Variable M differential lock

[Pete_vB]

Originally Posted by Dackelone

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.

That's exactly what I'd expect from any true LSD, and matches my experience. When you're spinning up an inside rear the outside stays stuck, and is able to contribute cornering force. As soon as both rear wheels spin up the rear swings wide. However the LSD doesn't effect constant speed sweepers, so it's a question of where the issue is occurring. Any video Scott?

The 944T seems a popular (and great) car. I had three- a chipped '86 with no LSD, an '87 with 340 whp plus full turbo cup suspension, and another '86 with a 968 6 speed, 3x motons and a number of goodies. I prefer the 1M as a street car and autocrosser, but as a track car the 951 has some significant advantages, particularly stability and the way it puts power down.

I've had both gear and plate type diffs, and have some experience setting them up with different ramp rates so have a reasonable feel for them, but I've never played with the M diff. I do suspect there are a few different things going on...

[scotth944]

Unfortunately I have not done any kind of logging of the issue at this point. I wonder if I might be able to get individual wheel speed reads off the OBD2 port using logging software which I would hope could help show how it manifests into a problem. I know there is a sensor on the rear end suspension to measure movement but I have know idea how it would be accessed, more research!

As far as limited slips are concerned, I had one in the 944 and was very in tune with being able to slid and spin the car around if a wanted to but I had to purposely force it into these conditions. There was one experience in particular with the 1M that I was not able to sense any of the wheel spin at about 80-90 MPH until the car started the spin. The instructor that was in the car with me also indicated that he did not sense any issues developing until the spin as well. There were other instructors and bystanders that reported seeing white smoke developing which they said is usually indicative of wheel spin in this situation. I had taken the turn-in slower than the car was capable of and went full power (long sweeper) by the apex which I had been able to do before but after the apex with no issues although MBM disable was always engaged until the run session when I spun. The car was holding without any issues through the apex and it was not until after I started tracking out when the spin occurred. I have looked at other cars going through the same turn and there appears to be a very slight bobble near mid turn based on wheel/fender observations. I think the inside wheel must have lost traction at the apex going over the bobble and it was not until after I had started the turnout a moment later that the LSD kicked in and I lost the outside tire, again, all theory.

I have been doing some research on LSD and I am wondering if WPC treatment (http://www.wpctreatment.com/index.htm) might help with how fast the LSD engages, Pete, have you experimented with this process yet?

[Pete_vB]

Originally Posted by scotth944

There was one experience in particular with the 1M that I was not able to sense any of the wheel spin at about 80-90 MPH until the car started the spin... I had taken the turn-in slower than the car was capable of and went full power (long sweeper) by the apex which I had been able to do before but after the apex with no issues although MBM disable was always engaged until the run session when I spun. The car was holding without any issues through the apex and it was not until after I started tracking out when the spin occurred. I have looked at other cars going through the same turn and there appears to be a very slight bobble near mid turn based on wheel/fender observations. I think the inside wheel must have lost traction at the apex going over the bobble and it was not until after I had started the turnout a moment later that the LSD kicked in and I lost the outside tire, again, all theory.

So you're after the apex, you've been at full power for how long? Accelerating, 4th gear? Are you unwinding the wheel yet? Certainly a strange place to snap-spin, after the rear is fully loaded up. Any video of the corner in-car from another car that you can find?

Originally Posted by scotth944

I have been doing some research on LSD and I am wondering if WPC treatment (http://www.wpctreatment.com/index.htm) might help with how fast the LSD engages, Pete, have you experimented with this process yet?

Never heard of anyone using it on a diff. Have you? Doesn't seem like the ideal application.

I'd think you could change fluid if you want to play with friction, I suspect it's sensitive and BMW is already putting in friction modifiers to cut down on noise. Would need to do some research, though...

[MDORPHN]

I believe there's a TSB (or somesuch) suggesting that dealers add a friction modifier to the fluid only for noisy diffs.

EDIT: Yep, referenced in the sticky -- http://www.1addicts.com/forums/showthread.php?t=620282


Neil

[scotth944]

"So you're after the apex, you've been at full power for how long? Accelerating, 4th gear? Are you unwinding the wheel yet? Certainly a strange place to snap-spin, after the rear is fully loaded up. Any video of the corner in-car from another car that you can find?"

