[Track/S]

Originally Posted by Westersund

I didn't know I had to proove myself to some random guy on the internet so I didn't record the draining process. You really think the oil isn't blended in the rear axle and the color would be uniform? I don't know how you use your car, my car is mostly a daily driver (I drive it hard sometimes) but it sees track time on the Nürburgring but not on conventional tracks.

That random guy could be your teacher, so relax a bit.

[vrooooom]

So for some reason I've kept the drained diff oil from the last time (a few months undisturbed) and when I checked it by pouring it into another bottle, it was still clear....until it got to the last 10% or so.

When I do my fluid change I usually drive the car around the blocks a few times before I start draining, starting w/ the engine oil. Before I drain the rear diff, I turn the rear wheels a few times. Car's on a Quickjack. I figured (I'm no mechanic) that should whoosh the oil around and allow the particles to drain out.

Tell us how that Motul do after. Would appreciate it. Apparently the rear diff is very finicky w/ oil.

[Westersund]

Originally Posted by vrooooom

Tell us how that Motul do after. Would appreciate it. Apparently the rear diff is very finicky w/ oil.

The BMW diffs aren't really finicky when it comes to oil, as long as you use 75W-140 with limited slip additives it will perform just fine. The added modifiers in the BMW branded oil is just a comfort thing to allow for more slipping of the clutch plates driving around tight corners/parking lots to reduce clunking noises. In fact it will reduce the effectiveness of the differential during spirited driving/track time because of the added slipping. Anyone who have driven with a race differential knows the kind of clunking I'm talking about, with a newly built race differential is sounds like the entire prop shaft is going to jump out of the car as the built up tension in the clutch plates let go.

I will let you know how the Motul oil performes.

I changed the gearbox oil today and it was really time to change it, the fluid was grayish. This MTF-LT 5 fluid is almost like water, makes for smooth gear changes but I'm not sure how good it handles abuse/heat.

[E90convert]

Originally Posted by Westersund

I can't explain how it could be that clean. I bought the car at 23000 Km and now it's at 43000 Km. I highly doubt the 2 previous owners ever changed the fluid and there's no record of it. I will change the gearbox oil today and see how that looks.

The Motul Gear Comp is 100% synthetic as it is ester based and the Castrol (that BMW use) is AFAIK mineral based and modified to be "fully synthetic". Don't get me wrong, I use the Castrol as well (Not BMW branded though) and it works excellent. I've used Liqui Moly 75W-140 LS in my E46 M3 race car that want's the same kind of oil as the M2 and it works great. But comparing the Motul to the Castrol, they have the same viscosity at 100C but the Motul has lower viscosity at 20C which makes it a better at cold starts.

Castrol SAF XJ is a PAO “true” synthetic.

[E90convert]

Originally Posted by Westersund

The BMW diffs aren't really finicky when it comes to oil, as long as you use 75W-140 with limited slip additives it will perform just fine. The added modifiers in the BMW branded oil is just a comfort thing to allow for more slipping of the clutch plates driving around tight corners/parking lots to reduce clunking noises. In fact it will reduce the effectiveness of the differential during spirited driving/track time because of the added slipping. Anyone who have driven with a race differential knows the kind of clunking I'm talking about, with a newly built race differential is sounds like the entire prop shaft is going to jump out of the car as the built up tension in the clutch plates let go.

I will let you know how the Motul oil performes.

I changed the gearbox oil today and it was really time to change it, the fluid was grayish. This MTF-LT 5 fluid is almost like water, makes for smooth gear changes but I'm not sure how good it handles abuse/heat.

The the electronic LSD, the calibration is specific to the BMW/Castrol oil. So using something different may result less than optimized differential performance in all operating conditions.

I’d bet the MTF-LT-5 will handle heat just as well as any other PAO synthetic. It’s the gears, bearings, and seals that are the question! I’m sure they are fine, it’s designed and for that oil viscosity. The transmission doesn’t get that hot on an M2 either, and the thin oil helps keep the temps lower.

