Bob,
Hydrogen and helium permeated o-rings can be damaged when the pressure is
released in <1 second. Surface blisters, cracks, and extrusion damage are
typical signs. Complete failure of the o-ring is rare, but I have
experienced it many times. Increased mechanical compression or reduced o-ring
cross
section reduces the effects.
I would not consider a 30 second blow down even a rapid decompression.
Krell
In a message dated 8/18/2015 6:52:09 P.M. Pacific Daylight Time,
rocket@xxxxxxxxxxxxxx writes:
I was planning on using a fluorocarbon (Viton) o-ring as the seal for my
flight helium tank (a D-size medical oxygen cylinder). I've been
reading a bit about explosive decompression where the gas permeates into
the o-ring and then can blow out during rapid decompression. I couldn't
find any good examples of what "rapid" means but in my case, the
pressure in the tank will go from 3000 psi to 0 in about 30 seconds.
The Parker literature has data on gas permeability rates for different
compounds and it shows that Nitrile is slightly better (7.4 e-6) than
Fluorocarbon (12.7 e-6) at room temperature but I would probably do
better with some field examples so see if I'm in the danger zone or
not. I'm currently working my way through Research Report 485,
Elastomeric seals for rapid gas decompression applications in
high-pressure services.
A couple of recommendations for minimizing explosive decompression is to
reduce the cross section and use a higher durometer compound . Instead
of making a custom fitting (which isn't actually that complicated), I
was thinking of using a -8 AN fitting with a 2-016 o-ring into the
cylinder. The cylinder port has a pretty good countersink so the o-ring
would be completely contained but it is not a standard boss contour (a
-908 o-ring is too large). A 2-016 o-ring is small enough to fit with a
bit of stretch although I haven't tested it under pressure yet. The
other option is to go with as-intended use of a large PTFE o-ring
clamped between the fitting and flat surface on the cylinder. It's just
that when I realized the cylinder thread was the same as a -8 AN
fitting, it seemed like a nifty solution.
-Bob