A Woodworker’s Guide to Gearbox Rebuilding

[Zombie Master]

This is excellent!

[Airbear]

Thanks, you blokes. Yep, gearbox fixing has always been portrayed as a dark art. On other forums I have seen reports of people paying or being quoted huge numbers of dollars or pounds to have experts fix their gearboxes, yet like everything else BMW most things can be done by amateurs with reasonable success, provided there is suitable information available. The internet is still a wonder to me.

These bikes have ways of getting their riders to know everything they need to know ...

[Airbear]

An ‘oops’. I probably should have shown this following bit before the previous bit, since getting the bearing off the back end of the input shaft needs to be done before rebuilding. A Senior Moment, I think such things are called. I have just become a great grandpa, y'know. Anyway …

Next was removing and replacing the other bearings. By the way, I washed and spun all the original bearings while they were still in place on their shafts. Most felt really good but a couple were obviously sloppy and worn in both boxes – I don’t recall which ones now.

Pulling the bearings using the puller set. I used a flanged nut from the junk box under the pointy end of the cranking bolt to prevent damage to shaft or tool …



… and set it up in the same inverted way in the vice, so I could catch the toothy bits in my hand as the bearing was pulled off …



… and pressing on the new bearing. Here I am using a socket to push on the inner race, squeezing it to its stop …



With the five new ball bearings in place it is time to move everything into the warm house and onto the kitchen table (yep, I share my house with only a cat, who cares little about such niceties). It is time to start re-installing shafts and getting ready for shimming.

My first attempt at installing the output and lay shafts didn’t go very well. Firstly, I had heated the case to install the outer race of the input shaft roller bearing, then put the case open side up in the gas oven to heat. The shafts I had put into the freezer a couple of hours before. Now, a thing about ovens that are not fan-forced is that there is a lot more heat at the top than the bottom. So I was checking the temperature with the infra-red thermometer and it was showing a nice 100C in the middle of the case. I took it out, set it on wood blocks and dropped the shafts in. Almost immediately they jammed in their pockets, slightly canted to the side and would not move from there. I checked the temp and found that the back end of the case was between 50 and 60C – not hot enough to get the required expansion. So, back to the heat gun. I got the shafts back out and back into the freezer. Had a break – coffee, food, beer, whisky, joints etc.

A couple of hours later the case went back into the oven, this time inverted (open side down).



A few minutes later I heard a clunk – the outer shell for the input shaft roller bearing had dropped out of its pocket. I retrieved that and put it safely aside. With the case at the proper temperature I sat it upright on wood blocks, dropped in the input shaft bearing and quickly followed with the cold output and lay shafts, which dropped beautifully into place with satisfying clunks. It really is good fun when things go well. Grin.

Edit: Another senior moment - when fitting the output and lay shaft clusters one must also have the shifting fork slide onto the fixed shaft at the same time, as in this pic ...



It can be tricky. The case is hot enough to burn skin so you are holding the clusters and fork together in heavy gloves and trying to align the three and lower them quickly into place. End edit.

So now, with everything at room temp, including my IQ, it was time to think about shimming.

To be continued …

[Airbear]

So, the fun bit – shimming. Once I got my head around this it really was a joy. Ok, I already owned a decent 0 – 25mm micrometer and Jeff lent me a depth micrometer and the measuring plate. FYI, a reasonable depth micrometer can be had for around $80 and the measuring plate is available from Cycleworks for US$88.

This is what we need to figure out. I hope this diagram makes it clear enough (in conjunction with the spreadsheet below) …



Now, I have been teaching myself spreadsheeting (I think that is a verb nowadays) and this seemed like a good opportunity to do some practice. Got my head around averaging and simple subtractions and came up with something that worked (this is the result sheet):



So, I was comfortably set up in a warm kitchen (as god intended) it was time to start measuring. Note the plate on top of the lay shaft. There is one at the bottom end too - yes, you have to put everything back where you found it. I took a few minutes to familiarise myself with the depth micrometer – the numbers are sort of in reverse but it all makes sense.



I took 6 measurements around the circumference of each bearing, filling in the spreadsheet as I went. Did the same with the pockets on the cover, which I clamped by its ‘horns’ in a little drilling table vice …



Filled in the spreadsheet and it provided the shimming thicknesses I needed for each shaft. Then I set to measuring shims. This was interesting – there was so much variation, even with the new shims from Motobins. They all had surface ‘scale’ on them and could vary by up to 0.04mm, often just in one small area. Another whisky or three and a bit of rubbing with some 800 grit wet’n’dry and they were more predictably measurable. It was helpful to clamp the micrometer (gently) to a block to make measuring easier …



To be continued ...

[Rob]

Whiskey! Now I know I am out of my league.

I think my I.Q. i closer to room temp. where you live, than where I do, Charlie.

