Press fit between Hub (Celcon M140) and Shaft (T6 6061) Question 2

[Core2]

All,

First time poster and long time reader of Eng-Tips.

If I have a hub made of Celcon M140, with an OD 14.00 mm, and an E of 2500 Mpa
A shaft made of Aluminum T6 6061, OD 9.2mm, ID 7.36mm, Length 10.00 mm
I need to figure out the interference needed to accomplish the following:

A press fit that will create a hermetic seal, during temps of -20F to 130F, and last for 10 years. I only mention the time constraint because I am trying to consider creep rupture of the Celcon in this calculation. This is only a static configuration, with no external torques. However, it will encounter vibration.

I keep calculating a ridiculous interference fit of 12.57mm. Absurd!

Links to information that I have been using:
Designing with Celcon:

http://www.kmsbearings.com/pdf/Celcon_Design Guide_3.9.07.pdf

Analysis of Press Fits:

http://www.faculty.fairfield.edu/wdornfeld/ME311/PressCylinderHam.pdf

Fundamentals of Machine Elements: Chapter 10 by Hamrock, 1999 (Found some parts on the itnernet, I don't have the book)

Thanks in advance for any help you may provide.

 

[Pud]

You have not considered the fact that plastics exhibit creep (not to "rupture" specifically). This is the the "killer" in your requirement of ten years...

Try and find the isochronous creep curves for acetal (Celcon in your case). Unfilled polymers are pretty poor - although acetal is one of the better ones.

Your calculations may be correct - in practice they are a no-go!

Filled materials will be better, but 10 years stressed is an awful long time for polymers.

Have a look at elastomers, which may be better suited to retain a seal...

www.tynevalleyplastics.co.uk

It's ok to soar like an eagle, but weasels don't get sucked into jet engines.

 

[Core2]

Thanks Pud for the fast reply

I will research what you pointed out about the isochronous creep curves.

I should have also mentioned that the hub will be submerged in gasoline.

Would you suggest a specific elastomer?

Thanks!

 

[Cockroach]

I had a little uncertainty with Celcon M140 mechanical properties, went with a Poisson's Ratio of 0.37 as an average. I also used a yield strength of 8850 psi which seemed to be consistent with industry specifications. Converting your Aluminum T6 6061 shaft to imperial units, 0.36220 inch OD, 0.28976 inch ID and the Celcon M140 hub as 0.55118 inch OD, I transferred the shaft to hub interference length as 0.39370 inches. I assumed the worst case, no lubrication along the mating surface so the coefficient of friction (static) is 0.15, probably a little on the high side (worse case).

Using Lame's Equation as reduced from Pressure Vessel Theory, I have a Hub ID=0.35720 inches, which is 9.076 mm. This would give you an axial force press fit of 113 lbf with a sustainable torque capacity of Hub to Shaft of 21 inch lbf. You can play around with Mohr's Circle and find the shear due to torsion is 757 psi, principle stress of 653 psi for an element rotation into the principle plane of -12.303 degrees. As a result, your factor of safety is 2.063 which I think is sufficient for this application.

So the interference fit between Shaft and Hub is 0.005 inches on diameter or 0.127 mm.

Hope this helps you out.

Regards,
Cockroach

 

[Core2]

Cockroach,

Thanks for the time you spent on working this out. I am currently reviewing my work to figure out where I went wrong.

Your answer is more appropriate than what I was calculating.

Thanks again!,
Core2

 

[Core2]

Cockroach,

I have just found the time to return to this problem and I am still not calculating a reasonable answer. Although it is better then what I was calculating since, Thanks to you, I realized that I should be using the yield strength.

Could you possibly point me in the right direction? The following pic explains what formulas and values I am using and my answer. Not shown are the dimensions of the shaft and hub. I am using the same dims as shown in my first post.

Picture

I tried to embed a picture without success. The link works, but it's not convenient.

Thanks,
Core2

 

[Core2]

Still tinkering.

Does anyone see something that I am missing?

 

[Cockroach]

It's difficult to see what equations you are modeling from when given a spreadsheet. I have attached the model regarding Thick Wall Pressure Vessel and the reduction to Lame's Equation. Let me submit this first, unfortunately I can only get one attachment at a time.

What I would do is be sure you are using the correct quantities, I've seen interns using diameter and radius interchangeably, which mucks up the outputs. Also check the mechanical properties of your materials, be sure there are no errors in the quantities put into the program. Quite often it is easy to overlook that, messes a guy up pretty badly.

Regards,
Cockroach

 

[Cockroach]

Okay, attached is my file regarding the computation using the information you have provided. I have great faith in the output, having measured several interference fits with a hydraulic press following machining of Shaft and Hub to these outputs. Typical errors are those of measurement, around 5%. Clearly the big issue with this is friction factor, so I use a manufactured spec lubricant like EquiLube 80. Alternatively a guy could go to the web and check out various results that were lab tested. This is perhaps more practical in order to get order of magnitude numbers.

I generally work in imperial, convert metric to english units for this kind of computations. Old school, most of my equipment is imperial and I have a greater feel for the outputs as such. Otherwise, you can convert back and forth as I do in thermodynamics and chemistry projects. But I would use your model with my inputs and then see what numbers get kicked out. Could be a software related issue. Check your equations.

Bet you it is something simple like that, the hardest thing to find is the old perverbial needle in the haystack.

Regards,
Cockroach

 

[Cockroach]

Finally, attached is a Press Fit Tolerance Chart which is typically used to find your class of interference. I have you sitting around K6, which is kind of high for your situation. Noting that the plastic will give a lot more than a carbon alloy hub, I would think most of the deformation will be to that component, you may wish to lower it slightly depending on what kind of torque is needed. Remember that tolerances are usually measured in the fourth decimal place, we're talking very find detail work for the machinist. So shaft and hub diameters are somewhere like +0.0005/-0.0000, +0.0000/-0.0005 inches respectively for carbon steel alloys. You would never get that accuracy with a plastic hub, so I think this contributes to the difficulty in your problem, hence my move towards a high tolerance press fit.

Hope this helps you out.

Regards,
Cockroach

 

[Core2]

Cockroach,

Thanks again for the posts. I plan to dig through my work and find the source of my error the next time I have a spare minute.

This has been fun. It has been a while since I have worked on a problem like this. Reminds me of school.

At my place of work we had some press fits that use the same material and the interference you calculated was similar. I wanted to calculate it myself to confirm the already existing press fit tolerances were good. I don't like relying on dimensions that I don't know the source of.

I will post my findings as soon as I find what went wrong.

Thanks,
Core2

 

[Core2]

Cockroach,

The parts I received parts. Their ID is .15mm under sized. When I press them together I measuree 125 lbf. So your 113 lbf was a good calculation.

Core2

 

[Cockroach]

Yeah, that would also vary with they type of grease used between mating surfaces. But happy she worked out for you!

Regards,
Cockroach