A Tale Of Two Transmissions - Airhead Style

We can't wait!

Good stuff Mark, keep us posted.




Mike

Joerg, who posts on here frequently, is kind of enough to provide the drawing on his most excellent website. There is a proviso that it not be used commercially without permissions. http://jhau.maliwi.de/mot/

His article on 5-speed overhauls is my primary guidance on this project, and deserves all the credit. I'm just having fun sharing some of the little wrinkles that come up as i learn the process.

WRT to the shim plate tool, I'm very much interested in testing the oversized bridge style depth guage attachment I mentioned above against the plate tool.

Mark:

As a favor to all of us, could you take it apart again and take pictures this time. It's important.

Tom

OK.

Alright, here's the real update for today. I'll try and keep it order, even though it only involved two bearings and some cleaning.

I started with the 6403 coming off the output shaft. Can you say "skin of the teeth" grip on the bearing to get it started?


Once I got the bearing started with a little gap under it I readjusted the jaws and pulled it right off. Heating to about 50-80C with a heat gun does indeed help slide these bearings off.

Same thing in reverse, let the shaft cool. Heat the new 6403 C3 and start it on the shaft. This time rather than wrestle with trying to keep everything in column and start on the press, I tapped the bearing with a socket/driver to get it started true and then moved to the press to drive it home.

Notice on the new bearing(SKF from Great Britain) that there is very little chamfer on the inner bore. Way less than the old FAG that came out and still less the the same bearing in the original transmission which had the circlip and spacer wire ring.


In fact the wire ring spacer would not fit under the circlip, so I ended up eliminating it and fitting only the circlip which has a perfect flush fit with no gap. Another instance of variations one can run into with transmissions from this circlip transition period.

Here is the circlip and spacer wire ring that would normally be used.


As per Joerg's advice, pull the other bearing with the first gear pinion. Again, heat to about 80 C and they slide right off. Watch for the orientation of the gear, inner bushing and washer under the gear that will stick to something and fall randomly as yo lift parts off.








Then when its time to put it back together, a few taps to get the heated bearing started true, and the press is your friend.



Everything cleaned, bearings changed, seals laid out. Everything is ready to reassemble and play the shimming game....



Some notes on arbitrary decisions made along the way... I already addressed the differences I found among the 6403 bearings and the decision to omit the wire ring spacer. No brainer - it just didn't fit with the new bearing. Coincidentally, when I pulled the old bearing off which had no circlip fitted - it wasn't that hard to pull. No extra interference fit as has been alluded to in some articles. The shifter detent roller - the original tranny had the solid aluminum one. The replacement tranny (subject of this rebuild) has the nylon roller. After spending a fair amount of time playing with both shifters in my hands and watching/feeling the actions, I elected to stick with the nylon roller. I fell the nylon has better wear characteristics and better shock absorbtion in this particualr application. For those of you who really want to switch to a SKF 688 roller bearing, they are a very common bearing in inline roller blades (along with 608 bearings of slightly different dimensions).


Progress over the next few days may be hit or miss. I'll be packing and heading to Bike Week At The Scrub on Friday thru Sunday, so my focus may more on getting my camping gear and such sorted to match the weather forecast, etc. For those that don't follow the craziness at the Scrub, keep an eye on this thread http://www.advrider.com/forums/showthread.php?t=429309

And I have no idea who that guy is in the first pic wearing a Chain Gang shirt and eating a BBQ rib

Just a bump to say this thread hasn't died on the vine. I just wrenched my back over the weekend and am getting it sorted so I can stand in one spot and bend over the work bench again. Progress will resume soon.

It's killing me every time I walk by the parts laid out neatly in a box and knowing they only need to be installed, shimmed, and tossed back into the bike.

More pics coming. Yesterday I got all the shafts/forks/shifter back in and it shifts up and down. When I post the pics soon I'll elaborate on some of the things I tried that helped the process along and some better ways to do the things I tried.

I convinced myself to veto a previous decision regarding the nylon roller for the shifter detent and swapped the aluminum roller in from the original gearbox.

Unfortunately for me, despite all the previous measuring of the Lexan plate I made my shim tool from, I rechecked it and found a thickness difference in two spots of a few thousandths right where it sits on the housing. So I finished milling the alternative bridge-style depth gauge stop for my digital caliper and just have to fit a clamp to it for locking the pieces together. Pictures will make this a lot clearer than me writing a thousand or more words. All is not lost, though. I plan to try and mill the shim tool perfectly flat so I can still compare measurements between the two types of tools. My instincts (which have been wrong before) tell me the large bridge-style depth gauge may be the better type tool. But only an actual trial will tell.

Stay tuned.

OK, drank some beer, took my nap, uploaded a few pics...


