Shrink Fit Failure 1

[CodeRed7]

I am currently designing an assembly that requires a sleeve to be shrink fitted onto a potentiometer shaft. I am having issues with analyzing the stresses caused by the fit. I have tried using an online calculator, however no matter what I do I always seem to have stress levels that are too high. I have summarized the information below.

Material

Hub Material: 360 Brass Alloy
Shaft Material: 303 Stainless Steel​

Dimensions

Shaft Dimensions:

Maximum Shaft Diameter: 0.1248"
Minimum Shaft Diameter: 0.1245"
Length of Shaft: 0.313"​

Hub Dimensions:

Maximum Inside Diameter: 0.1244"
Minimum Inside Diameter: 0.1242"
Outside Diameter: 0.25"
Length of Hub: 0.4375"​

Fit Specifications:

Type of Fit: Light Drive Fit [FN1]
Maximum Interference: 0.0006"
Minimum Interference: 0.0001"
Engagement Length: 0.1309"​

If I need to specify any more information, let me know. Any help would be appreciated, thank you.

 

[desertfox]

Hi
Try the link below

faq404-1230

 

[CodeRed7]

Thanks for the help. One quick question though, what exactly is the δ symbol? Is that the symbol for the amount of interference?

 

[desertfox]

If your seeing that symbol on the tribology site I don't know because my comp won't open the calculator sorry

 

[dlybbert]

I am always really hesitant to use online calculators. I remember one that converted 1 ton to 64400lbs (2000*32.2). Remember that they may or may not be anywhere near close. That being said, this one looks pretty good.

The symbol is a small "delta", and as best as I can tell, denotes the interference in your fit.

A shrink fit is a function of thermal expansion and elasticity (so long as you stay in the elastic range), and generally need not be overly complicated. The indeterminate solving assumption is (generally) that the stress at the interference faces is the same for the outer and inner components (i.e. the inner pushes out as hard as the outer pushes in, given identical dimensions). A usual mistake is the assumption of identical deformation between the parts. This is not a correct assumption.

Machinery Handbook and other design books (Norton, Shigley, Roark) layout typical methods for designing shrink fits.

"If it has been done... there is a better way."

 

[3DDave]

δ is noted on the website: "The value of the diametrical interference is typically about δ/d=0.001. The calculator is based on elastic deformation (Lame's equation), i.e. the stresses should be smaller than the elastic limit Rp0.2 of the elements."

I think I would just stop that sentence at 'diametrical interference.' as the value they show for δ is (d*.001)

Based on the online calculator, it looks like you need
nominal dimensions,
amount of interference,
Poisson's ratio shaft and hub,
elastic modulus shaft and hub,
elastic limit shaft and hub and
friction coefficient.

From this you get
how much torque the joint can carry,
whether the parts fail due to plasticity, and
how much axial load will be required to install/remove, and
the side knowledge of what the contact pressure is between the bushing and the shaft.

 

[kingnero]

one thousandth of the diameter is a good place to start, it will be the correct order of magnitude (at least that is my experience when working with two steel components).

 

[CodeRed7]

All right, after playing around with the calculator fit, I have found that polycarbonate (40% glass filled) would be a suitable material for hub, with a FN4 fit. Would using that polycarbonate be a good idea? I'm not too familiar with the material.

 

[desertfox]

Hi CodeRed7

What's the application?
The materials you started with are metallic, polycarbonate is a plastic used vandal proof windows and doors (in place of glass).

look at this link:-

http://www.teijinkasei.com/drilldown.asp?page=applications.asp

 

[CodeRed7]

Hello DesertFox,

Yes I started with both metallic options, however fiddling with the online calculator led me to plastic material for the hub, since metallic material caused stresses that were large. I am now onto plastic materials, and am trying to find the right material. I am designing a gearing assembly, that would require the hub to be shrink fitted onto a potentiometer shaft. The hub will have a knurled exterior, and be press fitted onto a gear. I would attach the potentiometer shaft to the gear directly, however I cannot modify the shaft (in order to knurl or have a key or spline), and cannot find a gear with a bore size that would have the right dimensions in order to permit a shrink fit. I need the fit to be able to withstand a load of 40 oz-in.

 

[desertfox]

Hi CodeRed7

Well you can have polycarbonate gears, do a search and you'll find loads of them.
You could try and buy a blank gear in Nylon or polycarbonate and machine in the bore for the shaft.
What is the environment for these gears?