[F87source]

Originally Posted by E90convert

The the electronic LSD, the calibration is specific to the BMW/Castrol oil. So using something different may result less than optimized differential performance in all operating conditions.

This is untrue for so many reasons:

1) The actual OEM fluid for these differentials isn't castrol's/bmw's SAF-XJ, it's actually BMW's MSP/A, and MSP/A is unlikely to be a Castrol product (I believe Castrol I'm Germany is based in Hamburg, while MSP/A is made in Munich). The only country that uses SAF-XJ is the USA (well Canada will soon move over to SAF-XJ too, because of logistics but when I changed my diff oil in 2020 and 2016 I was given MSP/A).


2) ISTA specifies bmw approved differential oils, and liquimoly among others were approved as suitable replacements for bmw's diff oil.


3) the differential isn't as tunable as you think it is, the 2 servos are driven by a pwm signal. The higher the pwm the more force the servos put onto the clutch packs which locks the rear axles tighter. This tuning is quite granular and not based on 1% increases in pwm (because the changes in clamping forces is way too small to make a difference). Instead the pwm changes is likely in the 5%, 10%, 25% increments. What really makes the tuning work is how these servos respond to yaw angle, and wheel speed - pretty much the dsc and stability control system. Then it determines if it wants more wheel lock or not, and if it still needs more wheel lock itll request more pwm from the servos. In addition to that it can apply that pwm to the left or right wheel independently to vector torque. The ghas coding likely helps it determine how agressive it vectors torque. So pretty much it's the car's vehicle dynamics tuning that determines how much the diff engages and if it needs to engage more to meet targets (yaw and wheel speed), the diff itself is pretty granular in terms of tuning (pwm and left or right).

So the idea that it's calibrated only for the Castrol/BMW oil is absurd.

Originally Posted by E90convert

I’d bet the MTF-LT-5 will handle heat just as well as any other PAO synthetic. It’s the gears, bearings, and seals that are the question! I’m sure they are fine, it’s designed and for that oil viscosity. The transmission doesn’t get that hot on an M2 either, and the thin oil helps keep the temps lower.

I doubt it, bmw oem fluids are designed to meet 2 main things: oxidation resistance to last "a lifetime", and low viscosity (mtf from bmw is super thin) to reduce drivetrain losses and meet emissions targets. This means they're not going to be able to bring a lot of Pao's and esters to the mix because it will break down too fast to be a "lifetime fluid", and so they have to use more group 3 oils. We know for a fact esters and Pao oils tolerate heat way better than group 3 oils, so there is no way bmw's mtf is going to tolerate heat better than something like a redline dctf which is all esters and Pao's.


The transmission may be designed for that oil viscosity, but that doesn't say anything about sheering and viscosity decreases due to heat.


What is your proof the 6mt doesn't get hot on the m2? Have you ever done an oil change on it before? Because I have and it gets unbelievably hot even with just city driving for an hour I can barely touch the transmission case. When I drained the oil itself, that was after jacking up the car, and doing an oil change so ~1hr total before I drained the mtf, and it was still as hot as a cup of coffee. You get from the store.

So im willing to say when on track that mtf is easily exceeding 100C.

[E90convert]

Originally Posted by F87source

This is untrue for so many reasons:

1) The actual OEM fluid for these differentials isn't castrol's/bmw's SAF-XJ, it's actually BMW's MSP/A, and MSP/A is unlikely to be a Castrol product (I believe Castrol I'm Germany is based in Hamburg, while MSP/A is made in Munich). The only country that uses SAF-XJ is the USA (well Canada will soon move over to SAF-XJ too, because of logistics but when I changed my diff oil in 2020 and 2016 I was given MSP/A).


2) ISTA specifies bmw approved differential oils, and liquimoly among others were approved as suitable replacements for bmw's diff oil.