Set the shafts in place. I used a baby nose syringe loaded with water to check for when the housing reached 100C from the heat gun. A couple of drops every so often as it got close and when they sizzled - voila! I also had the two shafts in the freezer to shrink the bearings. It all slid together quicker than in the video on Joerg's webpage about assembling the parts.

However, putting the gear shafts in the fridge makes them get really wet with frost/condensation. I spent a good while warming them back up with the heat gun to make them dry off. Lesson learned: Best bet is to just chill the bearings by placing them on a Ziplock full of crushed ice rather than chilling the whole shaft(s)




Here's a trick to help you... When diddling the shifter into place and engaging the forks with their respective grooves, stuff a clean rag in to keep everything engaged until you get the bolts in.


Even though it's easy to get everything to mesh with the shifter in neutral, as I pushed and shoved to line up the bolt holes the shifter kicked itself into second gear which made it much easier to line up the bolt holes. That may or may not be just an oddity of how the gears were aligned at that moment.


Now we're talking...



Almost forgot - the detent roller swap....



Sorry for the blurry photo. Marking the bearing height to see how much material to mill out of the bridge-style depth gauge stop I have alluded to in previous posts. More detailed photos to come after I finish the clamping part and actually test it out.



Stay tuned.

Wowzer, this is great. Can you expound a bit on the "detent roller swap"?

T.

Yep. Tomorrow. I'm getting ready to pass out right now. :huh

And Hey Kevin! I see you're back home and watching this thread. Great to see ya and ride with you again at the Scrub.

Is there a specific reason you're not using Mold-a-Gauge to to measure the endplay of the shafts??

Good question. I'm not using Plasti Gauge or similar products because of the effort required to get the rear cover back off of the bearings.

I'm guessing that's the reason that normal practice is to use some type of measuring plate and do a little math.

The roller swap....

The transmission originally in the bike when I bought it had the solid aluminum roller in the shifter mechanism. At some point in time that gearbox had been opened and serviced. So I don't know if the aluminum roller was factory installed or an upgrade performed later.

The '84 transmission I bought on Ebay and overhauled had the plain nylon roller. After working them both by hand while both transmissions were apart, I planned to leave the nylon roller in place as it felt like it was ever so slightly smoother.

Then when I started reassembling everything I took one more close look at the rollers and noticed some wear on the nylon roller. The aluminum roller had no wear marks at all. So I swapped it onto the shifter mechanism.

After finding a couple of places of unequal thickness on the Lexan shim plate despite the sides both measure nearly flat, I tried to mill the offending high spots with limited success. I got it closer to parallel sides, but still not perfect and I discovered while tramming the cross slide with the milling vise mounted that i had .001" runout from front to back, but .000" along the ways. So I adjusted my cutting pattern to capitalize on the zero runout axis, but still had problems as the milling setup doesn't have to travel to cut the required area in a single pass.

So I focused on finishing the huge bridge-style depth gauge stop.


It was a cut and try design, but with a lot of thought and trial fits before marking and cutting the relief/rise part.

It does not flex or deflect with normal handling and I'm satisfied that the measurements are accurate AND repeatable.

First off measuring the bearing heights above the flange. Because the bridge was designed (intended) to have a 20.00mm height above the working surface, the math should have been easier. In practice I could only index it securely at 19.93mm. Still easy math.




Then take direct measurements of the bearing pockets in the rear cover.
This is actually a staged photo after I had calculated and put two shim stack into the housing with tape to avoid accidentally spilling tthem back out. I have to run over to the dealer tomorrow to pick up the one odd size shim I need (0.200mm)




The math is simplified with this bridge - measure the bearings above the housing fange and subtract from the baseline setting (in this case 19.93mm). Measure the depth of bearing pockets in the cover and subtract the baseline 19.93mm from those numbers. Calculate for gasket thickness and desired clearance of .05mm and you easily determine the shim stack needed.

So far so good.

Actually, I think the reason for using a measuring plate is the time involved, not necessarily the effort.

The reason I ask is, whilst your measurements and setup--which is VERY NICE--will work for your rear cover, what about a rear cover from a different tranny??

Given the extremely tight tolerances spec'd for the shafts, I am thinking that there might be significantly more variance 'tiwxt covers and measuring devices, which might be why a lot of the learned will tell you that BMW never sent out a properly shimmed tranny from the factory. Factor in some measurement uncertaintity, and you will go well beyond the allowable tolerance.

My thinking is that by using the Mold-a-Gauge you tailor the shims to each sceanrio--including any gasket squish--independent of any extraneous variables, which, while more time consuming, might just be closer to spec in the end.

Thoughts??

kix

I just wanted to explicitly say that you have really done an excellent job w/ your setup.