3.06 ATB Differential
 

After recently winning the bid on a Quaife QDF5V ATB differential, I broke my 3.06 open diff down in preparation. This will be going into my 90 300CE, which currently has 3.27 gears. The donor 3.06 is from my 91 CE parts car. The Quaife is currently being shipped so as progress happens on the diff I'll post pics and update the post.


The breakdown is pretty straightforward. I'm not messing with the pinion other than to replace the seal so back cover off, clips out and hubs out. Unbolt and remove the side bearing carriers and the open differential falls out. After the bolts are out the ring gear, a little heat applied and the ring falls right off.

I'm a bit nervous about setup in that special shims are used to set the carrier in position. Most of these shims are NLA from Mercedes so anybody with spares give a shout out. I've got 1.13mm and 1.93mm and in a perfect world they would get the job done. I'm planning to replace the side bearings. They're not cheap but without the special tool to pull them it doesn't seem worth the effort. So considering it will have new bearings and a carrier manufactured by a 3rd party, it will be a miracle if my original shims get it setup right.


Here's the beast stripped down:

Despite tracking showing a sorting error, the differential arrived a day early. :eek: It's a nice unit but for the outlay expectations are high. Not disappointed.

Side bearings are on order but are still a couple days out. Setup doesn't begin until they're in and pressed on. Circlips to retain the hubs arrived today as well. As advised the hubs will need a slight trim on the lathe to give the circlips enough depth to work in.

Ain't she pretty?

Following!


Sent from my iPhone using Tapatalk

I have to go to work now but I will try to get back to this tonight. I may be able to help with shims and side bearings.

If this is a 210mm diff I have 6 sizes of side case shims.

1.526
1.202
1.863
1.908
1.490
1.509

I guess what you are calling the hubs are what I call the axle flanges. I also have 6 sizes of axle flange rings. No need to put your axle flanges on the lathe. These rings are also selectable and I have 6.

3.072
2.928
2.797
3.110
3.103
3.019

No sense in getting MB side bearings as they use standard Timken available everywhere. You will probably have to buy the race and cups separately.

As far as removing side cover bearings the OTC puller 4534 will do the job with a few minor modifications.

I had to make a new center ring with studs to pull the jaws in closer or spread them out in the case of the outer races. See the first post in my differential tool build ideas showing this puller being used to pull the races out of the side covers.

Let me know if you need anything I can work something out.

The bearings came in today and I got them pressed on. This amount of progress allowed me the first opportunity to measure for setup beginning with case spread. The spec is to have .15-.20mm of spread. This is essentially the preload on the side bearings. I stuck my original shims on the side bearing carriers and torqued them into the case. Instead of .15mm preload, I had .15mm clearance. I managed to score another two shims from a junk diff, one of which is 1.0mm and swapping that one for my 1.15mm got the preload to 0.0mm. If I can locate another shim in the 1.8 to 1.85 range I can get the preload set and then add the ring gear to the mix. Hopefully backlash ends up ok after splitting the difference in preload between sides.

Looks like you were typing while I was typing.


I went ahead and ordered the bearings from Mercedes. They weren't exactly cheap but rather than hope I cross referenced them right I bit the bullet.


I'm very interested in the 1.8xx shim though once I mount the ring gear my backlash could be way the hell out. I could still be futzing around to find a combo of shims that measure 2.8xxmm collectively that gets the ring where it needs to be. At least I know what backlash needs to be. The number is stamped on the pinion gear. Mine says .12 and I measured .13

The lathe work for the hubs is just to deepen the groove that the original retaining circlip held. The groove gets cut a little deeper allowing later model captured circlips to be installed. As it stands the replacement circlips sit a bit proud of the root of the splines and they need enough room to compress under that dimension.

Once you get the preload correct with some amount of backlash at that point the total thickness of shims remains the same and it becomes a mater of decreasing thickness on one side and increasing by the same amount on the other. Yes you need an unlimited supply of shims and a surface grinder is a huge help.