[Major Softie]

Whiskey! Now I know I am out of my league.

Yes, but he just got Veg's and my attention.

I plan on copying the whole thread when it's done. Like all of us that have been around here awhile, I've stopped trusting that the page will be available if I just bookmark it.

[Bob Bennett]

I plan on copying the whole thread when it's done. Like all of us that have been around here awhile, I've stopped trusting that the page will be available if I just bookmark it.[/quote]

Good idea, Thanks Major.

Regards Bob.

[sprints@pldi.net]

Thank you so much Charlie - I've just begun a rebuild on my /5 4-speed so this is perfect timing. I feel like I know what I'm in for now. Keep up the good information - what a reference tool - Rod

[Airbear]

Thanks gents.
I've been writing the text in Word and adding the images as I go. I can publish this as a PDF - it'll be about 3MB, and the spreadsheet is about 35KB. It would be good to have a website where folks could download this sort of stuff as required. Anyone got one? I'll be happy to send these to interested inmates via email once it's finished.

Rod, have fun. And post what you find. I imagine there will be differences between the 4 and 5 speed boxes.

[Airbear]

A little diversion:

Testing and selecting the best shifting mechanism was part of the process along the way. I found it best to mount each of the shifters that I had onto a bit of wood, fit a gear lever, clamp it down solidly and run it through the gears.



It is a beautiful little machine in its own right. Moving the gear lever pulls or pushes a pawl which has a sort of hook at each end. This engages with four little posts – you can see their riveted ends in the big cam plate on the right – and the plates rotate, causing the shifting forks to slide into the different arrangements required to engage and disengage gears. Great fun to play with. I chose the better of the two I had – it just felt more solid and the little white wheel snicked more nicely into its notches. I’ve read that the white wheel can be replaced with a ball bearing quite inexpensively and I’d like to try that one day. The other matter is the notorious pawl spring. When this spring breaks the pawl drops down under gravity and will no longer engage with the posts on the cam plate and your box is stuck in whatever gear you were in when it broke. I would have replaced this but Motobins had no stock at the time I was ordering parts.

Other matters: I had been wondering when one should install the new seals. Could they be fitted before applying all that heat? The answer is yes, I found, though if you are heating with a gas torch I think the answer may well be NO unless you are really careful. I had no problem, and it is certainly easier to fit the input and output shaft seals while the box is disassembled.

Back to the box on the kitchen table, which is now almost complete. With the measuring done it is time to select the best arrangement of shims and put the cover on. This is the range of shims I had to play with …



… and it was simple enough to select the best combinations to achieve the correct axial clearance for each shaft. I put a smear of gear oil on each one and sat them in place on top of the bearings. The cover went into the oven until the temp reached 100C on the thermometer gun. Glove on, grab cover and drop it on. Give it a gentle tap with a hammer on a block of wood. Quickly fit screws (I used an Allen bit in my cordless drill) and torque them down. Turn the box over and give the input shaft a similar gentle tap. Spin the input shaft with your fingers – how does it feel?
Mine felt great at that stage, but …

The following morning I was troubled, because the input shaft did not spin freely at all. I spoke to Jeff, who felt that this was a good excuse for a ride. He turned up at mine and we pondered the likelihood that it might ease in use but came to the conclusion that there was sufficient doubt …

The upshot was that I pulled the cover off again. This time I was more careful about bolting the measuring plate down to the specified torque (I had guessed before and was clearly wrong). I filled out a new spreadsheet and found that the new shim measurements were 0.03mm less than what I had found previously. I selected new shim sets, put it back together and went to bed. Next morning, with ambient temperature at 11C, it felt tight again but not as bad as it had previously. I put the box on a chair near the wood heater and found a short while later that the input shaft was again spinning freely. Temperature of the box was around 18C and all was good. Amazing what a difference a few degrees can make.

Next is fitting the output flange. The tapered part of the output shaft and the hole in the flange must be super clean and dry. I washed them in petrol (gasoline), dried them and washed another two times with methylated spirits (grain alcohol) and dried with paper towel. I dropped the flange on, screwed down the nut and set the box up as before with a post clamped to the side of the bench as a stop for the holding bar.


With a borrowed, good quality and heavy duty torque wrench set to 220NM and a length of water pipe I cranked that sucker down. That heavy, well laden hardwood bench was just starting to slide on the floor as the wrench clicked.

Bench testing: I dug out an old clutch plate, fitted a long bolt with lock-nuts and a length of garden hose as a handle …



… and started cranking and selecting gears. When doing this it is important to also turn the output shaft (it is revolving when the bike is moving), so a helpful extra hand is required. I was alone but managed ok. With pressure on the gear lever, a turn or two of the clutch plate and a twist of the output flange I was able to select all gears including neutral easily enough. Time to fit it into the bike and road test ...