[Enabled]

BMW ATC 35L Transfer Case Repair and a Prevention Method




Hello everyone, I believe I have figured out why our transfer case output flanges fail.



Background story:
I noticed that I had a mild-moderate oil leak around the transfer case area. I put it on a lift and I could tell it was coming from the rear output flange seal. I ordered a new seal (just in case) and then days later went to try to install the seal.

2014 328Xd wagon, 89,000 miles, maximum power tune for the last 18,000 miles. This is the factory original transfer case, with a date of late 2013. No symptoms or weird drivability issues.


I put the car on the lift the second time, took off exhaust, all the panels, and the driveshaft partially. I got to the output flange, and immediately noticed a problem. A wobble in the output flange!
That explains the leak...


I pulled off the circlip and pulled out the flange by hand. I noticed that the splines were quite worn down:







From inspection of the splines, this is NOT instant shear force that wore them down. This looks to be from constant vibration.
For proof, look at the symmetry of the wear on each side of the spline. It's even on both sides with sharp peaks. This means they were wearing clockwise and counter clockwise equally, even though the flange spins mostly in one direction for the life of the car.

Splines/output shaft was still holding down the torque, and I was still driving around.
This is not a metallurgy problem or too much torque.

More pictures here https://f30.bimmerpost.com/forums/sh...&postcount=133



I changed my mind about changing the seal and went and picked up my new transfer case from home, which I have been saving for an 'event'.
I transferred the transfer case mount from the original case to the new one, because it was in great shape.






I couldn't just install this new case in, and call it a day... but rather, I pulled off the output flange.

I started with the circlip:




And the flange pulled out... BY HAND. I didn't use a gear puller.




That was huge alarm for me, because it's NOT pressed in nor held in place by a huge nut with high torque.
Essentially, the circlip is the only thing holding it in place.
I believe this is where ATC (designer and manufacturer of these ATC 35L cases) made the mistake regarding the output flange.



I discussed this with a seasoned transmission and drivetrain specialist.

This design makes it very susceptible to vibration, with eventual wear down of the teeth. It wouldn't matter how much power the car is making; more importantly, how much vibration the engine and drivetrain is putting on those spline teeth.

It makes complete sense why the 328xd is the most common of the xDrive vehicles to have it fail, followed by the 328xi. These engines simply generate more vibration.

Sitting at a light, idling with brakes pressed, the engine vibration is always hitting against the splines of the transfer case, the output flange not really secured to the shaft.




Prevention:
A common successful solution for many manufacturers is to use a compound on the splines to prevent vibrations and movement. Ferrari, Lotus, and BMW themselves use this method on splined shafts of axles to keep them from having play.


There is a spline shaft specific Loctite 660 that is probably the best for this:
https://f30.bimmerpost.com/forums/sh...8&postcount=37


I first cleaned the output flange and shaft with brake cleaner. I sprayed it on a rag and cleaned the shaft carefully without spraying inside the transfer case.




At around 5pm it was extremely hard to find Loctite 648/680/660 locally, so I used high temp Loctite 272 instead. This is rated at 3000psi of shear force. The full cure time is 24hr + depending on temperature.





Install the flange:




Clean excess, put o-ring back in, tap in the spacer, and lift the shaft up with flat screwdrivers, to fit the circlip:





Install to car:



Additional Prevention Step:

If your original flex disk (giubo) is aluminum with inserts (BMW 26117605629 , 105mm bolt circle to driveshaft, for cars before 07/2015)





You should update it to the superseding part number BMW 26117610061 (all rubber giubo, 105mm bolt circle, for cars before 07/2015)



4 hour install time, with my wife helping (best helper ever), on a lift, indoors away from the chilly weather. Started swapping it around 4:30-5pm, and we were packing up the tools at 9pm.
I drove it home gently as the Loctite cured some more, no issues.


I will clean up the old transfer case and take pictures of the not-fully worn splines, and of the signs of vibration on the spacer and circlip. Having caught it pre-failure and comparing to the brand new transfer case allows for quite a bit of insight to be gained.





** UPDATE 6/26/20 **
I've received some more insider info regarding this and the Service Bulletin SI B27 01 20 from BMW.
BMW does indeed blame the aluminum flex disc, the counter vibration onto the output flange, and the abrasive action of the worn flange rust speeding up the process of wear.
Interestingly, they didn't propose this SIB for the gas 320xi/328xi which would benefit as well.

The proposed fix to go to a rubber giubo and stopping the wear on the flange is still very relevant and valid.

