Variable M differential lock

[scotth944]

Does anyone know how the variable M diff used in the 1M is capable of achieving a 100% lock condition which the BMW documentation claims it is capable of if it is dependent of the slip between the shear plates to create hydraulic pumping action it uses to engage the locking in the first place (at lease as far as I understand how it works) or is there some other mechanisms utilized which they do not mention? I would think 100% lock would mean no more slip is in progress between the shear plates.

[Pig Farmer]

Does this answer your question...

Sorry, I couldn't resist.

[scotth944]

I always did wonder how rainbows worked since I never did find the gold that I thought made it??

[scotth944]

Between trying to figure out the meaning of life and how the 1M diff works at 3 in the morning, I actually have bigger picture I have been sort out in my head. I always wondered why the 1M was prone to snap over steer given that I had a 944T setup a built which had many setup properties of the 1M including power to weight ration, 50/50 weight balance, turbo, tire sizing to weight.... and it never had an issue with snapping on me. If it did break loose, it was very controllable and predictable. Some on the forum have talked with me about the PS2 tires, short wheel base, spring/shock setup being the reason which makes since but this does not seem to get entirely resolved when changes have been made by various forum members based on all the threads I have gone through.

Based on what I have experienced and what others have reported, I have a theory and reasons I would like to through out for comment on which might explain the issue- the M-variable diff:
1- From what I can tell, the snap primarily happens when powering out/through a turn.
2- When inside rear tire breaks loose, the drive train back to the engine abruptly increases in RPM gaining more momentum. The variable diff quickly transitions from 0 to 100% lock which now overloads the outside tire with a RPM "spike" since the rest of the drive train had ramped up causing it to blow past it's adhesion capabilities that were already near the limits due to g-forces of the car it was holding.
3- The M3 is not as prone to the issue due to the increased drive train dampening due to more mass. car also has longer wheel base which also slows rotation of the M3 chassis.
4- I think there is also a great amount of rubber isolation in the rear end setup which would dampen the amount of feedback you can get from the rear tires meaning you will have more difficulty in catching the problem before it snowballs on you.
5- The factors I mentioned at the beginning such as tires, suspension help make the snowball effect more pronounced when it occurs.

Does this make since to anyone else or do I need to consider a different hobby such as collecting sea shells?

[Artemis]

Regarding the M differential lock on the E82 1M (E9X M3), have a look at the following docs straight from the horse's mouth:

• section 4.3. at pages 19-22 of the following document (1MB):
  http://www.1addicts.com/forums/attac...2&d=1396203865
• pages 8 and 23 of the following document (0.5MB):
  http://www.1addicts.com/forums/attac...4&d=1329761227

Both these docs are included in the 1M manuals sticky:
http://www.1addicts.com/forums/showp...82&postcount=1

[scotth944]

Thanks @Artemis , the graph in the first doc to me explains how the "pump function" of the diff would behave but it does not explain how you can achieve 100% lock up between left/right tires (if that is what they mean when they say 100% lock). The graph shows 0 Nm when there is 0 speed differential which you would have if the diff has achieved 100% lockup. I wonder if this is a marketing misrepresentation or is there another mechanism in play which they do not cover in their explanations??

I can tell that there is a slight lock up delay in my car when pulling out of a parking lot which the graph does indicate would happen. I think this delay can help contribute to rear end break away but I would like to think it is very quick at high road speed conditions?

[Pete_vB]

I agree the M diff can be abrupt, and could be a contributing factor, but my money is still on shock valving being the main culprit, followed by tires.

An experience last week reinforced this thinking.
I was driving a friend's car on radial slicks, 18" Avons. For those that haven't tried them grip and precision are immense, but grip builds very quickly and then falls off just as fast on the other side. Thus they are far "snappier" than a PS2 ever is, even when it's worn, though still very drive-able once you re-calibrate. You do need to be very quick with the wheel, however, like the 1M.

