FAQ: BMW eLSD - Not your dad's open diff!

[GaryS]

Originally Posted by int2str

I recently drove an older, open diff 528i again and the difference to even a stock 135i is shocking. I would find it hard to believe that the difference isn't obvious. With a completely open diff, the inside wheel lights up like a Christmas tree around pretty much every corner.

Hmm, not a comparison I can make. My 135i is the first car I've owned that didn't come with a mechanical LSD. Compared to them, my 135i would hang up terribly coming out of tight corners and sometimes driving over curbing and bumps. When that happened, it felt like an eternity waiting for the car to get going again. My Wavetrac completely fixed that. Again, I have no idea about the e-diff's advantages vs a regular open diff without the brakes intervening. I'm sure it helps something or BMW wouldn't waste money on it.

...adding: I mentioned BMW spending money on it, but thinking about it, I guess the e-diff doesn't really cost anything except sw development.

[FactorX81]

Originally Posted by GaryS

...adding: I mentioned BMW spending money on it, but thinking about it, I guess the e-diff doesn't really cost anything except sw development.

Never underestimate the cost of sw development and testing. In my experience it is usually quite a bit more than most people think.

[Diver]

Originally Posted by BrokenVert

It's not moaning when it's science.

What's your science, a butt dyno?

One of the marvels of the modding community is (except for the drag racers) there are never track times to back up the benefits of mods. People replace run flats then decide their run flat tuned suspension is too soft. Next the suspension gets replaced and finally they flip for a $3500 LSD. No lap times ever document the benefit of the changes. "It felt better." Even on the engine side I have a PM from someone who had a JB4, intake, down pipe, and exhaust. He said is car ran no faster than a friend's that had only a JB4.

Nobody actually races these modded cars, they just rip around in them. The mods put the car in a class where they are not competitive.

[BrokenVert]

^my science is my trade

Im an aircraft engineer through training, but I have designed racecars before (granted they were SAE cars while I was in school)

If you understand engineering principals - like I tried to explain on the last page - you'll understand why the eLSD is inferior to a mechanical LSD.

I also have much more experience with the tailout nature of the eLSD than most - because according to this thread that's all that matters apparently.

While the slide is more controlable than that of an open diff - its not up to par with the mechanical LSD cars that I have driven.

Also feel is just as important as setup. Feel brings confidence. And confidence brings speed. If it was just about setup why is Jensen button struggling to heat his front tires while using the exact same setup as his McLaren teammate who is much more competitive currently?

[int2str]

The intend of this thread was not to discuss the subtleties regarding ADB vs. LSD and which is better, but simply to clarify some common misconceptions.

To summarize:
- The 135i and other recent BMWs have an electronically assisted open differential that is significantly better than an open differential.

- Modern BMWs put power down to both wheels in most situation.

- No power is cut as a result of ADB.

- The electronic differential system works well for the vast majority of BMW owners and in most scenarios.

[GtiGyver]

Originally Posted by int2str

No. Only light brake pressure is applied to the spinning wheel this does not slow the car down. Mechanical limited slip differentials perform the exact same function, though instead of using the brakes to apply friction on the spinning wheel, mechanical friction is applied inside the differential. So the friction losses are similar.

This is incorrect. A true mechanical LSD has no losses in the sense you mention above.
Any friction inside a mechanical LSD is always between the two wheels, where the friction is always a ratio of one wheel to the other with the strongest case being a 1:1 lock.
With an eLSD, friction is used to slow down the spinning wheel against the frame of the car and transfer the remaining torque to the wheel that grips.

With an eLSD, the worse the difference in traction between the left and right wheels, the greater the loss in heat.

With a mechanical LSD, this would make very little difference by comparison.
Sure there will always be some heat generated by the gears/clutches in the diff, but this is negligible by comparison to wheel breaking torque biasing.

There is no question that eLSD is better than nothing, but by no way is it a replacement for a true hardware LSD, and this is the reason why M cars are still equipped with them. To BMW, an eLSD is a nearly costless addon that they can market as an alternative LSD.
The same goes with cruise control and traction control which in the day of drive-by-wire and ABS is another FREE addon for manufacturers.