I had transitioned from 50% throttle after the turn-in to 100% at the apex, I think I had been on full power for about 2 seconds, in 3rd and started unwinding the wheel when it let go. I was holding a line a little closer to the inside of the turn than most people take so that if the front started to push, I had room to recover which it never did. The spin snap was so tight, I stayed within a car length in the direction of travel as the car did a 360 at the end of the turn and going down the straight away.

Neil, I was wondering if the friction modifier would also change the rate in which the LSD locks up since it helps with the noise but I have not seen anyone make that claim so far; WPC creates a micro size divots in the surface to help hold oil so it will also slow down the rate of meshing with the plates and had been confirmed to work in some posted projects I found on the web for LSD setups.

Does anyone know of someone with a worn out M diff I can do a post mortem on to see what makes it tick?

[JimD]

Try this article:

http://auto.howstuffworks.com/differential4.htm

Or this:

http://en.wikipedia.org/wiki/Limited-slip_differential

This is the best but the others may help too:

http://www.autos.ca/auto-tech/auto-t...erential-lock/

The M differential is essentially a clutch type but unique in having the pressure applied to the spring pack by hydraulic pressure. The pump works based upon the differential in speed of the rear wheels.

You don't really want both rear wheels locked up very often. If you try to go around a corner that way, even a little, then a tire has to slip or the differential has to break. So worrying about whether it locks up solid doesn't make much sense to me. What you want is to transfer torque to the tire with traction.

[scotth944]

Believe me, I wish the LSD didn't lock to 100% with the stock power levels it makes and being a street driven car. My questions are 1- how it is even capable of achieving 100% given that it requires slip to work in the first place and 2- what can I do to change (slow) the process of getting to lockup so the behavior is not so abrupt and a unexpected spin follows.
I have driven a race car at the track with a welded up rear diff to lock it 100% of the time and (limited slip kept blowing apart at the power levels going through it) and handling is not a big issue nor breaking the rear drive train. The inside/outside tires will slip which is a normal part of their operation at their adhesion limits. The one thing to keep in mind with that setup is the car will tend to push on turn-in unless your good at trail braking to get the weight on the rear inside tire lighten up.

[Pete_vB]

Originally Posted by scotth944

Believe me, I wish the LSD didn't lock to 100% with the stock power levels it makes and being a street driven car. My questions are 1- how it is even capable of achieving 100% given that it requires slip to work in the first place and 2- what can I do to change (slow) the process of getting to lockup so the behavior is not so abrupt and a unexpected spin follows.

Assuming it's the diff, I'm not quite sure if you want less locking or more. One school of thought would say that if the differential pumping pressure is proportional to wheelspin it's always going to lock up, it's just a question of when. If this is the case, lower diff friction would simply let the inside wheel spin up more before it catches and locks, giving it much more momentum and potentially upsetting the car more. Better might be to reduce the speed differential the tires get to.

The main reason BMW claims the M diff in the 1M is better than a regular plate style diff has nothing to do with track performance. It doesn't lock uncomfortably on tight radius corners in a parking lot, nor does it have preload like a regular plate diff that would cause it to lock on ice. However preload is generally considered a good thing on the track. Makes me wonder if you don't want increased friction?

I was watching the Stig's off in the 1M again, can be found at 5:55 in on #3 here: Link
He goes past the tires- we know he cuts that corner significantly, so it's a decent bump, and if you look closely you can see the car's sideways from there on. This may have caused rear diff lockup, but you can see the instability from then on as he chases the rear- presumably he backed out of the power at some point, so it's not the diff, but he can't gather it up...

[scotth944]

Loved watching the video again, thanks!

Just for clarity, when I had referred to lockup, I was thinking there was no more RPM difference between left/right spindles and the diff was locked together. Maybe I misunderstood the term in the way they are using it which could be defined as 100% of the torque was getting transferred to the outside tire that was not spinning as fast??

The Stig was driving on a damp surface so this may have helped keep the snap condition from developing and he would certainly have been aware that he would upset the suspension going over the curbing and therefore been proactive with the throttle so the diff did not factor into the slide as much??

I wonder if there is or isn't different dynamics between a driver drifting the car verses the car snapping into a spin unexpectantly other than the driver being able to manage the situation in one?

[Dackelone]

What we need is THE "drift meister" @Advevo to chime in on this topic. He by far has teh most experience drifting various BMW's including his 1M.

[scotth944]

Originally Posted by Dackelone

What we need is THE "drift meister" @Advevo to chime in on this topic. He by far has teh most experience drifting various BMW's including his 1M.

