Welding an automotive gearbox mainshaft

[Huwpuw]

My question relates to removing (through grinding) 1st gear on a mainshaft, removing a 2nd gear with a lower ration from another gearbox, and welding it onto the original shaft.
I had originally hoped I could do this by tig welding, cooling, repeat. This would have avoided annealing shaft and gear.
However, trying this out has proved this to be easier said than done, and to get a decent fillet means the shaft and gear are softened.
This leaves me with a question, can the shaft, once welded to this new gear be heat treated back to something close to original hardness?

The application is a low mileage dirt track car, normally run in one gear, but it needs a "gate" gear to take it to the first bend, hence the fitting of a lower second gear in place of the original 1st.
Thanks,
Huw

 

[metengr]

What you propose by welding a set of gears and subsequent post weld heat treatment of the main shaft will adversely effect the dimensions of the finished gears, especially the gear teeth, and original main shaft dimensions.

I would consider fabricating a new main shaft from bar stock with new gear blanks, which are integral. This way you can heat treat the bar and gear blanks to what you need in terms of strength, and final machine the gear teeth and main shaft after heat treatment. If necessary, a surface treatment like nitriding can be done to the gear teeth as a final step.

 

[Huwpuw]

i looked into a new mainshaft, but the gears would have to be helically cut to match the secondary shaft, which makes it difficult, and more specialised.
I have considered building up the shaft with weld and getting a spline cut into it, with a matching spline in the gear. Maybe the best option.
I have also being doing some reading up on heat treatment of unknown steels, it really is a complex science. I do have one luxury though, i have a few spare, unrequired mainshafts, and could carry out some annealing, HT and tempering tests.
Is it worth trying do you think?

 

[Tmoose]

Do you have the pair of gears you'd like to use?
Could you post pictures of those gears, and the existing mainshaft and cluster gear?

http://www.morris4x4center.com/jeep-transmission-parts/images/t84-transmission-parts-5.jpg

 

[Huwpuw]

Yes, the secondary shaft gear is from a lower ratio box, its splined and just slips off, the synchro teeth line up, the baulk ring fits. The mainshaft gear (primary shaft if you want) is the one i'll be cutting off the mainshaft of that same gearbox. So i'm pinching a pair of lower ratio gears from another box, giving me a low and high second gear. I sacrifice 1st gear (pit gear effectively).
I'll see if i can take some pics later.
Thanks,
Huw

 

[Jughandle]

Whenever weld-heat is applied, all previous heat treatment and microstructure of the base metal is erased. So the best approach, rather than forging and fabricating new parts, is to connect the existing pieces without welding if possible. Could the desired gear be bored out to take some sort of adapter? Otherwise I think the next best chance for success would be a weld overlay on the shaft followed by machining splines as you already suggested. Check the shaft for runout before and after to see if gear mesh tolerances are effected. Assuming the shaft is weldable (low-med C), there still may be HAZ to deal with. Pre and post heat of a couple hundred degrees (to suppress formation of untempered martensite) may not hurt - just my seat of the pants opinion. Oh, the gearbox manufacturer would probably not recommend any of this rework.

 

[Tmoose]

I'm interested in the size of the gear to be welded relative to the mainshaft nominal OD. Maybe if a profile (spline, polygon, duodecahedron) could be machined when removing the integral gear, and the mate created in as the "ID" of the donor gear, the weld would be simpler or even could be eliminated.

 

[Huwpuw]

No i'm sure the original maker of the gearbox wouldnt recommend using 2nd gear where 1st used to be.
I'm warming to the idea of splining the shaft, it was my first choice way back, however the diameter of the gear that has to be removed is near a damnit the same as the bearing surface next to it (being 1st it's next to the input bearing).
Now, i have tried laying a bead of weld along the shaft between the bearing surface and the clutch spline - it seemed to work OK. So if i could build the area of the removed gear suuiciently, it should be possible to have a spline cut into it. Do the same on the inside of the gear liberated from another shaft and it might work...

 

[tbuelna]

After you build up the shaft surface with welding, machine the weld surface material to just above what your spline OD will require. Then perform a mag particle inspection of the area around the weld to check for defects or cracks. The cost to setup and spline a single part will be fairly high. You don't want to pay for cutting spline teeth on a shaft with structural defects.

 

[Huwpuw]

As soon as i work out how to post pics i'll do so.

 

[tbuelna]

One other thing to consider. If you decide to use a spline to attach the gear to the shaft, make sure to use a type of spline that provides a precise radial fit. Something like a major diameter fit spline is required. A side fit spline would produce far too much runout at the gear mesh.

 

[Huwpuw]

Nothing is as simple as it would first appear. Even a spline has issues, the one you mention presumably assures a good radial fit. however, i was intending to tig the gear to the shaft (just a small tack) to stop it moving along. If the spline isnt a tight rotational fit it's likely the weld will crack.
I'll see if its possible to set the gear position with a spacer fron the reverse gear, otherwise this is bad news.