[mike the snake]

My 135 when I gun it hard in 1st and 2nd (everything off) lurches from side to side when it spins the wheels, it goes anything but straight, and I am sure this is due to the eLSD.
Between the turbo power hitting so hard and unpredictably, and the eLSD causing the car to swerve under hard excelleration, I have all but stopped driving hard until I get a real LSD.
It's even worse in the rain.

To me, the eLSD isn't a performance thing, it's for super slow, tight turns like going up tight switchbacks where the inside wheel spins, or for mud or snow (again, slow traction conditions) because for performance driving, to me, the eLSD seems almost dangerous.

[int2str]

Originally Posted by GtiGyver

This is incorrect. A true mechanical LSD has no losses in the sense you mention above.

Why is this incorrect? Which LSD type are you referring to?
The two most popular LSDs for our cars (Quaife & Wavetrac) are gear type LSDs.

Here's an extract on how they work (extract from Wikipedia):
"As torque is applied to the gears they are pushed against the walls of the differential housing which creates friction."

To bias the torque, one output side has to be slowed down/restricted somehow. Whether you're talking internal clutches, gear types or in fact brakes, all three apply friction to achieve this.

With an eLSD, the worse the difference in traction between the left and right wheels, the greater the loss in heat.

With a mechanical LSD, this would make very little difference by comparison.
Sure there will always be some heat generated by the gears/clutches in the diff, but this is negligible by comparison to wheel breaking torque biasing.

In both cases friction is applied proportional to the input torque.
Btw. A lot of performance LSDs have cooling fins

There is no question that eLSD is better than nothing, but by no way is it a replacement for a true hardware LSD

That is subjective and highly dependent on the application. I'd still maintain that most forum members here couldn't tell the difference. Again this is not the point of this thread though.

[int2str]

Originally Posted by mike the snake

My 135 when I gun it hard in 1st and 2nd (everything off) lurches from side to side when it spins the wheels, it goes anything but straight, and I am sure this is due to the eLSD.

...

the eLSD causing the car to swerve under hard excelleration,

Absolutely incorrect.
If anything, ADB helps you stay straighter at that moment.

for performance driving, to me, the eLSD seems almost dangerous

*sigh*

This is the kind of FUD I hope to address with this FAQ.

[Red Bread]

Originally Posted by int2str

Again this is not the point of this thread though.

Oh, Obi Wan, could you remind us of what the point of this thread is again?

While you're at it, could you explain why a massively more advanced eLSD in the MacLaren MP4-12C has been nearly universally panned for handicapping an otherwise brilliant car?

[int2str]

Originally Posted by Red Bread

Oh, Obi Wan, could you remind us of what the point of this thread is again?

To clear up misconceptions and falsehoods spread about the adaptive differential braking on our cars and answer questions regarding the system.

While you're at it, could you explain why a massively more advanced eLSD in the MacLaren MP4-12C has been nearly universally panned for handicapping an otherwise brilliant car?

Because those journalists are still alive to write about it. The journalists who got to drive the open-diff McLaren MP412C-OD all got killed in horrific fiery crashes

[FactorX81]

Originally Posted by GtiGyver

To BMW, an eLSD is a nearly costless addon that they can market as an alternative LSD.
The same goes with cruise control and traction control which in the day of drive-by-wire and ABS is another FREE addon for manufacturers.

I'm not so sure this is the case, while not putting the lsd in the car saves "The Consumer" money by making the cars cost less I dont think it saves BMW any actual money. As far as Free addons the eLSD, I am sure, cost 10's of millions of dollars for initial developent and several million dollars on every model they adapt the software to as far as development and testing. I think the eLSD costs BMW more money than not putting it in and throwing in a mechanical LSD. However being integrated with the other systems DTC etc. it makes the car safer and easier to drive for your average lead footed consumer.

[Red Bread]

Originally Posted by int2str

Because those journalists are still alive to write about it. The journalists who got to drive the open-diff McLaren MP412C-OD all got killed in horrific fiery crashes

Haha, that actually would have probably been undrivable.

[GtiGyver]

Originally Posted by int2str

Why is this incorrect? Which LSD type are you referring to?
The two most popular LSDs for our cars (Quaife & Wavetrac) are gear type LSDs.

