Shaft Failure

[kgwhipp]

See photos of the setup.

http://i.imgur.com/WHNwACl.png

Hi Folks,

We have some equipment that used a stepped shaft to fit an off the shelf gear with a smaller bore. The shafts are failing right where one would expect. I'm no expert at metallurgy but thought someone might recognize the patterns. I assume that it's a cyclical bending fatigue failure onset by the stress concentration at the shoulder compounded by the shaft's diameter size reduction.

My proposed solution is to machine the entire unit from a single bar of 4340 steel and include the face relief as shown at the bottom rendering. The extremely tight space limits any dimensional expansion outside of the current envelope.

Thoughts?


-Kevin

 

[rb1957]

looks good to me. keep the heat treat down on the 4340 ... no more than 1/4 hard (something like 125ksi), maybe normalised (90ksi).

maybe shotpeen the relief cut ?

currently there are two pieces ... the gear and the shaft, yes?
how about shrink fitting them together ?
maybe put the relief cut in the shaft ??

Quando Omni Flunkus Moritati

 

[Tmoose]

What was the OEM shaft material?
Was the junction a very small radius or even a corner?

If the gear had a larger diameter square and flat face and could be clamped hard against a similar feature on the shaft the fillet would not be (would not have been) subjected to bending. I'd still incorporate a proper radius

 

[dvd]

That whole thing raises questions in my mind. Does the gearbox get precisely located in order to set the backlash between the pinion and the trunion? It looks like that is a difficult arrangement that may impose some high strains and moments. How is the perpendicularity of the reducer hollow-bore that the step shaft nests in? Is the baseplate grouted? What are you rotating? Is the rotation bi-directional? What is the radius in the step shaft fillet? Why is it so small? Can you chamfer the id of the spur gear and use a larger fillet radius? Was this arrangement used in order to avoid purchasing a rotary indexer? Did you follow the trunion bearing manufacturer's recommendations for thickness/flatness of the surface that the trunion bearing mounts to? That big guard over the reducer looks like it limits the cooling of the reducer - what does it do? ....

 

[kgwhipp]

looks good to me. keep the heat treat down on the 4340 ... no more than 1/4 hard (something like 125ksi), maybe normalised (90ksi).

I was advised by the gear machine shop to avoid hardening the 4340 too much as well. 300/340BHN (I assume this is approx 32-27 Rockwell C?) is what they specified in a quote.

maybe shotpeen the relief cut ?

This is interesting, I never investigated this treatment. Is it costly?

currently there are two pieces ... the gear and the shaft, yes?

Two parts, yes, a gear and a stepped shaft.

how about shrink fitting them together ?

The gear comes keyed from the factory and fully hardened. We'd avoided machining it.

maybe put the relief cut in the shaft ??

What was the OEM shaft material?

Original shaft material was probably 4140 or 4340 but it would have been machined from the normalized stock and not received any heat treatment. Would a heat treatment have been appropriate for this?

Was the junction a very small radius or even a corner?

Junction was a fairly sharp corner, with a radius no more than that of the cutting tool.

If the gear had a larger diameter square and flat face and could be clamped hard against a similar feature on the shaft the fillet would not be (would not have been) subjected to bending. I'd still incorporate a proper radius

I think I understand what you're describing, it was a very tight space which didn't allow any extra vertical clearance.

Does the gearbox get precisely located in order to set the backlash between the pinion and the trunion?

Yes, it's adjusted with threaded screws and locked down with a firm clamp.

How is the perpendicularity of the reducer hollow-bore that the step shaft nests in?

Admittedly, not that great.

Is the baseplate grouted?

Yes.

What are you rotating? Is the rotation bi-directional?

4,000 lb load, 180deg back and forth in 10 seconds, every 90 seconds or so.

What is the radius in the step shaft fillet? Why is it so small? Can you chamfer the id of the spur gear and use a larger fillet radius?

It's sharp from the cutting tool. Small chamfer already on the ID of spur gear, we were avoiding any machining on the spur gear so that it would be a drop-in replacement for wear but that would certainly have helped our fatigue.

Was this arrangement used in order to avoid purchasing a rotary indexer?

Back and forth motion or budget may have negated the use of a rotary indexer.

Did you follow the trunion bearing manufacturer's recommendations for thickness/flatness of the surface that the trunion bearing mounts to?

As best as possible.

That big guard over the reducer looks like it limits the cooling of the reducer - what does it do? ....

A protective kickplate. Personnel frequent in the area.




-Kevin

 

[MikeHalloran]

Next question: Are the reducer's output bearings sized for the overhung load applied? I.e. torque reaction, plus separation force, both multiplied by the geometry of the cantilevered gear. The step in the gear shaft suggests not.