3) the differential isn't as tunable as you think it is, the 2 servos are driven by a pwm signal. The higher the pwm the more force the servos put onto the clutch packs which locks the rear axles tighter. This tuning is quite granular and not based on 1% increases in pwm (because the changes in clamping forces is way too small to make a difference). Instead the pwm changes is likely in the 5%, 10%, 25% increments. What really makes the tuning work is how these servos respond to yaw angle, and wheel speed - pretty much the dsc and stability control system. Then it determines if it wants more wheel lock or not, and if it still needs more wheel lock itll request more pwm from the servos. In addition to that it can apply that pwm to the left or right wheel independently to vector torque. The ghas coding likely helps it determine how agressive it vectors torque. So pretty much it's the car's vehicle dynamics tuning that determines how much the diff engages and if it needs to engage more to meet targets (yaw and wheel speed), the diff itself is pretty granular in terms of tuning (pwm and left or right).

So the idea that it's calibrated only for the Castrol/BMW oil is absurd.




I doubt it, bmw oem fluids are designed to meet 2 main things: oxidation resistance to last "a lifetime", and low viscosity (mtf from bmw is super thin) to reduce drivetrain losses and meet emissions targets. This means they're not going to be able to bring a lot of Pao's and esters to the mix because it will break down too fast to be a "lifetime fluid", and so they have to use more group 3 oils. We know for a fact esters and Pao oils tolerate heat way better than group 3 oils, so there is no way bmw's mtf is going to tolerate heat better than something like a redline dctf which is all esters and Pao's.


The transmission may be designed for that oil viscosity, but that doesn't say anything about sheering and viscosity decreases due to heat.


What is your proof the 6mt doesn't get hot on the m2? Have you ever done an oil change on it before? Because I have and it gets unbelievably hot even with just city driving for an hour I can barely touch the transmission case. When I drained the oil itself, that was after jacking up the car, and doing an oil change so ~1hr total before I drained the mtf, and it was still as hot as a cup of coffee. You get from the store.

So im willing to say when on track that mtf is easily exceeding 100C.

1. Do you think Castrol makes MSP/A and it is just rebranded as a "BMW Genuine" oil? A VOA could at least determine if it's similar chemistry. I guess all BMW approved oils would likely have similar chemistry though.

2. That is nice that ISTA gives a list of approved oils.

3. Consider that the oil additives influence how much torque transfer you get through the clutches at a given PWM. For example, say you're on a corner, and the controller commands 60% PWM based on your mentioned inputs (yaw, steering, throttle, ect.). If you change the oil properties, you will get more or less torque transfer than intended by the original calibration. With that different oil, the controller may need to be recalibrated to send a different PWM signal to get the ideal torque transfer for that scenario. This may be more noticeable in certain scenarios. Others, maybe not so obvious to the average driver.

Regarding the MTF - oxidation and efficiency are important. I'm sure you are aware of many other functions of the oil that are also important. Gear and bearing, wear protection, synchronizer friction, foaming, elastomer and sealant compatibility, ect. The driveline fluids I am used to testing with at my work are all PAO based. Consider that BMW likely worked with the oil supplier (Pentosin?) for several years to fine-tune the oil and transmission mechanical design properties. Then spent several million dollars validating every functional and durability aspect of the transmission with that oil.

Unbelievably hot just because you can't touch it? You can't touch 120F for more than a few seconds, and oil doesn't even begin to oxidize due to temperature below 160F. I don't have data on how hot the M2/3/4 6-speed transmission gets on a stock car, but my DCT barely gets over 200F with hard track driving, and the rear diff temped at 170F at the end of a session. A 6MT will have fewer thermal losses (more efficient) than a bevel gearset axle. Combine that with a greater surface area to dissipate heat, and it will run pretty cool maybe lower than the differential. Next time I'm at the track, I'll use my thermal camera to get the transmission case temperatures of an M2/3/4 after a session and reply with my findings.

[F87source]

Originally Posted by E90convert

1. Do you think Castrol makes MSP/A and it is just rebranded as a "BMW Genuine" oil? A VOA could at least determine if it's similar chemistry. I guess all BMW approved oils would likely have similar chemistry though.

