Repairing sheared impeller shaft - JD 47 snowblower attachment

[jthibodeau89]

Yes, some sort of stop is required. I was not aware that the bearings were not flush with the face of the gear. Your flange is a perfect solution. Since the bearing race cannot support yhe gear from axial movement, you should use the threaded bolt to lock together. Once torqued, cut off, and dressed, you can secure with a TIG tack weld.

Your fit numbers are correct. You can undercut the corner radius 0.010 and grind the shaft OD.

 

[DanMc77]

I completed machining of a mock-up gear and shaft, just to experiment with heating the gear and cooling the shaft while I take measurements to determine the changes. The mock gear was set on top of the wood stove for an hour or so, while I took measurements. The shaft was cooled in a bath of liquid nitrogen until the boiling liquid stabilized.

It's looking good.

Mock Gear ID inches:

66 deg: 1.7095
200 deg: 1.7200 << Edit: should be 1.7120
400 deg: 1.7350 << Edit: should be 1.7135
Gear temps were measured with an optical thermometer.

Mock Shaft OD:

80 deg: 1.7130
Cold: 1.7110
The cold temp may be approaching -300 deg F. Liquid nitrogen is at -320.

With these numbers, I would have .0035" of interference. I'm wondering if I would risk cracking the gear with that much stress. Maybe .0020 would be better. Anyway, tomorrow night I'll repeat the heating and cooling at the same time, and then slip the two parts together. I'm expecting that no amount of beating with a hammer will get these parts to budge after assembly.

Mock gear and real gear


Mock gear and shaft:


Heating gear:


Cooling shaft in liquid nitrogen:

 

[jthibodeau89]

Based on your dimensions i don't think that you need to cool shaft at all. The 0.022 loose fit at 400° will drop together. The extra 0.002 clearance due to the LN2 on the shaft is not necessary. Just set the gear on a wooden block a set the shaft in place. If you have a way to keep light press together during cooling, that would be good to ensure the shaft stays seated. Looking great, congratulations.

Reducing the interference to 0.002 should be fine. Especially if you still intend to apply the 10-32 threaded dowels at the interface between the shaft and gear.

 

[DanMc77]

Based on your dimensions i don't think that you need to cool shaft at all. The 0.022 loose fit at 400ｰ will drop together. The extra 0.002 clearance due to the LN2 on the shaft is not necessary. Just set the gear on a wooden block a set the shaft in place. If you have a way to keep light press together during cooling, that would be good to ensure the shaft stays seated. Looking great, congratulations.

Reducing the interference to 0.002 should be fine. Especially if you still intend to apply the 10-32 threaded dowels at the interface between the shaft and gear.

If I keep the interference at .003, do I run the risk of cracking or fracturing the gear? I'll still add the cut off bolts.

If the shaft is fully cooled with LN2 (~ -300) and the gear is at 400, could the assembly of such disparate parts create a thermal gradient in the gear that could be destructive?

I don't have a press, so I could tap the shaft lightly as it cools and starts to bind up to keep it seated on the flange.

I'm inclined to skip the grinding step. It's a bunch of extra work to set up and perhaps adds risk since I have zero direct experience with this. And I don't have a diamond dresser for the wheel.

All comments are valuable.

 

[DanMc77]

I just realized that I made an error when I copied my notes into the email from last night. I can blame that on the rum I was sipping..... Or at least that's the excuse I'll use.

Mock Gear ID inches:

66 deg: 1.7095
200 deg: 1.7120 << Corrected from 1.7200
400 deg: 1.7135 << Corrected from 1.7350
Gear temps were measured with an optical thermometer.

Mock Shaft OD:

80 deg: 1.7130
Cold: 1.7110
The cold temp may be approaching -300 deg F. Liquid nitrogen is at -320.

I think I'll need to cool the shaft if I want this to slip together.

 

[jthibodeau89]

Ahhh, i thought the gear growth numbers looked odd, but they were linear, so i disregarded. The revised numbers are more reasonable.

Instead of LN2, just set the sample in the freezer overnight and measure. -300°F is pretty extreme.

I would try to reduce the interference fit to 0.0010~0.0015" if you are using bolts to lock together. Too many variables (alloy, hardness, case depth, etc) to say if the 0.003 interference fit will cause the gear to crack. You can try the -0.003 install on your mockup and see of you can get things flush before they lock up.

You may want to consider a 1-2 ton arbor press to aid in this repair. They are reasonable at HF with a 25% coupon. It would also help with bearing and race installs in the future.

More importantly, what brand of rum? I like the Black Seal 140 proof spiced rum.

 

[DanMc77]

Ahhh, i thought the gear growth numbers looked odd, but they were linear, so i disregarded. The revised numbers are more reasonable.

Instead of LN2, just set the sample in the freezer overnight and measure. -300ｰF is pretty extreme.

I would try to reduce the interference fit to 0.0010~0.0015" if you are using bolts to lock together. Too many variables (alloy, hardness, case depth, etc) to say if the 0.003 interference fit will cause the gear to crack. You can try the -0.003 install on your mockup and see of you can get things flush before they lock up.

You may want to consider a 1-2 ton arbor press to aid in this repair. They are reasonable at HF with a 25% coupon. It would also help with bearing and race installs in the future.

More importantly, what brand of rum? I like the Black Seal 140 proof spiced rum.

Captain Morgan Spiced Rum. I'm not real picky, it's reasonably priced and is great in egg-nog!

I'll try to take .001 off the shaft and slip the parts together. I mostly want to get an idea of whether I can bang the parts apart.

 

[jthibodeau89]

I love spiced rum 'nog. I sdd the Black Seal to my Evan Williams holiday edition 'nog.

If you want to get back apart, heat assembly to 400 and dip end of shaft into the LN2. As shaft rapidly cools, the fit will loosen. But you still may need the arbor press to push apart.

 

[DanMc77]

Today I heated up the mock gear to about 360 degrees and put the shaft into LN2. The two parts slipped together very easily and I was able to rotate the shaft in the gear for perhaps even a minute before it locked up. I could notice the parts starting to bind up. The shaft was tapped with a small hammer as it was binding up to attempt to keep it bottomed out on the "flange".

I put the gear in the vise and beat it hard with a small sledge no budge, as expected. Now on to the real stuff.

 

[jthibodeau89]

Great news, but photo did not come through.

Can you check mock up for runout?