shink fit problem

[fe16260]

Knowns:
D2 tool steel Sleeve (Rc 62)
4140 Shaft (Rc 28)
This is a fairly large part and the interference fit is calculated/specified to be 0.008"

Problem Statement:
Have had (2) issues: (1) the sleeve cracked (like an eggshell). I assume it's due to high stresses since the D2 sleeve is heat treated. I say this b/c there has been times when it has worked without cracking. It was stress relieved but still cracked. The (2) issue is the sleeve slips on the shaft when backing off of the interference fit (0.008" down to 0.005"). It is obvious that the fit needs to be 0.009" or greater but I need to address the cracking issue so that i can get better reliability out of the part.

Question:
1. Could i use a different material for the shaft (like stainless) to where once the sleeve shrink fits it will gal (seize). My thought here is if it gals/seize (the sleeve to the shaft) then it would help keep the part from slipping without using a high interference fit. I would be ok if the part didn't seize until it slipped a little.
2. HOw do you go about selecting the materials (sleeve) and (shaft) to get optimum surface area/grip??? In other words, will the interference fit (resistance to torque) be best if the two parts have the same material? different materials?

I appreciate in advance for any comments and/or direction.

Thanks
Andy

 

[desertfox]

Hi fe16260

What are the dimensions of the sleeve and shaft? what are the stresses due to the interference fit and what torque are you trying to transmit.

desertfox

 

[fe16260]

Desertfox,

Thanks for replying - here's the information you requested:

Transmitted Torque = 900,357 in*lbs
Internal Diameter interface(shaft-sleeve) - 16.9 inches
External Diameter of sleeve - 26 inches
Radial pressure at interface to transmit torque (assuming a 0.003" radial interference fit) - 3087.2psi
coeff. of friction assumed to be 0.35
Tang. force at interface due to transmitted torque-106,551 lbs
Thermal expansion - 0.006 inches
Operating Temp - 130degF
Assy Temp - 75degF
Therm expan for steel (alpha) - 6.6E-6 in/in degF
Calc. stresses due to internal pressure of 3087.2psi - 411.2psi

I hope this helps. Like i said, i would like to use a stainless shaft and steel (D2) sleeve and maybe let the two gall together - will this happen with such a press fit? Ultimate, i am trying to solve the issue of the sleeve slipping on the shaft; hence, is why i want to see if i can use galling in my favor to help with this.

I am looking at cryogenically treating the sleeve after final machining/grinding process so that i can relieve all stresses from the part - hope this will allow me to get a greater shrink fit without cracking the sleeve.

 

[desertfox]

Hi fe16260

I'm looking into it

 

[desertfox]

Hi fe16260

Right I've done some calculations on this which I need to double check yet, however I need the length over which the interference is maintained.
My calculations show at this point that using a mu of 0.35
the maximum torque that can be transmitted is just over
640,000 lbf.in based on a one inch long engagement.
I don't think relying on galling will help its like the friction coefficient highly variable, that said I think your best bet would be to do some practical measuring tests and establish a more realistic coefficient of friction you can use in the calculations.
The stress at the interface of the sleeve and shaft is about 10000 lbs/in^2 and that at the outer diameter around
3100 lbs/in^2.

desertfox

 

[desertfox]

Hi fe16260

Just to add about friction coefficient, here are some values for steel on steel:-

http://www.roymech.co.uk/Useful_Tables/Tribology/co_of_frict.htm

scroll down till you see the title press fit joints.

Here are some formula which I checked my calculations against:-

http://www.roymech.co.uk/Useful_Tables/Mechanics/Cylinders.html#Interference

How have you established you're friction coefficient because I believe that is the problem ie its too high, also the 10000psi stress doesn't seem to high to me so I'm not sure why the sleeve is cracking unless there's a problem in the heat treatment method.

 

[fe16260]

THANK YOU so much sir for checking this for me. I will check the length of the engagement on the print when i get back to work tomorrow.

As to establishing my friction coefficient (it was just assumed for the calculations), i have specified to the fabricator to have a very rough finish on both the shaft and sleeve. There's no way of me to measure this. I know that the slicker the finish the lower the coefficient so i am trying to rough it up to gain a better coefficient of friction.

I believe the sleeve cracked due to heat treatment. I plan on using -325degF cryogenic process to remove all stresses from the part after the final machine process so that any residual stresses will be moved.

Here's the current process:
1. Rough Machine blank
2. heat treat
3. double temper & stress relief
4. final machine
5. hone ID of sleeve to size
6. Assemble shaft to sleeve using liquid nitrogen on shaft and heating up the sleeve

Here's the new modified process i am going to use:
1. Rough Machine blank
2. heat treat
3. double temper & stress relief
4. final machine
5. hone ID of sleeve to size
6. Cryogenic treat sleeve to -325degF
7. Assemble shaft to sleeve using liquid nitrogen on shaft and heating up the sleeve

Thanks again and i will get you that information so you can confirm it.

So did you use 0.009" as the press fit value or did you happen to calculate what the press fit should be to transmit the torque?

Thanks
Andy

 

[desertfox]

Hi fe16260

I used 0.008" for the interference fit.
If you can wait till weekend I'll post my calculations.

desertfox

 

[fe16260]

Hi Desertfox,

The length the interference is maintained (length of sleeve) is 36.846 inches.

Yes that would be great to post your calculations when/if you get the chance. Thanks again - sorry for the delay on my part but i have been swamped at work .

Best regards,
Andy

 

[desertfox]

hi fe16260

According to my calculations you're 0.008 or 0.005 thou interference should be more than adequate for the torque requirement over the length you've specified for a mu of 0.35, in fact it should be still ok at a mu of 0.15.
I would question though the length of intereference and its consistancy over such a long length.

 

[fe16260]

We are honing the bore out and they are holding the ID true to 0.001" of total runout. Maybe that's the issue were having is maintaining that interference over the length of the sleeve.

Again, thank you very much Desertfox - look forward to seeing your posted calculations when you have time to post them. Thank you!

 

[desertfox]

Hi fe16260

Here are my calc's.
By what means are you measuring the bore diameter over a 36.8" length and verifying the tolerance and runout?