Reassembly of Unimog 404 portal boxes

[Bohalrantipol]

I have been slowly getting my 1968 unimog 404 hard cab back on the road, and one of the issues I was addressing was the sludge and muck that was in the portal boxes when I first tried to change their oil. Suffice to say, all the seals were clearly the original and all of the bearings had suffered for all the crap they had been letting in. Also, the cardan joint looked like it had been encased in grease. So, I decided to replace all of the bearings and seals on all of the portal boxes.

 

[Bohalrantipol]

First there are several places that also talk about rebuilding 404 portals here:

Rebuilding 404 Portal axles!
Replacing Axle and Hub seals in a Unimog 404.1


While these cover some of the rebuild there were several points that held me up. So, I wanted to write up a little explanation of what I did for reassembly. Use this advice at your own risk, but I was pretty happy with the process.

The most important tidbit I found regarding disassembly was that in some earlier unimog axles if you do not engage the locking differential the slider would fall into the axle when you take the axle out of the housing. That would have been a real PIA, so thanks internet for that one.

 

[Bohalrantipol]

First thing was to buy all of the seals and bearings. Most of these are available from retailers, but that may change in the future.

The bearings I used were:

Back Top: 30309

Back Bottom: NJ210EG15

Front Top:30306

Front Bottom:C36215


There are some variants of the above that are also probably acceptable. I tended to stick with brands I have had good luck with (SKF, FAG, and TIMKIN).

With these in hand I began to reassemble.

First thing to note is that the only precise bearing freeplay in the portal needs to be on the two taper bearings on the top. So if any dimensions are adjusted between the front of the housing and the rear seal holder then you are supposed to check and adjust this distance. In one version of the Unimog manuals I have they go so far as to say that just replacing the seal requires you to measure and use a similar thickness gasket to not throw off this measurement.

Well I was replacing all the bearings, all the seals…..

So they want you to use fancy Mercedes tool number 401 589 00 21 00 to find out what the total height of your input shaft is from the back of the rear portal housing in order to determine what shim is required under the front bearing race to establish the correct "bearing free play". Clear?

So I got it out of my tool box……


No, no I did not. Who has one of those? Anybody? I actually looked around thinking maybe someone somewhere had one or a reproduction. Could not find it.

What I did was to reassemble the input shaft with the two bearings and portal gear.

Then I put a new gasket with some Curil T2 coated on both surfaces of the rear seal carrier, and bolted that down.

I then inserted the front race into the front portal housing without any shims, and the rear race until it firmly butted up against the rear seal carrier

It’s also interesting to note, that if you actually have that fancy Mercedes tool you also need the depth of the front portal housing to figure out what the "bearing free play" number will be. Stamped on each casting is the measurement from the factory right below the bearing race to be shimmed.

That is a little hard to see, but it is a “50” indicating the total depth is 78.50mm. Anyway, that’s totally useless to me, but interesting.


So, I carefully inserted the input shaft into its bearing race on the rear, through the rear seal (using some grease to prevent damage to this), and installed the main housing gasket.

I then torqued the front housing on to the second “uuungh”, flipped it over and slapped my dial indicator on it.

Then I jiggled it up and down to see what the lash was. The "bearing freeplay" number you are looking for is around 0.08 +/- 0.02 mm. I was around 0.15mm so: 0.15 – 0.08 = 0.07mm. So I needed to add a 0.05mm and 0.02mm shims to get to the prescribed "bearing freeplay" on these bearings. Coincidentally the only ones I had! What a life. So I pulled it apart and added the shims, and permanently installed the front bearing race in the front housing.

 

[bdkw1]

Uhhhhh, if you were measuring the freeplay with the indicator, to get .08 of pre-load with .15 of freeplay you would need .23 of shim.

Right now you have .08 of freeplay and zero pre-load.

 

[bdkw1]

Also, a wide base dept mic would do the same thing as the mercdes tool.

 

[Bohalrantipol]

Uhhhhh, if you were measuring the freeplay with the indicator, to get .08 of pre-load with .15 of freeplay you would need .23 of shim.

Right now you have .08 of freeplay and zero pre-load.

Sorry I misspoke "bearing freeplay" is what you are establishing. fixed

Also, a wide base dept mic would do the same thing as the mercdes tool.

The Mercedes tool is a fixture that is bolted to the four retaining studs and uses the gasket as its base, so that wouldnt help unfortunately, I tried.

 

[Bohalrantipol]

So the next step was to complete the reassembly. The manual has you assembling the entire portal box with axle and cardan joint and then sliding it back into the axle housing. Well this thing weighs a decent amount, and I wasn’t looking forward to man-handling it back in there. You can just assemble the back half of the portal box and then reattach it to the axle, but that carries the possibility that the input axle in the portal will slide around ruining the rear seal. So I farted around on my computer a little bit and came up with this “fixture”.

And then 3D printed it.