Well, after J. Roncallo shipped me his collection of shims I went ahead and bolted the ring gear to the carrier. With it installed the closest shim combinations were tried. As luck would have it, my closest combinations were either too loose or too tight. 1.8 and .95 mm shim combo doesn't quite hit the spread spec and ends up .21 backlash. Next closest is 1.45 and 1.3mm and they end up in the negative backlash range and pretty much bind the assembly. If I could locate a 1.7 to use in place of the 1.8 it likely will be golden. The search is on. There are a couple shops local that may have a pile of old diffs around. Worst case scenario I'll have to get one of the thicker shims ground.

There is a highly respected old timer in the Benz world that has a small shop nearby so I rang him up today. Turns out he had a significant collection of shims and was gracious enough to let me leave with the lot of them. A bunch of measuring, disassembling, re-measuring and disassembling again has yielded what I believe to be a well set up differential. I ended up with a 1.6 mm and a 1.1 mm shim to get 0.2 mm case spread and 0.11 mm backlash. The case spread is the tight side of the spec but considering I have new bearings I was hoping to hit the tight side. The backlash spec for this matched ring and pinion is 0.12 mm, but I feel ok with 0.11 mm.

The old timer offered to double check my setup with the rarer than hens teeth official Merecedes differential setup tools when I return the shims. Hopefully his numbers match mine as even in his collection, there were no 1.65 or 1.7 mm shims.

Actually, the +.12 that is etched into the face of the pinion is the variance from the "zero" position, and is used in computing the thickness of the shim that adjusts the pinion depth.
The good news is that by chance ".12" is within the range of .08-.14mm that is called out for the backlash, and so, in this case did not cause an error.

Good catch, I just glanced over most of the pinion stuff in the service manual. I feel even better about that 0.11 now.

Good to see you got it doing finding a guy like that with a stash of shims is a big help.

PM me the old timers name. I started working on Benz's in the late 70's in Denver maybe its my old boss.

I had the chance today to have the setup doublechecked with the official setup tools today. Where I had previously measured 0.11mm backlash, it was a bit tighter, 0.08mm to be exact. The spec is 0.08 to 0.14 so I'm really glad I had the opportunity to have it double checked. Had it been any tighter I'd still have thought it ok. The rotational torque was also measured and it too came in on the tight side but still in spec. With everything measuring up I went ahead and installed the seals and buttoned it back together.

Actual installation will be a while yet. I'm planning to do a complete subframe swap and not just a diff swap. I have a V8 subframe and matching back cover for the diff. The V8 subframe has larger diff mounting bushings that are spaced farther apart than the 6 cyl variant. I also think my current subframe may be tweaked. The right rear camber is a bit high and there is a some of bodywork in that corner. So it's gonna be a multiple birds / one stone kind of thing.

What made you decide to go to a 3.06 over keeping the original 3.27.

I'm currently running the 2.47 had a 2.65 in my M120 560.

Fuel mileage performance was pretty profound increase with the 2.47. But it does feel like its dragging a boat anchor after driving it with the 2.65. And it did loose about 1/2 sec in the 1/4 mile despite being an LSD differential, where as the 2.65 had a spring loaded phantom grip and still spun one wheel. The numbers below I actually find a bit unbelievable, so I want to do more testing before I swap in a 2.82.

Greensboro NC to Athens GA 297 miles, 12.539 gal, 23.7 mpg, average speed ~ 72 MPH, top down.

Athens GA to Atlanta GA 211 miles, 13.504 gal, 15.6 mpg, average speed, first half ~ 60mph top down, second half ~90mph top up.

Atlanta GA to Summerfield NC 331 miles, 13.072 gal, 25.3 mpg, average speed ~ 72 mph, top mostly up.

Quaife only fits up to the 3.06 gears. Then the carrier switches to a different offset for the ring gear to align with the pinion. Stupid Mercedes way of setting up differentials.

How are you going to handle the change in wheel speed signal ratio to the trans computer?

Diffs ( non brand specific ) with a wide range of ratios have always moved the ring gear mounting surface but not for every ratio like MB. An example of a USA unit would be 3.27 and smaller number is one carrier and 3.45 and larger another.

