Topic: DIY back-cut tranny questions.

[crazypj]

You'll probably find normal HSS milling cutters bounce off gear dogs and will have to grind them.
You would probably be better off making some sort of fixture to mount a die-grinder.
Dovetail cutter will be way too steep, you'll get one gear but never get another without stopping engine then rocking bike back and fore .
I think back cut angle is around 3~5 degrees total, just enough to pull dogs into engagement (I used to know but I forget things when I no longer need to know)

[iron_worker]

Ah I see now. I'm in agreement with the others that the dogs will likely be case hardened as well as the gear teeth. You'd probably need carbide cutters to even scratch em and even then I think it would be tough. I think grinding is probably the way to go. Why do you want to do rounded contact faces? I think this makes the issue even more complex... and makes getting matched contact faces even harder.

I think I would try attaching something like a tool post grinder near the spindle of your mill/drill. You'll need a diamond cutter so you can true the grinding wheels you use also.

I would then set up a rotary table on the on the table of the mill/drill. The rotary table would have to be shimmed or otherwise mounted such that the rotational axis is offset from the spindle of grinder by the desired back-cut angle. Will need a very high degree of precision on this angle ... some type of sine bar measuring method will likely be needed to confirm it's accuracy.

You would then need to offset the tool forward and back with the mill/drill table to the front and back face of the dog. Then rotate the rotary table to bring the next tooth around into position.

I don't think the rigidity of the mill/drill would be too much of an issue during grinding since you would take very light cuts and the loads would be relatively small.

The other thing to think about would be the cutter itself. If you used just a regular cylindrical grinding wheel then as you offset the wheel to get the desired backcut angle the side opposite the cutting side will tend to dig into the side of the gear. A grinding bit shaped similar to the dovetail cutter would relieve this problem I believe.

It's an interesting problem and I hope you have success with it. Definitely take pics of your progress and keep us in the loop.

IW

[iron_worker]

I just had a couple of other thoughts... You really got my brain engaged here. lol

If you had a diamond wheel dresser you could potentially turn the cut back angle you desire into a cylindrical grinding wheel... essentially making it tapered. Then there would be no need to shim the turntable and there would be no wheel interference on the side of the gear.

OR

Mount a grinding wheel similar to one from a tool sharpening grinder or a surface grinder perpendicularly to your mill drill spindle.   The rotary table would once again need to be shimmed to a precise in this method. Then the face of the grinder would have to be cut at an angle such that there is clearance for the side of the gear. (Angle must be at least as large as the back-cut angle). The problem with this method I suppose is you are somewhat cutting with the side of the grinding wheel.

Anyway, I think there are many ways to do it. Good luck!

IW

[iron_worker]

More thoughts....  ... once you figured out how and set up your machine to cut a lovely back cut gear dog you also have to keep in mind that all the faces must be in exactly the right angular position around the periphery of the gear. If you cut too much or too little you will have one dog that engages too early (taking all the force) or does not engage at all (does not share any of the load).

I believe this could be measured by setting up a dial gauge that is fixed to the table and the button is as perpendicular to the dog face as possible. Then bring each face around using the rotary table and you would want to achieve the same reading on the dial indicator. This would indicate that each dog face is in the same position relative to the dog centerline.

I realize most of this probably doesn't make any sense since without diagrams. lol

IW

[754]

 Makes perfect sense to me..

[crazypj]

The real problem, (why I've never even tried it) getting one dog and contact face to match isn't really too much of a problem, the real 'fun' is indexing all the others and matching them up. If you don't, you may well end up with only one dog doing all the drive (until it fails and takes out ..........................everything)

[scottly]

It ain't rocket science, guys. Using a dremel, grind a slight under-cut on the thrust side of the male dogs; it doesn't take very much, unless the dogs are extremely worn. There are twice as many holes, as dogs, so they don't tend to wear as much. Check the contact area with Prussian Blue. The gears are not a precise fit on the shafts, so any minor differences between the three dogs will equalize somewhat. 

[754]

 Using an indexer makes it it more accurate, and more chanceof spreading load equally.

 If it were only loadbearing on one tooth, the chances of grenading a gear are far higher..