There are no updates to the transfer case nor the output flange.
Loctite and rubber giubo should fix this one right up for its lifetime.

**

[McQueen2014]

I would be interested to see if the G20 has an upgrade in this area.

[328iX]

I really hope it gets recalled, what a silly failure

[Enabled]

Unfortunately it's not exactly life threatening like an engine fire, or an emissions issue, so a recall would be quite difficult.
It simply sucks. It can be prevented.


I have a good contact at BMW corporate, I will see at least if they can put in a TSB for it.


BMW has a similar TSB for noisy/rattly E90 axles.

[Enabled]

Additionally, I am slightly concerned as to why the xdrive 328d uses the 105mm aluminum front Giubo (flex disk, sometimes known as guibo), instead of a softer rubber one (96mm) the rwd 328d uses for front Giubo. Rubber one would absorb those vibrations a lot better, at the expense of some flex.

[FaRKle!]

Wow, interesting find. The vibration aspect of it definitely explains why the diesel seems to be more affected than the other versions. I wonder what the fit "slop" tolerances are that ATC uses.

[kdog_x]

Very interesting, nice find! When you replace the output flange with new it comes with a dozen or so flat spacer washers of varying thickness that sit behind the flange to take the play out. You would hope that the manufacturer would install the best one, but for every size that they are off you could expect another thousandth of play. Which could explain the really short lived transfer cases. If I end up having to install a fourth transfer case, I may give the loctite a shot! Curious if the rubber guibo off the RWD could be swapped in?

[Enabled]

Originally Posted by kdog_x

Very interesting, nice find! When you replace the output flange with new it comes with a dozen or so flat spacer washers of varying thickness that sit behind the flange to take the play out. You would hope that the manufacturer would install the best one, but for every size that they are off you could expect another thousandth of play. Which could explain the really short lived transfer cases. If I end up having to install a fourth transfer case, I may give the loctite a shot! Curious if the rubber guibo off the RWD could be swapped in?

Did you buy 2 last summer?
I kind of regret not getting 2 at the $600 price. They're $4000 now. Price fluctuates with stock, so I will be watching it.
I will likely get a new shaft and output flange to rebuild the original Tcase. The output shaft has fine wear, another contra for the metallurgy theory.


The spacers are there to prevent axial movement (front to back), but they cannot do anything for the rotational vibration.
My original circlip and spacer had very specific wear "milled" into them.


I am now searching for a rubber 105mm giubo for that location. That by itself may be enough.

I planned to take a video of the giubo under the car, with car in drive and brakes pressed, to see the amount of vibration.. but ran out of time. I would have gotten more pics too.
A comparison video between the aluminum one and a rubber one would be awesome. Sitting at a light today, I put the car in neutral and it does reduce vibratory load quite a bit.
I think sitting still in drive is what is what is wearing these the most.
Once there is a little bit of wear, the damage following that happens at a faster rate.


One more modification I would do to my procedure, besides running Loctite 648, I would put a special type of vibration absorbing grease on the transmission output shaft before installing transfer case on, for protection. This is a common method on Ford spline shafts where there is no contact with oils, without hindering removal.
I will ask my transmission expert contact which grease that is exactly.

[kdog_x]

I wish I bought two last summer. I put in an order for a second one about three months later exactly when they raised the price. It got rejected with no notice. about a week later When I checked on it I saw that they upped the price to $1800 and had cancelled my order. I have about 20 K on the most recent one, so we'll see how long that holds up. At the current price I think I will rebuild it if it goes out again, I have the input shaft on hand I would just need to buy the flange. Not sure if I would replace all the bearings again, it doesn't seem likely that they would be worn. I might try to clean them somehow, in case of metal shavings.

[Enabled]

I will have to get your source for a new shaft and flange, once I open this up (for fun).
I wouldn't change bearings unless absolutely necessary.

[FaRKle!]

Originally Posted by Enabled

I think sitting still in drive is what is what is wearing these the most.
Once there is a little bit of wear, the damage following that happens at a faster rate.

So using auto stop/start at long lights really does have benefits... lol

Looking at a data sheet for loctite, it looks like 680 is only good to 2800psi for steel/steel in the table (even though the description says up to 4000psi). 648 is good to the 4000 (3900psi) though. I didn't see the shear strength for 272 on there though. I wonder how the more common 242 holds up. That'd be a lot easier to separate if you had to later.

[n00bkiller944]

This is sweet. Do you have any pictures?