In order to accomplish this wide range, the ring gear must be thicker as the numbers get larger. This leads to more rotating mass / material used. MB kept the ring gear thickness the same across all ratios and just moved the flange. We also don't know if there was a thickness limitation in their gear cutting machines when this was instituted.

While this causes issues for the microscopic number of people that swap ratios, it works out better for the 99.99999% that drive the car and not change things.

I use tone wheels bolted to my axle flanges. Changing the number of teeth on the tone wheels fools the trans control module. Been working fine for about 6 months now.

And I guess Quaife doesn't make one for the 3.27. what about Wavetec.

As stated by others, it was the only choice. I built the diff from my 91 300CE and that car never felt over geared. To be honest though, that car had the best off idle torque of any 104.980 I've ever driven, and that's more than a few. Don't really know why, and never put much energy into finding out. The '90 CE has the '91's fuel distributor and EHA, its injectors, and its EZL, but the '90 still doesn't pull the same off idle, not even close. Long story short, in its current form its not a really fast car, so a bit of gearing change wont be a big deal. The '91 met its demise in a full throttle turn under adverse traction conditions when one wheel spun up and the other didn't have enough grip to keep from spinning out. If it had a limited slip diff, I doubt that would have happened.

A limited slip might have helped as long as both tires don't break traction, however, once traction is lost the tail will come around. RE the loss of traction limit with a LSD might have been higher and more severe due to the loss of lateral grip where an open diffs loss of traction is lower and more stable due to a loss of forward motion.

A limited slip tends to break traction of the inside wheel during a turn under power. ( Think a solid non diff axle / go kart ) This is usually OK unless the outside wheel has limited traction. I'm betting this is why MB decided to use ASR rather than a LSD. Going the ASR route also gives the possibility of ESP.

Will the change from 2.47 to 2.82 result in an even tone wheel tooth count change?

2.65 is the baseline with a 48 tooth wheel.

To use a 2.47 ratio multiply tooth count by 2.65/2.47 = 48*2.65/2.47 = 51.49 Round to 51.

To use a 2.82 ratio multiply tooth count by 2.65/2.82 = 48*2.65/2.82 = 45.11 Round to 45.

Didn't think this would work but it does. Still looking for any adverse issues at this time but none so far.

2.47 is about the worst case scenario. I could probably use a 51 or 52 tooth wheel. That is one of the reasons I elected to try it first.

In another thread I recall whipplem104 saying that the room for wheel speed error is very small, looks like there is some room.

I don't know what the threshold is but 1% appears to be acceptable, 7% is not acceptable.

% error threshold is somewhat vehicle speed dependent. At low speeds in low gear you can have more slip from driveline components and resolution. And your rpm differential is low. At high speeds you may find that error of a small percent is large enough to trip a fault. I do not think 1% will ever throw a fault. I do not know the exact number or actual fault threshold. But I would say probably up to 3% would work but maybe not sustained. 1-2% is pretty typical variance in meaurement though.

1% seems to work well for at least up to 100MPH at this point, with at least 10 min sustained speeds of 90MPH last weekend and has worked well since September.

I dug out my V8 rear diff cover and snapped a couple of comparison pics for those unfamiliar with the difference. The V8 part has a 129 part number. It also has cooling fins and vertical internal baffling in addition to mounting points spread several inches further apart compared to the 6cyl part.

After cleaning it up good I ran a bead of sealer and bolted it up. I also cleaned up the ABS sensor bracketry and bolted that to the housing. At this point it is finished. Still no Idea when I'll get it swapped in with a V8 rear subframe. Rear lower ball joints are also on the to do list so that is a whole bunch of labor added all up. Just in time to loose one of my weekend days to the farmers market my nursery sells at starting this weekend. Not enough hours in the day or days in the week.:rolleyes:

Just to round out the thread , please post the part numbers on the covers when you get a chance.

Please do elaborate with pictures. How id ABS handled on this car. Is this a 4 channel ABS or 3 channel. Is ABS on the input or output of the diff. Can you post pictures of the brackets.

