[PSUBS-MAILIST] O-rings and cold temperature

Wed Jan 28 20:15:48 EST 2026

I could be missing some information. Obviously NASA and MT engineers would be the authoritative word on the issue, but to me, it actually makes sense that a temperature swing in the positive direction would be more likely to induce compression set. Here's why: Viton is generally pretty resistant to compression set, but in any seal design, you have a target initial installation "squeeze", or compression of the O-ring, which should be enough to maintain an effective seal in the absence of differential pressure. If you want to maximize that initial seal force, you increase the initial squeeze to some value below the value that would cause compression set in your elastomer, so that it will recover fully if you remove the O-ring from the groove. Now, when you increase the system temperature, the groove bore and plug diameters of the metal parts increase, so increasing the height of the groove, but the elastomer cross-sectional diameter also increases, and the coefficient of thermal expansion of the elastomer is something like six times that of the steel, so you end up with a larger O-ring being crushed beyond design squeeze in the groove, and if the resultant squeeze is greater than the value at which compression set begins, then when the temperature cools and everything goes back to normal, the O-ring having suffered compression set will exert less sealing force in the groove.

Now, how it could work the other way, is that in addition to cross-sectional "squeeze", an O-ring also nominally experiences a circumferential "stretch", which is nominally slight in most seal designs, but of course as temperatures drop, this stretch is lost in favour of increased cross-sectional diameter of the O-ring, which increases squeeze sympathetically with simultaneously shrinking groove dimensions. If the dimensions of the groove are tight (e.g. 90+ % groove fill), this could induce unrecoverable compression set which would result in lower nominal squeeze once the temperature returned to normal.

Many variables here. You'd really need to look at the specific groove design and elastomer to analyze the behaviour.

https://proton.me/mail/home

As far as PSub applications are concerned, submerged temperatures have a pretty consistent and narrow range that is relatively easy to design seals for. If I were trailering a sub through severe cold in advance of a dive, however, I would be inclined to give it time to equilibrate with the water temperature for a while before diving it. This is also prudent for your acrylic windows if they have been allowed to become thoroughly cold.

Sean

-------- Original Message --------
On Wednesday, 01/28/26 at 16:42 Alec Smyth via Personal_Submersibles <personal_submersibles at psubs.org> wrote:

> Thanks Sean! That's a bit different reasoning than the one given in the article (that is wasn't the cold on launch day that did it, so much as a permanent set taken on days prior when it was colder.)
>
> I do think that, fortunately for us, there's something fundamental that makes O-ring seals problems more forgiving on subs than on spacecraft. Its the fact that if an O-ring is going to seal poorly it tends to happen at low pressure, which for subs means on the surface. But spacecraft start with "high" pressure that tapers off as they climb. I'm guessing that's why Challenger blew up a little over a minute into her flight.
>
> On Wed, Jan 28, 2026 at 4:20 PM Sean T. Stevenson via Personal_Submersibles <personal_submersibles at psubs.org> wrote:
>
>> The problem with the shuttle O-rings was their dynamic performance charactetistics, as opposed to the quasi-static ones.
>>
>> Interestingly, there was no actual functional reason for the SRBs to have the O-ring sealed joints at all. The contract for manufacture of the SRBs was awarded to Morton Thiokol out of state for somewhat suspect reasons, but regardless, this necessitated sectioning the booster in order to accommodate inter-state transport by rail, instead of manufacuring them as a single part in a facility adjacent to the launch complex.
>>
>> So, the SRB sections were designed with those joints, comprising a double radial O-ring seal protected from the propellant exhaust gases by a non-incendiary putty and insulation on the inside. The cold temperature on launch day, in addition to shrinking the O-rings and reducing the static sealing contact pressure as installed, also reduced their resiliency by increasing their stiffness and how readily they could both deform under applied pressure and return to their original shape when that pressure was relaxed.
>>
>> The SRB is not simply a tube filled with solid propellant that burns away from the bottom end. In order to maximize specific impulse, you want to burn as much propellant as possible in as little time as possible, so the SRB actually has a hollow core which is ignited from the top down to set the entire core alight, and furthermore, instead of being a simple cylinder, the shape of that core void is convoluted in order to maximize its surface area for the initial burn. The result is very high exhaust gas pressures achieved very rapidly. Ordinarily, the O-rings would rapidly expand and deform into their service (sealing) shape in the presence of the applied pressure, but in the case of Challenger, this pressure was applied too quickly for the cold and stiffened O-rings to keep up with the deformation rate required to maintain the design sealing force through their transition from initial installation shape to their maximum service pressure sealing shape , resulting in blow-by of the hot exhaust gases which ultimately failed the joint structurally.
>>
>> Sean
>>
>> https://proton.me/mail/home
>>
>> -------- Original Message --------
>> On Wednesday, 01/28/26 at 12:45 Alec Smyth via Personal_Submersibles <personal_submersibles at psubs.org> wrote:
>>
>>> Here's one for the PSUBS brain trust.
>>>
>>> I was reading today about the Challenger accident 40 years ago. The O-ring that failed was Viton, and the failure was caused by cold weather in the days prior to the mission. The low was 18F. Per the engineers interviewed for the article, when an O-ring contracts due to extreme cold it sets at its newly reduced size, without regaining its prior shape and elasticity as temperatures go back up.
>>>
>>> Here's my question. The variant of Viton used on Challenger has a service temperature range with a lower end of -20F, way colder than +18F. Why would 18F have damaged it?
>>>
>>> The O-rings on Shackleton are Buna-N, which is rated to -40F, but I'd still like to understand the failure mechanism, since the weather forecast calls for 3F (-16C) in a few days' time, and I'll be diving Shackleton in Florida two weeks from now.
>>>
>>> Best,
>>> Alec
>>
>> _______________________________________________
>> Personal_Submersibles mailing list
>> Personal_Submersibles at psubs.org
>> http://www.psubs.org/mailman/listinfo.cgi/personal_submersibles
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.whoweb.com/pipermail/personal_submersibles/attachments/20260129/e5aa6d59/attachment.html>