Hopefully he has experienced both situations to comment on since one would be anticipated and theoretically quicker to respond to vs. the delay reaction time I would expect to see from being surprised even if it is small.

Keep in mind that some form of water was again involved with these videos.

[Dackelone]

Originally Posted by Ian///M

Isn't the 1M's diff tuned differently between EU & US markets?

No. the LSD is the same. Your thinking of the MDM mode. The electronic coding is different. On EU cars(or coding) BMW allows for more "yaw" or tail out, than on US spec cars. On the e9x M3's a lot of guys have their MDM mode coded to EU yaw specs.

[scotth944]

Originally Posted by Dackelone

No. the LSD is the same. Your thinking of the MDM mode. The electronic coding is different. On EU cars(or coding) BMW allows for more "yaw" or tail out, than on US spec cars. On the e9x M3's a lot of guys have their MDM mode coded to EU yaw specs.

And MDM keeps the snap from getting out of control when used.

[Dackelone]

Originally Posted by Ian///M

So with all nannies off there is no software involvement, just purely mechanical?

Yes, I believe so.

I can't be 100% sure about the MDM mode on the 1M. I only remember clearly -that it's different on the e9x M3's US vs EU specs. I would think that BMW also did this on the 1M, but I am not 100% sure.


Dack

[ozinaldo]

Originally Posted by Ian///M

So with all nannies off there is no software involvement, just purely mechanical?

With M3 only ABS remains active, rest is like you say with DSC off. With 1M CBC also remains active at all times unlike M3.

Neil sorted it out through coding, had posts about this.

But in principal what you say is correct, with DSC off it is purely a rwd with a variable mechanical differential. Of course, there must still be 'some' electronics involved while deciding the percentage of lock.

[scotth944]

The M diff has no electronic component to it, it is strictly mechanical with an internal hydraulic component.

MDM off and DSC on, the car can start to snap on me but the electronics will catch it. With DSC off, it's up to the driver to recover from it at least in the US 1M.

[M3 Adjuster]

Originally Posted by Pete_vB

So you're after the apex, you've been at full power for how long? Accelerating, 4th gear? Are you unwinding the wheel yet? Certainly a strange place to snap-spin, after the rear is fully loaded up. Any video of the corner in-car from another car that you can find?


..

Originally Posted by scotth944

I had transitioned from 50% throttle after the turn-in to 100% at the apex, I think I had been on full power for about 2 seconds, in 3rd and started unwinding the wheel when it let go. I was holding a line a little closer to the inside of the turn than most people take so that if the front started to push, I had room to recover which it never did. The spin snap was so tight, I stayed within a car length in the direction of travel as the car did a 360 at the end of the turn and going down the straight away.

hmmm this sounds eerily familiar....

Originally Posted by Pete_vB

I was watching the Stig's off in the 1M again, can be found at 5:55 in on #3 here: Link
He goes past the tires- we know he cuts that corner significantly, so it's a decent bump, and if you look closely you can see the car's sideways from there on. This may have caused rear diff lockup, but you can see the instability from then on as he chases the rear- presumably he backed out of the power at some point, so it's not the diff, but he can't gather it up...

since you asked for video..... I actually have some... Clearly I'm no better than the Stig....I couldnt gather this one up.... It's literally one of the only offs I have had in the 1M and of course it was conveniently caught by a friend. At any rate... It definitely has had me scratching my head, just like you are over your spin Scott... so here it is...

This is a third gear corner... i entered it a little bit early.. got out of the throttle a little... and then found I was riding a pendulum... and as you can see.. I nearly had it gathered more than once.... and I swear that there was something going on from the car.. perhaps corner brake control?

Originally Posted by Dackelone

No. the LSD is the same. Your thinking of the MDM mode. The electronic coding is different. On EU cars(or coding) BMW allows for more "yaw" or tail out, than on US spec cars. On the e9x M3's a lot of guys have their MDM mode coded to EU yaw specs.

Correct. I recently had my car coded to the Euro version of MDM. It absolutely allows for more yaw and allows the car to be rotated without cutting the power nearly as soon. I've only been to the track a few weekends with it and autocrossed once and played with it a little bit, but I am far happier.

I can lap around a track fairly quickly with my car in the US MDM mode by driving smoothly within the limits of MDM. In spots where the power is being held back in the US version.. there is no power cut in the Euro version and the slip angles that can be generated are much larger.