Is it possible to fit a radial bearing in the kickplate to support the distal end of the shaft?





Mike Halloran
Pembroke Pines, FL, USA

 

[Tmoose]

"Original shaft material was probably 4140 or 4340 but it would have been machined from the normalized stock and not received any heat treatment. Would a heat treatment have been appropriate for this? "

I hate buying fancy material to improve reliability when the fatigue strength could easily be boosted 10X or more with improved geometry.

Page 82 here has some data about general relationship of hardness (heat treatment) and fatigue life improvements.

http://www.timken.com/en-us/Knowled...s/Timken_Practical_Data_For_Metallurgists.pdf

Compare the possible improvement to hit by geometric stress concentrations that can increase local stress 3X, and surface conditions that can lower fatigue strength 30%.

The improvement shotpeening can make on fatigue strength can be remarkable.

http://www.metalimprovement.co.uk/documents/english_lit/MIC_Green_Book_9th_Edition.pdf

 

[Tmoose]

Tmoose (me) said Quote: "If the gear had a larger diameter square and flat face and could be clamped hard against a similar feature on the shaft the fillet would not be (would not have been) subjected to bending. I'd still incorporate a proper radius."
kgwhipp said "I think I understand what you're describing, it was a very tight space which didn't allow any extra vertical clearance."

In your first image it appears the shaft extended a little beyond the face of the gear. If the shaft was shortened about 4X the protrusion a thick stepped or T shaped washer could be made that would be no taller than the shaft protrusion.

http://www.amainhobbies.com/images/large/yei/yei-nd-yr7-as1078.jpg

it would have a counterbore for the head of a SHCS bolt. The end of the shaft could be tapped and counterbored to receive a bolt with 4 diameters grip length to permit some healthy elongation when torqued to provide 1000s of pounds of clamping force to hold the gear against the shaft shoulder, and the washer would bear in the gear face and not touch the shaft end.

http://www.beaudaniels-illustration...stock-automotive/engines/Toyota-FJ-engine.jpg

 

[kgwhipp]

Mike:

The smaller section of the shaft went into the pinion. The larger section is in the gearbox.

I'm not familiar with the separation force you mention- where does this come into play? I definitely want to include this in future calculations.

There's only 1/8" of clearance above the pinion before we encroach into the area of the spinning surface above, so definitely not fitting anything else in there.

Tmoose:

Nice find on the Timkin reference, I would have somehow imagined the opposite was true, but I agree that geometry improvements and shotpeening are the correct route to move forward with. I'll probably keep the fancy alloy for a consistent heat treat on the hobbed part, some of the heat treat folks by me don't like 1045, which would be my go-to less exotic alternative. That said... I'm probably remembering incorrectly and might have been talking to a flame hardening technician.



-Kevin

 

[MikeHalloran]

Separating force is defined in this page:

http://en.wikipedia.org/wiki/Design_and_manufacturing_of_gears

Mike Halloran
Pembroke Pines, FL, USA

 

[MikeHalloran]

If you bought the gear with a tapered bore and a keyseat, you could use a shaft with full diameter right up to the gear hub, a tapered section, and as suggested, retain it with a SHCS and a thick washer engaging a c'bore in the gear hub at the distal end of the shaft. Sort of like a B&S flywheel.





Mike Halloran
Pembroke Pines, FL, USA

 

[tbuelna]

kgwhipp-

A cantilevered pinion is almost always a bad idea. Your shaft likely failed due to a combination of bending stress concentration, torsional shear stress, full reverse cyclic fatigue bending in the shaft at 1P frequency, and surface fretting damage at the gear/shaft interface.

If you're going to redesign the pinion shaft, I would suggest adding an outboard bearing to support the pinion at both ends. The approach of using a stronger material that you are proposing is just a band-aid fix.

Good luck to you.
Terry

 

[dvd]

The components look tightly arranged and thus changes are not easily done. As MikeHalloran suggests, the overhung load capability of the reducer should be investigated. If you can install a shaft and pinion that will handle the forces but the reducer cannot, then you have another problem to solve down the road. I've been in the situation where an indexer was the right solution, but wasn't chosen due to the expense. This appears to be the case here. Costs for downtime and repairs/modifications add up. Good luck!

http://www.sankyoamerica.com/cam-indexers_dial-index_design-for-specials-DT_specs.htm

 

[Compositepro]

I'm coming in late but it appears to me that the shaft failure is mainly torsional rather that bending, but both will be related. What is your motion profile? Are you using constant acceleration (sinusoidal)? Anything else will greatly magnify the load.

 

[desertfox]

Hi

Looking at page 14 on this site and studying your pics it looks like a rotational bending failure to me.

http://www.rexnord.com/sites/process/services/renew/documents/108-010.pdf