Here's an extract on how they work (extract from Wikipedia):
"As torque is applied to the gears they are pushed against the walls of the differential housing which creates friction."

To bias the torque, one output side has to be slowed down/restricted somehow. Whether you're talking internal clutches, gear types or in fact brakes, all three apply friction to achieve this.



In both cases friction is applied proportional to the input torque.
Btw. A lot of performance LSDs have cooling fins



That is subjective and highly dependent on the application. I'd still maintain that most forum members here couldn't tell the difference. Again this is not the point of this thread though.

The difference is that the mechanical diffs apply friction on the diff housing, the housing being the part attached to the driveshaft, and not the bellhousing (or pumpkin as some call it).
This means that it applies friction against the wheel with most grip and the driveshaft.
Its really difficult to put in words but you are making an incorrect assumption about the losses when comparing both types of LSD.

It is true that mechanical LSDs produce heat, however we are talking orders of magnitude lower than break power biasing on eLSDs.
The reason being that they designed to limit the difference between the speed between output shafts and not the frame of the car like an eLSD.

eg:

left wheel: 15rpm
right wheel: 40rpm
In this scenario, the right wheel is losing grip.
The delta: 25rpm inside the the LSD.

For an eLSD, it will need to slow the wheel directly (40rpm) and modulate the brakes until the delta is within limits.
For a hardware LSD, it only needs to account for 25rpm, and that while using the engine driveshaft as a reference, not the motionless frame of the car.

Hope this clears things up a little on this rather cloudy topic.

Whether a hardware LSD should be on all BMWs, or could be noticed by the average driver is another discussion altogether.

[GtiGyver]

Originally Posted by FactorX81

I'm not so sure this is the case, while not putting the lsd in the car saves "The Consumer" money by making the cars cost less I dont think it saves BMW any actual money. As far as Free addons the eLSD, I am sure, cost 10's of millions of dollars for initial developent and several million dollars on every model they adapt the software to as far as development and testing. I think the eLSD costs BMW more money than not putting it in and throwing in a mechanical LSD. However being integrated with the other systems DTC etc. it makes the car safer and easier to drive for your average lead footed consumer.

LSD hardware is far more expensive than a software solution so eLSD is definitely saving money for BMW, otherwise they would still install them in all their cars as they once did.
This is why nearly all car manufacturers today feature eLSD in their cars. VW already had it in the early 2000s.
Very little research needs to be involved, you are grossly overestimating the complexity of this kind of software system. A cheap easy solution to a known problem with open differentials.

[int2str]

Originally Posted by GtiGyver

Its really difficult to put in words but you are making an incorrect assumption about the losses when comparing both types of LSD.

I'm still not quite sure which part is incorrect.
If there's a mistake, I'd like to clarify it in the 1st post. So maybe you can help me clarify that point.

The reason being that they designed to limit the difference between the speed between output shafts and not the frame of the car like an eLSD.

I'm not seeing a difference here....

eg:

left wheel: 15rpm
right wheel: 40rpm
In this scenario, the right wheel is losing grip.
The delta: 25rpm inside the the LSD.

For an eLSD, it will need to slow the wheel directly (40rpm) and modulate the brakes until the delta is within limits.

I would argue against usage of the word "modulate" here, since the brakes aren't applied in a "binary" (all on/all off) fashion.

Also, applying some friction will slow the fast wheel down, but not to 25rpm. As friction is applied to the 40rpm wheel (using the brakes), the 25rpm wheel will speed up, since it now receives power from the driveshaft.

For a hardware LSD, it only needs to account for 25rpm, and that while using the engine driveshaft as a reference, not the motionless frame of the car.
Hope this clears things up a little on this rather cloudy topic.

No, sorry. Doesn't clear it up for me...
Inside the diff, the half-shaft is pressed against the left or right inside of the diff, causing friction on the output of that half-shaft (right wheel spins, right wheel output half-shaft friction increases).

I'm still not sure what the difference is between applying friction on the inside end of the half shaft (by applying friction to the diff housing; attached to the "motionless frame of the car", btw.) and applying friction to the outside end of the half shaft (by applying the brakes).