We then played with sway bars, which adjust mid-corner balance. While it was easy to make the car oversteer, that too is fairly slow and predictable, with the back end simply coming around sooner rather than surprising you.

Finally we adjusted shocks. This is where the car really got snappy: by tuning one end separate from the other it was very easy to make the car undrivable with snap over-steer on both corner entry and power on at the exit. This was on a flat course with no bumps, and my previous experience is that bumps magnify the effect. As I've said previously the 1M shock dynos suggest BMW was tuning to improve turn-in, which is exactly what we were doing last week, but this had the unintended effect of making the car too quick to rotate on corner exit particularly if the tires are less than ideal.

In my experience the M diff comes into play most abruptly on tight, slower corners. At higher speeds I notice it less, and my gut feel is that it comes in more progressively. The control unit is supposed to "anticipate" wheel spin, presumably using throttle position and Gs. I would suspect that you could either disconnect it or have someone program it out for a test. You might lose some stability on corner entry in the process?

Originally Posted by scotth944

Between trying to figure out the meaning of life and how the 1M diff works at 3 in the morning, I actually have bigger picture I have been sort out in my head. I always wondered why the 1M was prone to snap over steer given that I had a 944T setup a built which had many setup properties of the 1M including power to weight ration, 50/50 weight balance, turbo, tire sizing to weight.... and it never had an issue with snapping on me. If it did break loose, it was very controllable and predictable. Some on the forum have talked with me about the PS2 tires, short wheel base, spring/shock setup being the reason which makes since but this does not seem to get entirely resolved when changes have been made by various forum members based on all the threads I have gone through.

Based on what I have experienced and what others have reported, I have a theory and reasons I would like to through out for comment on which might explain the issue- the M-variable diff:
1- From what I can tell, the snap primarily happens when powering out/through a turn.
2- When inside rear tire breaks loose, the drive train back to the engine abruptly increases in RPM gaining more momentum. The variable diff quickly transitions from 0 to 100% lock which now overloads the outside tire with a RPM "spike" since the rest of the drive train had ramped up causing it to blow past it's adhesion capabilities that were already near the limits due to g-forces of the car it was holding.
3- The M3 is not as prone to the issue due to the increased drive train dampening due to more mass. car also has longer wheel base which also slows rotation of the M3 chassis.
4- I think there is also a great amount of rubber isolation in the rear end setup which would dampen the amount of feedback you can get from the rear tires meaning you will have more difficulty in catching the problem before it snowballs on you.
5- The factors I mentioned at the beginning such as tires, suspension help make the snowball effect more pronounced when it occurs.

Does this make since to anyone else or do I need to consider a different hobby such as collecting sea shells?

[eeghie]

Originally Posted by Pete_vB

... We then played with sway bars, which adjust mid-corner balance. While it was easy to make the car oversteer, that too is fairly slow and predictable, with the back end simply coming around sooner rather than surprising you. ...

Can you please share what (generally available) sway bar solution(s) you'd suggest for improved 'predictable coming around sooner' behavior?

[scotth944]

Thanks Pete.
So if you were to put a stock setup mod list together for a streetable DE setup by order of priority, would it be the following:
1- tires such as RE-11, R-S3, AD08...
2- camber increase, 2-3 degrees
3- shock/struts such as Ohlins R&T
4- sway bar... less rear or more front, to be based on tire size change??
5- suspension bushing upgrades for camber link??
6- diff swap to 2-3 disc, 40% lock??

[Pete_vB]

Originally Posted by scotth944

Thanks Pete.
So if you were to put a stock setup mod list together for a streetable DE setup by order of priority, would it be the following:
1- tires such as RE-11, R-S3, AD08...
2- camber increase, 2-3 degrees
3- shock/struts such as Ohlins R&T
4- sway bar... less rear or more front, to be based on tire size change??
5- suspension bushing upgrades for camber link??
6- diff swap to 2-3 disc, 40% lock??