This car only has one sensor on the pinion. The previously mentioned bracket is pretty much just a metal shield covering the sensor keeping random road debris from causing havoc.

The old ABS cars dont care if you change the differential because they changed the tooth count on the pinion. So you just swap a whole differential in and done.

So it bolts into the side of the case like the 107 but has an additional shield.

Those were the days.

Here's a pic of the cover/bracket. Also a better picture showing the cooling fins on the V8 diff cover.
As requested the diff cover part number 129-351-16-08

After putting it off for too long, I finally got the diff installed with the V8 subframe. It was a straightforward swap. The only discrepancy was due to the V8 subframe being for dual channel ABS and my car is single channel. This leads to the bracket that routes the ABS sensor's cable not lining up. Nothing a zip tie couldn't fix.


First pic is the V8 part ready to install. Second pic is the job ahead. Its a lot of work to pull a complete subframe. 3rd pic is the undercarriage with the subframe removed. 4th pic is the rear lower ball joints. These were particularly stubborn to press out, even with the special tool. Given the poor shape they were in, it was a must do while everything was apart.


After buttoning everything up I put it up on the alignment rack hoping to just square up the rear toe. Nothing is ever that easy, and a full alignment ensued. The right rear camber was too negative before, and it remains so. I suspect its an artifact of the bodywork in the right rear quarter. I had hoped that the subframe swap would have resolved it. Apparently it will require a spindle. Not a priority since it isn't wearing the tire bad from it.


A quick test drive did reveal a leak and on closer inspection the left stub axle had pulled out enough to clear the seal. After popping it back in it felt solid against the circlip so I went ahead and drove it home. Still ok so I'm not worrying about it but will keep an eye on it. I suspect that during assembly I may have popped it loose. I installed the spindle, links, and spring as an assembly so it was like wresting an octopus. I could have easily pulled the axle enough to compress the clip without realizing it.:rolleyes:

The good news is that it's dead quiet at speed. I can already tell the difference in the mounting. The old rear bushings were on their way out and I can feel that its tighter now on hard shifts even though I didn't do any real beating on it yet.


FWIW, I've got a 3.27 ratio diff for sale if anyone wants to make an offer.

Thought I'd post a pic of the old subframe. Compared to the pic of the V8 subframe above you can see the difference in the rear mounting. Another big difference is that the front cross bar is stamped sheet but on the V8 part it is tubular.


Still haven't beat on it too much yet but it does give me a bit more confidence in full throttle kickdown in hard turns. :D

After about 600 miles I've pulled the diff back out. There was a slight hissing sound coming from the left side bearing. Couldn't hear it in the car but with the windows down, driving past a fence or like, I could hear a noise reflected that didn't sit well. Running it up to speed on the lift it was clear the noise was mainly the left side bearing and maybe a little on the right.


So today I swapped it back out for the 3.27 while I get the LSD sorted..... again.:rolleyes: So upon disassembly it looks like the preload was too tight on the side bearings. Upon inspection you can see fine scoring of the race with pitting just starting. Damn! I consistently measured 0.19 to 0.20 mm for case spread and the spec is 0.10 to 0.20 mm. I figured that the tight side of the spec would be ok. I figured wrong. My next move here is to go 0.05mm thicker on both side shims to put the case spread on the low end of the spec while keeping the backlash where its at.


I would have set it a little looser if I had shims in the range needed to do it. My selection stepped from 1.65 to 1.8mm At this point it looks like I'll have to get a thicker shim and have it ground to 1.7mm. The right side shim will have to go from 1.05 to 1.1, which I recall was available in my selection. This is assuming that the replacement bearings install to the same dimensions these last ones did. One odd thing I noted was that the wear path between the left and right sides were in different spots in the races. Given that, I'm gonna have to check and re-check everything again on the next assembly. Nothing will be assumed. I'm specifically not looking for a charming 3rd try.:D

I received and installed the new set of side bearings. Measured the setup again and it was still pretty much spot on. No change from installing the new bearings. A revelation did occur during measuring though. The screw in tip of the dial indicator was missing. It screws in to the indicator and was nowhere to be found. I have a sneaking suspicion it came out during my last measurements and caused the trouble with the first set of bearings.


