surprising that users who just drive it casually on public roads would see a "benefit" from "opening up" the diff if they had a malfunctioning unit. That's exactly what the E-diff was SUPPOSED to do .... no hydraulic pressure with minimal bind between left/rear axles when you don't need the axles locked, and progressively binding them when you DO want them locked (as determined by sensors + software + manatino mode)

It's the same reason why cars with mechanical LSDs are harder to park (those who remember driving cars with LSDs AND w/o power steering can testify to this), the preload partially binds the left and rear wheels together making turning difficult.

 

All good points here

A mechanical LSD isn't a magic speed button, it's a very niche device which gives you some benefits sometimes, and some detriment sometimes. Good/keen drivers using it for a specific purpose in a certain way would be able to make the trade off "make sense" but for many others it's actually more downside than upside.

The E-diff was designed to solve this problem. It was supposed to be as close to an open diff as you reasonably can be when limiting slip was not neccesary, but to be able to do it on command when needed and to be able to get around the limits of what a mechanical LSD has in terms of setting up ONE specific "lockup profile", electro-hydraulic control effectively frees up the entire range of lockup vs applied torque universe to be used whereas a mechanical LSD limits you to a tiny subset of this.