That sounds like a pretty good list, with the caveat that the 1M is not my track car, so I have only done #1 and #2.

For DE I'd probably do R-S3s over RE-11s or AD08s; they like heat and have a progressive breakaway. However I'm really hoping the correct size Pilot Sport Cup 2s come out before I burn through my latest set of PSS- wouldn't be surprised if those are even better.

I might do adjustable swaybars before adjustable shocks, just because they are cheaper, easier to tune and directly address most balance issues.

I'm not sure what I'd do to the diff, if anything, until I'd tried #1 through 4. I do understand they wear out pretty quickly, so it's possible that mine is no longer working as its supposed to anyway.

As for the swaybar, I should note my friend's car is a Porsche. I'm not certain which swaybars I'd get for the 1M, but I've noted these before, they seem to check most of the boxes:
http://www.rdsport.com/store/e9x-m3-sway-bars.html

[scotth944]

The specs on the bars say they are 13-33% stiffer, I take it you would leave the rear full soft an set the front as needed to balance the mid corner handling? Would the stiffer sway bar setup be used with softer springs if possible?

It would be nice to have a compiled list of car setup details that are being used with notes on how they behave compared to stock so we can track a trend on what might fix an issue we are having and to assist in planning a project build. I know Neil should be just about ready to weigh in on this!

[Pete_vB]

Originally Posted by scotth944

The specs on the bars say they are 13-33% stiffer, I take it you would leave the rear full soft an set the front as needed to balance the mid corner handling? Would the stiffer sway bar setup be used with softer springs if possible?

I'd probably start from the middle setting on the swaybars and then tune for balance- looks to me like the kinematics are better there than at full soft.

I don't think you'd need to go softer on springs, in fact for track work slightly stiffer would probably be fine in the dry.

We didn't mention alignment, but you might try a little more toe-in at the rear, less toe up front.

Have you ever tuned adjustable shocks before? Can be a little tricky...

[scotth944]

As far as the bar adjustments are concerned, I have not figured out what determines when to loosening up one end vs stiffening up the other end, do you have a guide to determine which end to start with first?

I was wondering about using lower spring rates to help with maintaining traction on irregular road surfaces which I would think is a bigger issue for stability concerns.

I considered alignment as part of step 2, I was thinking 0 toe in on the front and +2 to +3 degrees in the rear, caster would be maxed out up front but not sure what to do with the rear on this.

Adjustable shocks are another element I have not had the opportunity to work with yet so just like the sway bars, knowing which end you start with for a specific issue will be another opportunity for me to learn. I know the text book part of how they work but real world application is not always the same thing.

[Pete_vB]

Originally Posted by scotth944

As far as the bar adjustments are concerned, I have not figured out what determines when to loosening up one end vs stiffening up the other end, do you have a guide to determine which end to start with first?

Stiffen or soften based on grip. If grip is great, with a sticky surface and tires, the car will wallow too much and feel like it reacts slowly. If grip is low you need to soften or the car will feel like it's skittering and skipping over the surface. A 30% bump in stiffness isn't all that much- for real track cars with track tires manufactures will often bump spring rates by 5x or more.

Originally Posted by scotth944

Adjustable shocks are another element I have not had the opportunity to work with yet so just like the sway bars, knowing which end you start with for a specific issue will be another opportunity for me to learn. I know the text book part of how they work but real world application is not always the same thing.

For shock tuning, I would make sure you get shock dyno sweeps so you know how clicks map to forces. I would then start with the shocks set proportionally equal to stock (stock shock dynos have been posted here), or stiffer in proportion to any increase in spring rate.