Not trying to be dense here, but I don't see the difference just yet.

[BrokenVert]

Originally Posted by int2str

I'm still not quite sure which part is incorrect.
If there's a mistake, I'd like to clarify it in the 1st post. So maybe you can help me clarify that point.

I'm not seeing a difference here...

I would argue against usage of the word "modulate" here, since the brakes aren't applied in a "binary" (all on/all off) fashion.

Also, applying some friction will slow the fast wheel down, but not to 25rpm. As friction is applied to the 40rpm wheel (using the brakes), the 25rpm wheel will speed up, since it now receives power from the driveshaft.

No, sorry. Doesn't clear it up for me...
Inside the diff, the half-shaft is pressed against the left or right inside of the diff, causing friction on the output of that half-shaft (right wheel spins, right wheel output half-shaft friction increases).

I'm still not sure what the difference is between applying friction on the inside end of the half shaft (by applying friction to the diff housing; attached to the "motionless frame of the car", btw.) and applying friction to the outside end of the half shaft (by applying the brakes).

Not trying to be dense here, but I don't see the difference just yet.

Its very hard to understand unless youre an engineer.


Mechanical systems are very confusing and false assumptions (such as yours) are easy to make. Its not a problem, but dont fight the people that actually know what theyre talking about when youre making the wrong assumptions.

[int2str]

Originally Posted by BrokenVert

Its very hard to understand unless youre an engineer.

Wow, not condescending at all there

Mechanical systems are very confusing and false assumptions (such as yours) are easy to make. Its not a problem, but dont fight the people that actually know what theyre talking about when youre making the wrong assumptions.

WHICH assumption is false.
What makes you think I'm fighting anything?

[BrokenVert]

Originally Posted by int2str

Wow, not condescending at all there

WHICH assumption is false.
What makes you think I'm fighting anything?

Not condescending at all. Its complex stuff and unless you have the training that we have its very difficult to grasp.

Nobody would expect me to understand anything about Orgo - but thats second year stuff for doctors - premed students that is.


The assumption that applying the brakes and friction on a shaft yield indiscernible results - thats the false assumption.

[int2str]

Originally Posted by BrokenVert

The assumption that applying the brakes and friction on a shaft yield indiscernible results - thats the false assumption.

I wouldn't argue "indiscernible", but certainly similar (which is the word I use in the first post).
Please explain how this is false. Thanks.

[GtiGyver]

I'm usually very good at analogies, but one eludes me on this topic.

We agree than a hardware LSD works completely independently of the frame of the car. In fact, imagine only the driveshaft, the LSD itself without the cast iron housing and the half-shafts to the wheels.
Now, in order to equalize the speed of both halfshafts, the LSD will internally increase friction, potentially all the way to locking the shafts together.
Any friction here is applied to the system that is already in rotation.
Minimal losses.

For an eLSD, as you know, the brakes need to be applied. How well an eLSD can do this is highly debatable since the wheel speed needs to be measure via the ABS wheel speed sensors, compared with the other wheels, then brake pressure be applied, which is far from instantaneous.
Too much brake pressure could result in the wheel that was slipping to completely lock up, thus transferring in this specific case all the power to the other side causing it to lose grip. This is why the eLSD has a very limited range.

Now back to the initial explanation, we agree that the LSD itself in the eLSD does no work. The work here is strictly performed by the brakes on the wheel with least grip.
Since the diff is open, we agree that the total power on both wheels must equal the input, and this is why locking one wheel causes the other to spin twice as fast
.
In a wheelspin scenario, the brakes will attempt to simulate the same friction on the slipping wheel as the one that grips reducing the differential.
This friction on the brakes is a pure loss, since it will only result in heat, but in no way will propel the car forward, and thus a loss of power.
Since this occurs in limited grip scenarios, the driver will never feel it. The car is already losing power by roasting up its tires.

[BrokenVert]

Originally Posted by int2str

I wouldn't argue "indiscernible", but certainly similar (which is the word I use in the first post).
Please explain how this is false. Thanks.

I did explain it already on the first page and GTIGyver added even more to my initial explanation.


It has to do with moment arms and frames of reference.