[ivanhoew]

fang , looks like you have managed to epitomise the usual  thread split here ,95% posts telling you its all doomed ,4% saying it might work but dont do it ,and 1% saying hell yeah give it a try .

i will now raise that to 1.1% by saying... damn right ...have a go see what happens , life too short to not  have a go at something just because you havent done it before and so on .big up to you, oh yes !!:)

[dave500]

i think its a bloody good idea,you could probably resurect a jumping/slipping gear?you will need a degree of precision though,as crazy said they all have to share load,if one does it all itll snap off ill bet?taking out everything,,i think the factory stock dogs have never sheared,ive never heard of it,but if you unload one or two the working one wont flex itll snap aswell as put pressure on its gear to crank over.

[70CB750]

More thoughts....  ... once you figured out how and set up your machine to cut a lovely back cut gear dog you also have to keep in mind that all the faces must be in exactly the right angular position around the periphery of the gear. If you cut too much or too little you will have one dog that engages too early (taking all the force) or does not engage at all (does not share any of the load).

I believe this could be measured by setting up a dial gauge that is fixed to the table and the button is as perpendicular to the dog face as possible. Then bring each face around using the rotary table and you would want to achieve the same reading on the dial indicator. This would indicate that each dog face is in the same position relative to the dog centerline.

I realize most of this probably doesn't make any sense since without diagrams. lol

IW

Isn't it all case hardened?  Meaning that he will take the hard layer off and be stuck with soft surface?

[iron_worker]

That's a good point. I guess it depends how deep the hardening goes and how much material you have to take off. If you wanted to do it all correctly you'd probably look into having the gears hardened again when you're done.

IW

[754]

 Pretty sure they are hardened through, then tempered..

[iron_worker]

Just with the mechanics of hardening it is almost always "case" hardened to some degree. The steel is heated to a temperature that allow the iron and carbon molecules to arrange a certain way and then it is quenched in water/oil/blast of air (the rate of the cooling determines how hard the material will get ... ie how much martensite will form). Since the outside of the gear hits the cooling fluid first it cools quickest and gains the most hardness. The inner core of the gear will cool more slowly and thus the hardness reduces towards the center of the gear (where it is surrounded by the most material).

The tempering process is to ensure that the material not too brittle. It reduces hardness slightly but increases ductility significantly. I believe it also relieves some of the internal stresses cause by quickly freezing the molecules in place.

This is true for high carbon steels. You can also harden low carbon steel by heating in placing in a high carbon environment. The carbon will then migrate into the steel. I believe this is what they refer to when talking about case hardening.

IW

[scottly]

fang , looks like you have managed to epitomise the usual  thread split here ,95% posts telling you its all doomed ,4% saying it might work but dont do it ,and 1% saying hell yeah give it a try .

i will now raise that to 1.1% by saying... damn right ...have a go see what happens , life too short to not  have a go at something just because you havent done it before and so on .big up to you, oh yes !!:)

I did it because my trans had a chronic case of popping out of gear: it was either back-cut the dogs, or replace the whole trans. The dremel and stones don't remove material very quickly, so you can sneak up on it. 

[70CB750]

Just with the mechanics of hardening it is almost always "case" hardened to some degree. The steel is heated to a temperature that allow the iron and carbon molecules to arrange a certain way and then it is quenched in water/oil/blast of air (the rate of the cooling determines how hard the material will get ... ie how much martensite will form). Since the outside of the gear hits the cooling fluid first it cools quickest and gains the most hardness. The inner core of the gear will cool more slowly and thus the hardness reduces towards the center of the gear (where it is surrounded by the most material).

The tempering process is to ensure that the material not too brittle. It reduces hardness slightly but increases ductility significantly. I believe it also relieves some of the internal stresses cause by quickly freezing the molecules in place.

This is true for high carbon steels. You can also harden low carbon steel by heating in placing in a high carbon environment. The carbon will then migrate into the steel. I believe this is what they refer to when talking about case hardening.

IW

Actually the high carbon environment treatment is called cementation.  Just saying:)

[fang]

I am enjoying this conversation.  Subscribed   

As far as cutting through the 'hardening' goes, remember that more or less ALL the race bikes have their transmissions back cut.  At the top of the 2°-3° cut almost no material is removed.  At the bottom of the taper about 1mm is removed.  I believe that while some folks 'hardness treat' their transmissions after cutting, the vast majority do not.  APE and Cycle-X, for instance, do not even mention re-hardening on their tranny cutting pages:  See HERE and HERE.

Once again, I need round cuts because my gears have round synchro dogs. 

Here are few pics, just for kicks and grins.