From there you are doing two things, balancing body control vs grip and tuning for turn-in. Grip is similar to spring tuning above- too stiff and the car gets skittish and looses grip. There are a few tuning procedures to use, such as this one:
http://www.koniracing.com/rrtuningguide.cfm
I recommend intentionally going past the ideal values by a few clicks so you know what too stiff and too soft feel like and roughly where the line is.
With the shocks roughly in the right range for grip you then tune for balance, which is the critical thing in terms of stopping the rear stepping out at the wrong time. The link below is the method I use to do that:
http://www.ozebiz.com.au/racetech/th...hocktune1.html
Think about where you're feeling the back end step- phase 4?

Make sure you do things in order: tune the springs/ swaybars first with the shocks set relatively soft, then tune the shocks for grip, and finally tune the shocks for entry and exit balance. And I might do the tuning at an autocross with lots of runs first, rather than trying to set it up initially at the big track. It's both easier to make adjustments and to push the car past the limits safely.

[scotth944]

Thanks for the links, very informative.

Did you have any opinion on the amount of chassis feedback you are getting from the rear end and still keeping is streetable, have you contemplated doing any of the bushing upgrades for your 1M setup that we have discussed for other serious track cars?

[Pete_vB]

I haven't and won't do bushing upgrades- the 1M is mainly my street car and will stay that way.

If the 1M was my track car the main thing that would push me towards rear bushing upgrades would be tires and/ or stiffer springs. Stickier tires both increase loads on the bushings, resulting in more toe steer, and simultaneously often operate at lower slip angles, magnifying the effect of that toe steer. If I was running radial slicks, Hoosiers or significantly wider than stock tire sizes I'd think rear bushing upgrades would be highly desirable. With lower grip or tires with higher slip angles tires they become less necessary.

Each tire type has difference characteristics, so I'd decide on bushings based on feel. I generally gravitate towards tires with wide, predictable slip angles on street cars partly for this reason- Rivals, RS-3s, RA-1s, etc.

[scotth944]

I was thinking in terms of connecting the driver to the chassis with the question, I find the feeling I sense as the rear tires as they transition to breaking loose to be very muted in comparison to other cars I've driven. maybe a subframe bushing upgrade would get me what I'm wanting in terms of feedback?
in the past, I have been known to weld up the rubber isolation used in the 944 tie rods to get more feedback from the car and it did not bother me as a daily driver.

[Pete_vB]

Originally Posted by scotth944

I was thinking in terms of connecting the driver to the chassis with the question, I find the feeling I sense as the rear tires as they transition to breaking loose to be very muted in comparison to other cars I've driven. maybe a subframe bushing upgrade would get me what I'm wanting in terms of feedback?
in the past, I have been known to weld up the rubber isolation used in the 944 tie rods to get more feedback from the car and it did not bother me as a daily driver.

I'd need to defer to others who have done it on that. Certainly on other cars, ie 997 vs 997 GT3 with aluminum rear subframe bushings, you do get a slightly better feel for what the rear tires are doing as you take out rubber, however on the 1M I can't say I find the transition particularly muted, just occasionally snappy.

Doing the rear swaybar is quite involved, so if you're in there you might look at doing some bushings at the same time.

[Dackelone]

@scotth944 I asked a friend of mine who works for BMW AG about this subject. He also owns a 1M. He think its the design of the limited slip differential. It is true that the 1M is capable of achieving a 100% lock condition - but that 100% is only in theory. And when it "locks up" he told me that's when the rear axle (will)step out. He gave me this article on lsd types - it's in German, but you could use Google to translate it.


http://www.limitedslip.de/varianten.html

My friend also said that IF you were to take a 1M to track events or racing... he suggested to replace the 1M's LSD with another one that doesn't lock up so abruptly.


I also remember seeing an BMW AG video from a few years back... they tested various LSD types installed in various BMW ///M(e39, e46, e92, e60 and F10) models. I tried to find it, but it is eluding me right now. Maybe someone else remembers this video.

[Pete_vB]

It would be interesting to test a different diff certainly. Do you know if the M diff can be easily disabled somehow?

