Deflection/Stress from Compressed O-Ring

[KevinH673]

This may seem like a trivial problem, but I'm a new Engineer and the stress I am calculating from an O-Ring compressed roughly 30% seems a bit high, and am looking to see if I did this correct.

I have included the .pdf of my hand calculations

My stresses calculated are 16 ksi and 28 ksi, which is a significant difference...

Deflection is only 6/1000's this seems fair.

 

[KevinH673]

Forgot to mention:

The O-Ring goes into the 0.070" gap. The O-Ring is 0.070" wide, and is compressed to 0.050"

 

[Dtuck]

O-ring compression sealing is generally calculated by percentages and not stresses to start with. 18-20% for vertical compression and 93% for volume fill when compressed is a good starting point.

 

[KevinH673]

Thanks for the reply. I am actually not looking for the sealing of the o-ring, as it's not being used to seal a fluid, it is being used against a mirror to prevent direct contact of the mirror cover (the piece shown) and the mirror. The o-ring keeps the mirror in place. I am interested in the deflection and pressure that the cover is seeing.

 

[btrueblood]

I really can't follow your calculation past the equations you picked out from Mark's hdbk.

But it shouldn't matter - you know you are getting 14 lbs. of force from the 30% deflection of the O-ring, and want to know what the probable normal stress is from the O-ring to the surface it is pressing on, correct?

Face-seal loaded oring. Oring o.d. = 0.441, i.d. = .301. Projected area = pi/4 *(.441^2 - .301^2) = .0816 in^2
Force = 14 lbs. (assuming you limit the deflection to 30%). Probable mean normal stress across contact area is

14 lbs./.0816 in^2 = 172 psi. If you want to argue that there is a Hertzian contact stress distribution, and the peak stress could be 2x of that value, I could say maybe.

But your numbers, if I interpreted your problem correctly, is high by about 2 orders of magnitude.

 

[kevindurette]

$0.02: Don't forget that elastomers are essentially incompressible. There will be significant "squishing" sideways, which will alter your result.