In this picture you can see some 'typical wear' on some round transmission dogs -- the one on the left shows what no wear looks like.  The slightly rounded one up front shows typical wear on the load-bearing side (also observable on the dog above it...).  This tranny is one of my many "junk gear sets," and it came out of a 10,000 mile, un-modified, street-driven CB750.  Even under these relatively TAME conditions, you can casually visually observe the wear.


This is an example of the type of solid little Tungsten Carbide 15mm Hole Saw that I will be using to cut the male dogs.  I bought several of these from different sources.  All of them have carbide cutting teeth, are rated for cutting cast iron, etc, and all of them can be resharpened.   They have a 15mm OD, and approximately an 11mm ID.  Using a diamond cutting stone, I am going to carefully re-tool the hole saws so that the leading cutting edge is along the ID, instead of the OD.  With the Pilot Drill removed, I believe these bits will be able to slip OVER the dog (female fit) while I carefully and accurately remove the small sliver of metal required to back-cut the male transmission dogs. 



The cut will look something like this.

The female cuts will be just as simple.  I have a handful of tough, flat-tipped 10mm milling bits, and they should easily make the required cuts. 

The critical concerns here will be the accuracy of my jigs and the rigidity of my assembly.  I have a high degree of confidence that my equipment will exceed the requirements to deliver these cuts.  This ain't rocket science. 

If I jig it right, keep it cool, and take my time, this is going to be fairly simple.

Peace and grease,
-fang
 

[iron_worker]

Very cool. Keep us posted how it goes.

IW

[crazypj]

Looks OK to me, as you said though, depends how rigid the set up turns out 

[754]

 I doubt it will work with what you have shown, not trying to discourage you, but being realistic.
 If a hss tool goes dull you can usually sharpen it, a carbide in a no rigid setup will usually fracture, rather than just dull. (and be easily sharpened).
 I think that you would have better luck with a diegrinder and stone on homemade fixtures.

 Maching hardened parts is not hard with the right equipment, but it is specifically set up for grinding, and usually has good rigidity.

[754]

 Runnout is not the main factor here. I think you will have a bigger problem with spindle movement.
 If you put a dial on your spindle, put some blocks between the drill press coloumn, and spindle, now use a pry bar, or just pull on the spindle and see what you got.
 Hard to cut in a straight line if spindle is deflecting rather than actually cutting.

[solo 2]

I'm a machinist, I have everything needed to do this job correctly...I would send mine to APE . Forget HSS on hardened gear faces unlesss you own a tool grinder. Carbide works well if you go easy with lots of coolant, but if you chip the edges you're done. The hole saw idea is interesting, but keep in mind the radius of the cuts have to be the same ideally or make sure the radius of the male is slightly LESS than that of the female or it will catch the cut edges. Bluing/cutting/grinding can be done to great accuracy, but it's very labor intensive, a good rotary table would be better.

I'm not saying don't do this, many times I have spent double/triple or more just to DIY, I get great satisfaction from doing something myself but I'm also aware that it's usually much cheaper to have someone who has already perfected the process do it for me.

Good Luck, I hope you prove all the nay-sayers wrong.

[crazypj]

I trained as a precision machinist (in the 70's)
I've also spent crazy amounts of time and (rarely) money to say I did it myself
Even though I can machine to a 3 micron tolerance, I would send gears out
 As an experiment with worn gears, go for it, you have nothing to lose (except time and a little cash for tooling)

[ivanhoew]

i think i would agree with the above comments ,something that might fall a little inbetween them though , is to use a grinder to rough down to near where you want to be ,then use the cutter to finalise it all?

[lostinthe202]

aw comon' fellas, don't you remember what it was like to get something flat within a half a thou (or .015mm for you tea drinkers ) on a knee mill just to see if you could even though the tolerance was wide open?  Or how about cutting a multi-lead thread just to see if it would work?  Or squaring a block in a 4 jaw?

Or how about all those people who told you not to bother working on your own bike (or car or truck or whatever) because there were people out there that specialized in that sort of thing and it would just be a waste of time and money to do it yourself.  Did most of the people on this board listen to them?

It's true that Fang might not succeed, but that's hardly the point.  If all he cared about was back-cut gears, he would have sent them to a shop.  Obviously he wants the satisfaction of doing it himself.  So far it seems to me he's still setting up his process, so lets give him encouragement and advice when it's asked for, that IS the point of his post right?