For me what's pushing my strong suspicion in the direction of shock valving is the 1M's behavior around constant high speed sweepers over bumps.

The 944T slides as one piece, front and rear, when it hits a mid-corner bump at speed, sometimes moving sideways by a foot without getting unstable or unsettled or needing to change steering angle. The 1M in the same corner is quite different: the nose tucks slightly, the car gets slightly unstable, and a steering correction is required. In neither case does the diff or power come in to play. I indicates miss-matched wheel frequencies/ shock valving to me. No chance I can take a sweeper in the 1M as quickly as in an equally tired 944T.

I'm not sure if Scott notices something similar, or if the main issue he's feeling is in a different part of the corner...

[Dackelone]

Originally Posted by Pete_vB

It would be interesting to test a different diff certainly. Do you know if the M diff can be easily disabled somehow?

For me what's pushing my strong suspicion in the direction of shock valving is the 1M's behavior around constant high speed sweepers over bumps.

The 944T slides as one piece, front and rear, when it hits a mid-corner bump at speed, sometimes moving sideways by a foot without getting unstable or unsettled or needing to change steering angle. The 1M in the same corner is quite different: the nose tucks slightly, the car gets slightly unstable, and a steering correction is required. In neither case does the diff or power come in to play. I indicates miss-matched wheel frequencies/ shock valving to me. No chance I can take a sweeper in the 1M as quickly as in an equally tired 944T.

I'm not sure if Scott notices something similar, or if the main issue he's feeling is in a different part of the corner...

I used to own an '86 944Turbo with stage II APE chips: ~285 Hp. That's how I met Scott btw many moons ago. Anyways... a few years after owing my 944T... 2nd gear broke on the input shaft inside the trans. (that was one part Porsche beefed up on the TurboS's. That's when I had AutoThority(Scott) retrofit a LSD into my 944T. The handling changed more than I thought it would post LSD install.


Before when I tried powering out of a tight turn... the inside wheel would just (lift and)spin under power. With the LSD... what I noticed - was more (initial)understeer when I would start to turn the car into a turn... then, when I would power out of that same corner... the rear end would ever so slightly come out a foot or two. Nothing dramatic. The rear end would just slightly rotate the car out of the turn. I really wasn't expecting a LSD to change the handling traits of a car, but it did on my 944T. I think in the 1M is much more dramatic event.

[scotth944]

Thanks for the research David.
The issue has been produced under various speed and corner conditions which I believe an elevation change on the inside wheel came into play to some degree near or after the apex. Based on what Pete has weighed in with on this subject, my gut feeling is the problem gets initiated by the shock setup which then leads to the LSD kicking in abruptly unless you keep some of the nanny control on to detect the sudden RPM increase. From my experiences I think the higher the speed, the less it takes to snowball out of control which obviously could be my lack of recent track time but I have had plenty of time some years ago which I am using to base opinion and expectations on. I have also seen the issue under what I would consider reasonable street driving conditions when the tires may not be totally warmed up and making it more vulnerable to a runaway situation it is susceptible to when the two elements come into play with each other.

I think more track time would help me learn how to avoided getting into the problem area that I think is inherent with the 1M setup but I would rather change it's character to be more predictable instead of having to worry I will trip over it again and total my car. Since I have had other cars with setup parameters that are similar to the 1M without having a switch go off with no warning, I do not see why I could not achieve the same with the 1M setup as long as we can put our finger on what factors are coming into play in it's setup and then assess where to go with changing how they work together so that the car can still go fast and yet be well behaved. I think I have been able to get most of the items coming into play identified with a lot of help from Pete as well as others that have observed the situation and then ascertain whether it fits with all the scenarios I have experienced so far.

I would welcome anyone else who could weigh in on this with their opinion and observations even if it might indicate that I don't know what I'm saying, I would like to hope this discussion may someday help others in our forum have a better understanding of the car given that I do not believe drifting is the fastest way around the track . Thanks again everyone.