[Enabled]

272 had the 3000psi written on the back of the package. I saved the backing somewhere. Some online vendors say 2200psi. Which is still plenty.

272 also has a gap fill of 0.010" (0.25mm), vs. 648 at 0.006" (0.15mm). Should account for something. I still believe 648 is best for this, and it can be done with Tcase on car. Thick moly grease (as used on CV joints) on transmission output shaft will provide additional protection, if Tcase is removed.

A full rubber giubo should help out too. Part number 26117610061 brings up a rubber 105mm giubo.

I will also see about asking at BMW corporate about revising the transmission software to pull back how much clamping force happens at idling with brakes applied. The software already does it partially. More neutral would be even better.

[Enabled]

Originally Posted by FaRKle!

So using auto stop/start at long lights really does have benefits... lol

Looking at a data sheet for loctite, it looks like 680 is only good to 2800psi for steel/steel in the table (even though the description says up to 4000psi). 648 is good to the 4000 (3900psi) though. I didn't see the shear strength for 272 on there though. I wonder how the more common 242 holds up. That'd be a lot easier to separate if you had to later.

Haha, yep!



BTW, interesting find: LCI 328xd (from 07/2015 actually) and X3 28d have the 96mm rubber giubo installed at that front position. This explains the slightly different tcase part number for LCI.

[FaRKle!]

Originally Posted by Enabled

272 had the 3000psi written on the back of the package. I saved the backing somewhere. Some online vendors say 2200psi. Which is still plenty.

272 also has a gap fill of 0.010" (0.25mm), vs. 648 at 0.006" (0.15mm). Should account for something. I still believe 648 is best for this, and it can be done with Tcase on car. Thick moly grease (as used on CV joints) on transmission output shaft will provide additional protection, if Tcase is removed.

603 is rated for light contamination/oil, so if someone didn't want to be as thorough cleaning it off as you, they could also go that route. Not quite as much gap fill though (.0005").

Originally Posted by Enabled

A full rubber giubo should help out too. Part number 26117610061 brings up a rubber 105mm giubo.

I will also see about asking at BMW corporate about revising the transmission software to pull back how much clamping force happens at idling with brakes applied. The software already does it partially. More neutral would be even better.

Hrm, so LCI might be more tolerant of the issue.

[CleanKM]

@Enabled

Even though I'm not in the same boat, having mine be rwd, kudos to you for taking the time to document and write everything up here. Folks like you are what forums like this depend on. Appreciate the hard work.

[CTMike]

Originally Posted by Enabled

A full rubber giubo should help out too. Part number 26117610061 brings up a rubber 105mm giubo.

Checked RealOEM for my ‘15 F31 328d xdrive... it shows part number 26117605629 until 7/2015, which was then superseded by the part number above... the old part is heavier but doesn’t specifically say aluminum... is that “bad”? Is the rubber giubo and loctite something close to a fix, or is this sucker just dooooomed? Would love to tune it but not at the expense of eating transfer cases... thanks guys, great thread.

[Enabled]

The difference between the up to 7/2015 and post 7/2015 is the bolt circle diameter size.
Pre 7/2015 is 105mm, and post 7/2015 is 96mm. The door jamb date should tell you exactly, or the last 7 of your vin.

I have just ordered the 105mm Febi brand giubo in all rubber from a local friend with a Worldpac account, and I will be changing mine to all rubber.


To reiterate, this has nothing to do with power upgrades or tuning, but rather the natural vibration from the engine, and not enough absorption of those vibrations. The output flange not being pressed is a mistake in design for use with this engine. Loctite 648 on an unworn flange, plus a rubber giubo should prevent that wear from jitter/vibration at that location.

FYI, the US 535xd with 255hp/413tq uses the same exact transfer case, the ATC 35L.

[CTMike]

Great clarification, thanks!

[Enabled]

Snooped through realoem.. even the M550d (tri-turbo, 380hp/546ft-lbs) uses the ATC 35L transfer case, with the same part number as my new case installed above.

[328iX]

This doesn't really solve the whole clutch packs slipping issue, which is what happens to most of us too soon

[agentorange]

Originally Posted by Enabled

BTW, interesting find: LCI 328xd (from 07/2015 actually) and X3 28d have the 96mm rubber giubo installed at that front position. This explains the slightly different tcase part number for LCI.

This is some timely information. I've been looking to snag a X3d for my wife. Seems like a build date after 7/2015 is a must. Many thanks.