I had previously misspoke about the case spread spec. It is 0.15 to 0.20 mm, where I had previously thought the lower limit was 0.10 mm. My setup measured 0.18-0.19 mm which is about perfect. When ordering my bearings I also ordered a 1.70 mm side shim and was astounded that one came in today. I swapped it for the 1.65 on the left side, which puts the spread a bit under spec and adds a couple hundredths mm backlash. Also ok as backlash was at the tight side of spec.

I hung the diff this evening and ran it a few miles on the lift. The noise is still present, though to a lesser degree, and seemed to diminish a little after a few miles. Gonna have to keep an eye on it. If it doesn't get any worse, I can live with it. If it breaks in and diminishes more, so much the better. I'm gonna sample the oil after a few hundred miles. If its still clean, then I'm calling it good.

The missing tip shouldn't be a problem,the ball end is there more or less to protect the surface that is being measured , prevent the indicator plunger from wearing , allow for extensions to be added. A square plunger end " could " affect measurements if off angle to the measured surface but for what you are doing I'd expect the indicator to have more internal error.



After the change out, the ring and pinion are going to be running in a slightly different orientation / using areas previously not in contact so it may take some time to wear off high spots

The problem with the missing tip isn't measurement error. I think it may have been eaten by the diff. :eek: The first time I dropped the oil out the diff it was quite grey. A couple hundred miles later when I finally pulled the diff back out, that oil came out as clean as it went in.

Is this a clutch type diff and if so, are the clutches bare steel or coated with something? If coated, perhaps the gray was the initial break in.

The indicator tip's ball is going to be carbide or similar and the housing mild steel. I'm not expecting it to cause a problem even though I would not knowingly drop it in the diff.

If in the diff, it isn't going to be ground to nothing. At minimum it would be laying at the bottom or dented from one go round the gears. At max the housing would be hammered flat ( like engine bearings that have pounded to flakes ) and the carbide ball denting the ring and pinion.

Not clutch, its a torsen.

I cant see how an indicator tip could get into the bearings at least not in the race. It could get in the gear mesh but that would be obvious score on a gear tooth. Its possible its wedged in somewhere where its just rubbing some rotating member.

When you go to change the shims, remove the carrier completely and drag a coat hanger under the pinion. If I recall there is a drain back passage in the bottom of the diff. Also it you have a speed sensor pull that out and have a look.

I did inspect the ring very closely. There is 3 very faint marks on it. None of which give me any pause about the serviceability of the part. They could be orders of magnitude worse before worrying about them.

I did note a bit of magnetic foreign material on the tip of the speed sensor which is fairly strongly magnetic. I didn't notice it until after pulling the sensor out and fishing it over the top of the diff and out of the way. That leaves the opportunity to have picked up some other foreign debris, but unlikely.


If I had to describe the noise, it would be imagine a diff assembled dry and not filled with lube. Then imagine how it would sound before the bearings heated up and self destructed. Kind of a hiss. Its not real loud, you can't hear it in the cabin with the windows up. The noise doesn't travel through the chassis. Its unrelated to load at all.


I'm thinking to hook up a chassis ear to a pico scope and get a few captures. Being able to visualize the sounds may shed some light on it.

Sounds like pinion bearings from the description.

{Fingers in ears} La La La La La!



After pricing pinion bearings today I went ahead and ordered a set in. I really, really, reeeeally didn't want to have to futz with the pinion. It's just a simple carrier swap after all... right? Oh well, what's another couple hundred bucks? :rolleyes:

Good luck with this. When you put it back together you can put temperature labels near all the bearings. Here's a picture of two mounter nearest the pinion bearing.

These are available for about $5.00 + $5.00 shipping from McMaster. My car runs about 180°F on a long 7 hr cross country drive in 90°F ambient.

Have a look at the vids in this thread.

https://www.peachparts.com/shopforum/mercedes-benz-performance-paddock/395119-differential-cover-testing-gale-banks.html