2. That is nice that ISTA gives a list of approved oils.

3. Consider that the oil additives influence how much torque transfer you get through the clutches at a given PWM. For example, say you're on a corner, and the controller commands 60% PWM based on your mentioned inputs (yaw, steering, throttle, ect.). If you change the oil properties, you will get more or less torque transfer than intended by the original calibration. With that different oil, the controller may need to be recalibrated to send a different PWM signal to get the ideal torque transfer for that scenario. This may be more noticeable in certain scenarios. Others, maybe not so obvious to the average driver.

Regarding the MTF - oxidation and efficiency are important. I'm sure you are aware of many other functions of the oil that are also important. Gear and bearing, wear protection, synchronizer friction, foaming, elastomer and sealant compatibility, ect. The driveline fluids I am used to testing with at my work are all PAO based. Consider that BMW likely worked with the oil supplier (Pentosin?) for several years to fine-tune the oil and transmission mechanical design properties. Then spent several million dollars validating every functional and durability aspect of the transmission with that oil.

Unbelievably hot just because you can't touch it? You can't touch 120F for more than a few seconds, and oil doesn't even begin to oxidize due to temperature below 160F. I don't have data on how hot the M2/3/4 6-speed transmission gets on a stock car, but my DCT barely gets over 200F with hard track driving, and the rear diff temped at 170F at the end of a session. A 6MT will have fewer thermal losses (more efficient) than a bevel gearset axle. Combine that with a greater surface area to dissipate heat, and it will run pretty cool maybe lower than the differential. Next time I'm at the track, I'll use my thermal camera to get the transmission case temperatures of an M2/3/4 after a session and reply with my findings.

1. I highly doubt it, I think Castrol makes it in the USA because there might be a contract with bmw north America. Canada used to use the German stuff because we need both french and English on all of our packages, so it was cheaper to get the German stuff with pretty much 10 languages already printed on it vs. relabeling the US oil. But covid changed that, and we are moving over to SAF-XJ.

Bmw also says the msp/a is made in Munich Germany, and iirc Castrol doesn't produce their oils in Munich so it's gotta be a different OE, I know they have HQ's in Hamburg though.


I've heard discussions about the differences and people have said the FM additive packs are different. But that's all I could find. It's not easy to do a VOA because it's hard to get both oils in the same place.

2. Yep, ISTA gives approved engine oils and diff oils. Personally I would always stick with OEM diff oils and here is my reasoning:

The diffs on these f8x M cars have always been Uber fragile. From day 1 these diffs have been known to fail when the wrong fluid was used (SAF-XJ + FM for the e9x m3 was used instead of SAF-XJ, and this extra friction modifiers from the FM blew up diffs). This is also what prompted bmw to issue a recall for these diffs in 2016 and the tsb also stayed MSP/A is the fluid that must be used (well except in the USA). Anyways what this tells me is these diffs don't like extra FM because is can damage the clutch packs, but aftermarket diff oils keep saying it's compatible for the E9X m3 and the F8X M cars. But how can this be the case? One of these cars needs more FM than the other, so there shouldn't be a one size fits all fluid, and if there is, would there be too much FM for the f8x diff? And that's why I use OEM fluids for the rear diff, because even Castrol made SAF-XJ (now called syntrax) says it's e9x m3 compatible, so that worries me.


3) the FM has to be I'm a certain window to be permitted, and it's there more for noise and clutch slippage rather than engagement. But let's say you get less torque transfer, the yaw angles would still be too much, and then the car will request more pwm. There is still no way that there will be a massive difference in clamping load and torque transfer that 100% pwm on one fluid isn't going to be able to transfer the same amlint of torque as another fluid. Essentially we are working on an open loop system just like with afrs, because clutch packs and torque transfers aren't the only important thing in this equation, remember tire grips and surface conditions always change and that's a way larger factor to account for than clutch packs and oils. So that's why open loop tuning exists, and why feedback loops from sensors are way better than just referencing vs. tables in fully closed loop mode.