 

[3DDave]

What characteristic of the original selection seemed to be insurmountable?

One option not mentioned is using a shaft locking compound to tie the metal gear to the shaft without a press fit.
Another option is to have the gear supplier alter the bore to match - in many cases bores are purposely small to allow this option so the supplier can have in-stock gears requiring only this operation.

 

[CodeRed7]

Desertfox
The gearing assembly will be attached to a balancing valve that will be fitted on pipes in the mechanical room of a building. So the temperature would most likely be around 25-30 degrees Celsius.

3DDave
If I were to use brass as the material for the hub, the shrink fit would result in a stress that is higher than its yield strength. Playing around with the numbers in the calculator, I found that I needed a material with a lower modulus of elasticity, which made sense because lower modulus' cause lower stress when strained. I decided to go with plastic as it had a lower modulus.

The shaft locking option and bore altering option are both possibilities. My only concern with both would be the price for both. Would you happen to know generally which would be more expensive? I will do the research anyways, but would like to know a general opinion.

Thanks for all the help

 

[desertfox]

Hi CodeRed7

Normally you would use a key to fit hub and shaft together but that's not an option, so I can see only machining the gear bore to suit shaft. Machining a matching bore is done all the time and that's the way I would go.
One thing to mention what coefficient of friction did you use, I imagine that the friction coefficient is lower for plastic on metal; than metal to metal, if so it means the interference must increase.

 

[CodeRed7]

Hi DesertFox,

I used 0.35 as the coefficient of friction between steel and brass, and used 0.39 for the polycarbonate. That is the opposite of what you mentioned, but my sources could be inaccurate. I will double check, but would you happen to know of a reliable website?

 

[desertfox]

hi CodeRed7

I found similar figures to yours, I was surprised though.
just looking at polycarbonate and comparing it with brass it appears that brass has a better yield stress than the polycarbonate (

http://www.makeitfrom.com/material-data/?for=Polycarbonate-PC)

according to the figures I'm seeing.

http://www.efunda.com/materials/common_matl/common_matl.cfm?MatlPhase=Solid&MatlProp=Mechanical.

According to my rough calculation with an interference of 0.0006" and using steel and polycarbonate you will transmit a maximum torque of 9.41oz-in and that's nowhere near 42oz-in.
I'll check my figures again, is there any chance you can post your calculations?

 

[3DDave]

"I need the fit to be able to withstand a load of 40 oz-in."

What kind of potentiometer requires that level of drive? Is it an inertial transient load?

Give a Google a search for "Shaft Locking Compound" or "Shaft Retaining Compound"

 

[desertfox]

hi CodeRed7

I should of added in my previous post that you should design the system from the external load components first and then work backwards towards the driver, in this case a motor, but it needs to be sized based on the external load requirements and then you don't end up with a motor that's under powered.

 

[CodeRed7]

3DDave:

I misunderstood how the torque will applied to the load, and as a matter of fact only need the fit to withstand a load of just greater than 0.7 oz-in, so most plastics would work in this case. But yes, I will google that right away.


Desert Fox:

And yes, brass has a higher yield strength, but also has a larger modulus of elasticity, which results in larger stresses when the material is strained. I used the calculator mentioned above, and I will attach both the results for the brass and for the plastic I used, which I have changed now to nylon 6,6 due to the lower torque requirement. It's not important now, but I used 40% Glass filled polycarbonate, which has different mechanical properties. I'm sure your calculations were correct, as I got similar results when I used normal polycarbonate. Also, the motor has already been selected, and can already generate enough torque to drive the loads.

The calculations are all in metric, and converted all my measurements to metric as such. For reference, 0.7 oz-in is equal to 0.005 Nm, and the yield strength of Nylon and Brass are 63 MPa and 125 MPa respectively. As shown by the calculations, the stresses for brass are larger than 125, which is why it was ruled out. For the nylon calculations, I used interferences of 0.0001" and 0.00006" (since those are my interference limits) to make sure the allowed stresses and torques were within their limits for both.

I will double check all calculations by hand when the material is finalized, but am using the calculator for a general guideline.

Attachment Order
1) Brass
2) Nylon - 0.0006" Interference
3) Nylon - 0.0001" Interference

 

[CodeRed7]

Also I was wondering, would nylon be able to be machined to those tolerances? If not, are there any other plastics that I can use, or would I have to go with a different fit?