What this effectively does is hold the input axle in place so you can easily reassemble the cardan joint (a press is a requirement for this) without screwing up the rear seal. Now I originally thought maybe I could use something like this to measure the bearing freeplay too, but ultimately it is plastic, and I just wasn’t so confident that would give me reliable results.

So then with this fixture attached I added the second lower bearing to the back portal housing and then pressed on the first half of the cardan joint.

Then assembled the axle housing side of the cardan joint and put the two halves together

Then it was very easy to slide it back into the axle housing

I then installed the main output bearing and seal in the front of the portal housing and pressed in the output shaft/gear after THOUROUGHLY wire wheeling it and coating it in some silver permatex anti-seize. I am hoping if I ever need to remove this again the anti-seize will stop that from being another cursefest.

And then pressed on the output flange and bolted up the remainder of the portal box to the Unimog with the same gasket I used to establish the bearing freeplay, along with a thin coating of curil T2 to “seal the deal”.

Another deviation I made from the manual-dictated assembly is that I decided to use torx screws on the bearing caps for the cardan joint. Instead of then staking the screw I used some blue Loctite. I don’t think I would “recommend” that to anyone else, but time will tell if it holds up for me.

I also then sprayed some lanolin based protectant on the inside of the output shaft, where the main bolt goes. When I get the brakes back together and a tire on the thing I will then torque that to the specified “keep going until you feel your phalanges pop”. I think my torque wrench goes that high.

 

[TimJus]

there must be some mistake about bearing
setting.
you would never leave free play in tapered roller bearings. you need preload the assembly or gears will get focked up

 

[Bohalrantipol]

I am with you, thats why my brain keeps going to "preload".

 

[bdkw1]

Interesting, I've never seen play recommend for taper roller bearings.

 

[MoparMoe]

I worked in an industrial gearbox shop for a bit, they are running a clearance to accommodate linear expansion of the shaft.

If you get them a touch too tight you will pay , trust me . .08mm (.003 ) clearance is similar to what industrial shaft mount assemblies run.

 

[Bohalrantipol]

I worked in an industrial gearbox shop for a bit, they are running a clearance to accommodate linear expansion of the shaft.

If you get them a touch too tight you will pay , trust me . .08mm (.003 ) clearance is similar to what industrial shaft mount assemblies run.

I had to figure that Mercedes has some expertise here. Thanks for the information.

 

[TimJus]

it would boil the oil out of box before linear expansion would cover the .08mm clearance in the upper gear.

 

[bdkw1]

it would boil the oil out of box before linear expansion would cover the .08mm clearance in the upper gear.

This ^^^^

Pinion bearings see similar loads and higher speeds. They live just fine with preload.

 

[MoparMoe]

This ^^^^

Pinion bearings see similar loads and higher speeds. They live just fine with preload.

The cones are facing each other in a pinion , its a different application.
the preload is to combat the expansion.

 

[lt1yj]

First off thanks for an awesome write up!

I agree that endplay on tapered roller bearings is asking for failure.

I took Timkens bearing design class a couple times back in the early 2000's. They stressed that tapered roller bearings need preload on initial startup or the rollers would skid on the races and create rapid wear. They passed around several tapered roller bearings that had started to fail and it was easy to see wear on the cage, the ends of the rollers and the surface of both races.

The skidding happens when the roller leaves contact with the outer race and the cage forces the roller around the bearing rather than rolling on the outer race. When the roller comes back into contact the roller isn't necessarily in line with the race and it has to slide back into position and to get back up to speed with the other race. The larger the clearance the lower the number of rollers are in contact and the more sliding happens. Higher RPM's cause even faster wear.

I can see minimizing the preload but I can't see how tapered roller bearings can be made to work with endplay.

I think .008" preload would be way too high for those bearings as well. .0005-.0015" is more what I'd expect. Maybe up to .0025" if the housing is flexible enough.

 

[Bohalrantipol]

I appreciate everyone's sentiments. My previous experiences have also led me to always assume that tapered roller bearings require preload, however pretty much every one of those experiences were from obvious examples of why and how that preload impacts the overall assembly. These portal boxes are a little unusual because of the manner in which the loads are being transferred, and because of this I have to assume Mercedes knew what they were doing. Also, examining the Timken manual on setting tapered roller bearings they indicate you can establish a freeplay without major changes to the life of the bearing (see Figure 2).

Timken manual

 

[bdkw1]

I think .008" preload would be way too high for those bearings as well. .0005-.0015" is more what I'd expect. Maybe up to .0025" if the housing is flexible enough.

We're talking metric here. .08MM=.003

 

[MoparMoe]

Additionally, I currently work in a Papermill service job shop, we rebuild reel spool rolls ( they receive the paper that is produced on the machine ) that run Timken cup and cone systems these are assembled with .012 of clearance, the preload / clearance discussion is based on application. do what Merc says and you'll be fine.