ASTM C1055 says 140F is the temp that average humans can touch for 5 seconds before getting burned. A cup of coffee on average is 70C or 158F, and I have no problem touching that for quick moments. This transmission was pretty much on that level where I can barely touch it, and that's after sitting for 1 hour.

Don't forget the dct has active cooling.the 6mt doesn't, and the 6mt required active cooling on the last gen e9x m3, but bmw made the dry sump lube system and transmission efficient enough that this generation doesn't need it. So likely it is on the absolute limit temp wise.

[TUNEDM3]

Does anyone have the part numbers for both the oem manual transmission fluid and the differential fluid?

[TheHaze]

Originally Posted by TUNEDM3

Does anyone have the part numbers for both the oem manual transmission fluid and the differential fluid?

6MT fluid - 83222156969 (formerly 83222239654)
Rear diff fluid - 83222357992 (SAF-XJ)

[TUNEDM3]

Originally Posted by TheHaze

6MT fluid - 83222156969 (formerly 83222239654)
Rear diff fluid - 83222357992 (SAF-XJ)

Awesome thank you…MSP/A is not available in the US?

[TheHaze]

Originally Posted by TUNEDM3

Awesome thank you…MSP/A is not available in the US?

Nope, I was never able to source it. SAF-XJ is what the dealership used for the 1,200 mile break-in service and it's what I plan on using when I service it myself in a couple months.

[E90convert]

I just did a drain and fill with SAF-XJ 83222357992 at 32k miles. Differential performance was really impressive on the last few track days.

[vrooooom]

How do you know that the diff performed “really impressively”? As in just by feel? I feel some people are just more in touch with their cars, and I’m not.

I didn’t know the LSD on this car had servos and getting datas from several points. I guess it’s a lot smarter than just a “dumb” traditional LSD. A lot of info here.

[F87source]

Originally Posted by TUNEDM3

Awesome thank you…MSP/A is not available in the US?

Unfortunately no, it was readily available in Canada, but it's being phased out now. It's still available at some dealers but probably not for too much longer.

[F87source]

Originally Posted by vrooooom

How do you know that the diff performed “really impressively”? As in just by feel? I feel some people are just more in touch with their cars, and I’m not.

I didn’t know the LSD on this car had servos and getting datas from several points. I guess it’s a lot smarter than just a “dumb” traditional LSD. A lot of info here.

This video explains the f8x ediff:

Also it seems I was completely wrong in my pwm explanation, on the bmw specific ediff there isn't as much complexity in pwm control as other torque vectoring ediffs which I was comparing it to (also only 1 motor). It seems it has 3 positions, open, locked, and limited slip and it can transition in between those.

[vrooooom]

Thanks for the vid F87source! very interesting.

When our diff oil comes out dark, where's the majority of the material coming from—the friction discs on the clutch?

So based on that vid, only one side has that variable clamping force being applied. the left is always engaged w/ the motor.

Is it ok to stock-up on the MSP/A rear diff oil, good for 5 years let's say or does that have expiry too?

[F87source]

Originally Posted by vrooooom

Thanks for the vid F87source! very interesting.

When our diff oil comes out dark, where's the majority of the material coming from—the friction discs on the clutch?

So based on that vid, only one side has that variable clamping force being applied. the left is always engaged w/ the motor.

Is it ok to stock-up on the MSP/A rear diff oil, good for 5 years let's say or does that have expiry too?

Yup no worries, pretty much scratch out my previous explanation above because that was not entirely applicable to the f8x ediff. The feedback loop is though, because if the yaw rates dictate that car is slipping the diff will respond by going into limited slip mode, if that's still not enough it'll go full lock. And of course surface and tire conditions are far more important of a factor than the oil.


Now to answer your questions:

1) yes the friction discs is the source of the majority of the debris, it should be ferrous material so a magnetic drain plug like the tms titanium drain plug should help.