Originally Posted by scotth944

And MDM keeps the snap from getting out of control when used.

yes it most certainly does ! The US version of MDM is actually pretty good for the street because it does cut in before a driver gets totally out of shape. The Euro version of MDM will let you get the car significantly more sideways, perhaps that it may not cut the power in time to save a novice or hamfisted driver.

With track tires on my car, the car is pretty much worthless in US MDM mode, because you can get the light blinking warming the tires up in the paddock. With Euro MDM, one can get on track and lap fairly hard with R tires. Certainly 9/10ths.. A major improvement. Lapping at 9/10ths one will see the traction light blinking on occasion, but it takes a major error or a major change in slip angle to cause a major power cut. At the autocross course... I can actually almost lap as quickly with Euro MDM as I can with traction control off , due to less overall wheelspin.

Originally Posted by ozinaldo

With 1M CBC also remains active at all times unlike M3.

Neil sorted it out through coding, had posts about this.

I went and found the posts... here..

http://www.1addicts.com/forums/showt...&highlight=cbc

It looks like Mike Benvo has sorted out the dreaded CBC and can set it to be fully off when the DSC is fully off. I had my coded by another person, Alex @ Alpine who wasn't immediately familiar with the fact that CBC is there and can be turned off. I haven't noticed CBC yet.. but need more time on track and probably won't know for sure until I put track pads on again. now that I have found this thread I may see if i can press to find out for sure.

[Advevo]

The M diff.

First thing i have to say you need to get used to it. It s not an easy one.

My experience.

It s really feels like an open diff if you have no wheel spin.

When you start a drift you have half a second inside wheelspin and then diff locks quite extreme you need to be fast on the counter steering otherwise you loose the car.

This diff is a compromise to give comfort in daily driving and give pleasure if you want to drive extreme.

This diff is not the best solution for track days and drive with wide and high grip tires. It s a bit on off diff lock. The diff does not come in smootly on high grip surfaces. The diff is best if you drive on 18 inch winter tires in rainy conditions. The less grip you have you will enjoy the diff the best.

On a dry surface with ps2 tires you get some understeer on throttle but when you push harder the diff can lock suddenly when you are not expect it and backend swings out hard.

The best way the counter this. Try not to use the throttle to early midcorner. Go in fast and trail brake long into the corner and when the car is almost straight go easy on the throttle. The earlier you go on throttle and wheelspin comes up you get half second spare time but then rear diff will lock hard and send the back out. The more grip you have the faster you need to react and counter steer.

If you're in a drift never back off completely off throttle. What happens then. You're diff is going from lock to open again and you get instant grip in you're counter steering. The car wil go very hard the other way and you loose the back end. Be careful with this.

The diff works a bit like a rwd rally car. Most rally cars rwd have 70 procent lock. To get most traction out of corners. But mostly they set up the car on brakes or handbrake the car into a corner and need maximum traction out of the corner.

But on race cars i rather have a low lock on diffs or diffs which lock very smootly on the throttle input.

A 70% lock on a race car is to much it give lots of understeer. Then you always need to steer with the back end to get the nose in that means to much heat in the rear tires. This is rally style i like this a lot for fun. But for fast lap times it s a no go.

The best for track is to get as much grip you can get on rear and use as little rear diff lock as possible to counter understeer on throttle.

But the M diff is the best solution for road comfort and some fun. If you have 25 or 40% plated diff you get more cracking sounds etc. You don t want that in a road car. And if you're in a drift the nicest you can have is 100% lock it is sooooo nice if you can steer the car on throttle then.

But for a track day or permanent race car the best is always to go for a plated diff 25 or 40% lock. M diff is not very suitable for slicks.

The M diff is a compromise and a very good one. But you need te be aware if the diff locks. It takes no prisioners if you react to slow.

[scotth944]

This is some good feedback, thanks.

I have replayed M3 Adj. video over and over to see what initiated the slide. I can see a light black track from the inside rear has started to develop but unfortunately the left side is not discernible with all the other tread marks that had been left. As the slide progresses, you can see distinctive tread marks being left by both rear tires and this continues on and off as he is chasing down the sliding back and forth. Another detail I thought I could pick out was while some of the slides get started, the tread marks appeared to only be coming from one side of both tires which corresponded to the direction of the slide so I wonder if the contact patch is what makes it so easy to get started??