2) no don't stock up on MSP/A, it has a 4 year expiry date from the time of manufacturing. So by the time you use it, it'll be expired. Also since you're in the Philippines you should be able to secure MSP/A so long as Germany keeps making it. It's just the US and Canada that will switch to SAF-XJ (atleast from what I can tell).

[E90convert]

Originally Posted by vrooooom

How do you know that the diff performed “really impressively”? As in just by feel? I feel some people are just more in touch with their cars, and I’m not.

I didn’t know the LSD on this car had servos and getting datas from several points. I guess it’s a lot smarter than just a “dumb” traditional LSD. A lot of info here.

I've spent a bit of time in cars with mechanical LSDs (S2000, C6Z) and cars with an open diff (BMW 335i and Porsche Cayman S). The most noticeable action from this eLSD is it's ability to slingshot you out of corners in a controlled manner. I can trust that it will keep me going forward and not get squirrely. This probably has some to do with the rear suspension geometry as well. The Corvette if you get a lot of locking action on the corner exit, the rear would feel unstable. For the S2000 I had some wide 200 treadwear, and couldn't feel the LSD working all the much. The 335i would not rotate through corners with any consistency or controllability. The Porsche didn't need and LSD, it had so much rear grip. What a joy that car was.

[E90convert]

You are right about PWM. It has two hall effect sensors that will be PWM driven since there is no contact position sensing (like in a Ford transfer case actuator). It likely has TWO hall effect sensors because they are trying to meet an ISO 26262 functional safety ASIL (automotive safety integrity level) requirement requiring a redundant sensor. The controller probably sends a PWM signal based on a locking position requested from the calibration tables. The motor draws current to reach that position. It looks like it is a "dumb" DC motor where it just receives current from an H-bridge in a controller. Very similar to how the xDrive systems work with torque vectoring front to rear. F87source you have access to these calibration tables?

[E90convert]

Originally Posted by F87source

Yup no worries, pretty much scratch out my previous explanation above because that was not entirely applicable to the f8x ediff. The feedback loop is though, because if the yaw rates dictate that car is slipping the diff will respond by going into limited slip mode, if that's still not enough it'll go full lock. And of course surface and tire conditions are far more important of a factor than the oil.


Now to answer your questions:

1) yes the friction discs is the source of the majority of the debris, it should be ferrous material so a magnetic drain plug like the tms titanium drain plug should help.

2) no don't stock up on MSP/A, it has a 4 year expiry date from the time of manufacturing. So by the time you use it, it'll be expired. Also since you're in the Philippines you should be able to secure MSP/A so long as Germany keeps making it. It's just the US and Canada that will switch to SAF-XJ (atleast from what I can tell).

Actually... the friction material is MoS2... which is non-magnetic. The contamination generated by the LSD and more severe ///M duty cycles may be damaging to the bearings and oxidize the oil faster which necessitates the diff oil change interval. Anecdotally, magnets are always helpful to pull the iron out of the oil. Helps with cleanliness and oil oxidation.

[F87source]

Originally Posted by E90convert

You are right about PWM. It has two hall effect sensors that will be PWM driven since there is no contact position sensing (like in a Ford transfer case actuator). It likely has TWO hall effect sensors because they are trying to meet an ISO 26262 functional safety ASIL (automotive safety integrity level) requirement requiring a redundant sensor. The controller probably sends a PWM signal based on a locking position requested from the calibration tables. The motor draws current to reach that position. It looks like it is a "dumb" DC motor where it just receives current from an H-bridge in a controller. Very similar to how the xDrive systems work with torque vectoring front to rear. F87source you have access to these calibration tables? There

Yeah the rear diff is pretty "dumb", in that it has 3 modes (open, LSD, and fully locked) and these modes are activated based on what the car wants. Calibration is likely just how the car activates the 3 diff modes based on yaw and wheel speeds.

Unfortunately I have no idea if I have access to those tables, I'll have to check, but I'm currently still over seas so I'm still extemely busy with work. I still need to check those temp tables for you but that's also back logged behind the things I'm obligated to do first.