Pete, can you make out what the chassis may be doing as the slide gets started?

[scotth944]

has anyone out there used the Autoenginuity - OBD-II Scantool W/ Enhanced BMW Expansion logging, I spoke to someone at the company and he said they work with the 1M coupe but he could not tell me what the specific sensors are available via OBD2 in this car and what sampling rate they are capable of logging at. Some of the info I have found gives me the impression this might work for capturing the cause/effect relationship going on during a snap-spin which the on-board systems can track to some degree but I can't be 100% positive with the info I've seen so far.

Does anyone know someone who might?

[Pete_vB]

Originally Posted by M3 Adjuster

Clearly I'm no better than the Stig....I couldnt gather this one up.... It definitely has had me scratching my head.

Very interesting... Definitely looks nearly identical to the off the Stig had. So I don't know what's going on here, but I can theorize... It's very difficult to tell what caused it, but it's almost certainly magnified by some kind of interaction with the car that causes instability.

If you look closely at the video you can see both rear tires laying stripes, the rear end fairly loose as you'd get coming hot into a decreasing radius corner, which is expected. The main issue seems to be caused by the way the rear end "bites" when the car straightens out. Watch the skid marks and the path of the rear tires- the correction causes the rear to snap violently back in line.

In shock terms we're seeing load go onto the inside rear (bump), and off the outside rear and outside front (rebound).

In this thread I discussed what BMW appeared to be doing with the shocks and springs vs the M3: they significantly increased the stiffness of the rear shocks, probably in an attempt to promote both corner entry oversteer to help the car rotate, and corner exit understeer to help put power down.

In the case above, however, when you're not on power it looks to me like that same shock tuning probably becomes at least part of the issue. The 1M has maybe 15% higher rear low speed compression than the M3, 2x the low speed rebound, both of which increase rear end grip when the car is unloading (the table here helps explain why), so when the rear shocks start moving that way the rear gains lots of grip and "snaps" back into line and then overshoots. One part of the theory...

Originally Posted by M3 Adjuster

With track tires on my car, the car is pretty much worthless in US MDM mode, because you can get the light blinking warming the tires up in the paddock.

What sizes are you running, by the way?

Originally Posted by M3 Adjuster

I haven't noticed CBC yet.. but need more time on track and probably won't know for sure until I put track pads on again.

CBC is certainly another possible explanation, for this spin at least, but I now suspect the diff more:

Originally Posted by Advevo

If you're in a drift never back off completely off throttle. What happens then. You're diff is going from lock to open again and you get instant grip in you're counter steering. The car wil go very hard the other way and you loose the back end.

Thanks for jumping in here, very interesting also...

This is the other possible explanation for M3 Adjuster's spin, or at least that initial violent correction when the rear catches. That unlocking might also significantly increase rear end grip. The question is, was the rear end catching also accompanied by lifting out of the throttle?

The video shows the rear out of shape four separate times, and I don't know that every one of the slides could have been accompanied by a throttle lift because I suspect at some point you backed out? Thus I think there is more than just the diff going on, but it could certainly have contributed to the first jump.

Originally Posted by Advevo

When you start a drift you have half a second inside wheelspin and then diff locks quite extreme you need to be fast on the counter steering otherwise you loose the car.

This sounds exactly like what Scott originally described. That event sounds different than the Stig or M3 Adjuster's event above. Power on, then violent lockup punts the car sideways.

On balance I'm suspicious that in both of these cases we have two or more events conspiring together, as usual in an accident. It would make sense that in both cases the diff acted violently, either locking or unlocking, and then the stiff rear shock turning contributed to making the resulting slide much harder to catch.

The factory seems to know the diff can be an issue. Not only Dackelone's info from the factory guy, but also the M5's diff and new M3: they add an electric actuator to make the diff function active and under computer control. They would only add this weight, complexity and cost if they felt passive was a problem, letting them regulate the locking and unlocking.

It's just a theory, but it does seem to fit the facts- violent diff lockup and stiff rear shock tuning together could be expected to cause both issues.

So, Scott, my current best guess, for your use, is that you should be looking for another diff, though I suspect it's not a cure-all.

[scotth944]

Thank you everyone for weighing in to help sort out/confirm a plausible theory.

Pete, have you worked on diff setups as well? I have looked into some options available and the cost range is wide enough for me to want a cost/benefit assessment before I were to pull the trigger. At least for now, the cheapest solution is to not turn all nannies off.