[PSUBS-MAILIST] K-sub hatch o-ring

Wed Jun 9 20:49:57 EDT 2021

 No annotations for the groove in either set of plans.  Directly measured on K600; interpreted from drawing for K350.  The seat is marked #10 and in the materials list it shows 5/8 x 1.5 x 25 inches.  That dimension along with the drawing I attached earlier allows calculation of groove diameters.  The K600 drawing does not specify an o-ring, but the K350 materials list calls for a 471 o-ring.
You are correct on the 471 and 472 cross-section diameter at .275.  I typed it wrong in earlier message.
Sounds like either one will work.  I'm likely to go with the 472 as long as it's not too pricey.  It might be cheaper to purchase a kit and just make my own hatch o-ring.  I've only found pricing at one supplier which was $48 USD and that seems rather steep to me.

    On Wednesday, June 9, 2021, 08:19:45 PM EDT, Sean T. Stevenson via Personal_Submersibles <personal_submersibles at psubs.org> wrote:  

 That particular drawing does not appear to call out the O-ring groove diameter by annotation. Is that on another drawing? I guess you could scale it off. It also appears to show a half-dovetail groove, with a 10° taper on the ID surface, which would negate the need for stretch to hold the O-ring in as long as the narrowest width of the gland is narrower than the O-ring. Even still, you probably want a bit of installation stretch for reasons given earlier.

What reference are you using for O-ring dimensions? Per ORD5700, 400 series are 0.275" width. A 0.275" ring in a 0.215" gland depth at zero extrusion gap is 28% squeeze, which is acceptable but at the upper limit of the acceptable range for 400 series face seals. A few percent of installation stretch would reduce the cross-sectional diameter just enough to make the resultant squeeze comfortably within the range.

Extrusion should not be a problem provided the groove is sufficiently wide that when the flanges come metal-to-metal, you don't end up with an excess amount of gland fill. I might shoot for 75% - 85% maximum. Particularly with a dovetail on the inside per that drawing, which will push the elastomer against the acute angle. Also, with 0% squeeze (where the extrusion gap still exists), you don't want a lot of area subject to the external pressure compared to the hatch area acting axially. As pressure increases, you want the squeeze to win over the tendency to extrude. 0% installation stretch on a 0.275" thick O-ring would start you at 0.060" gap. With 5% installation stretch you would start at 0.053" gap. Not significantly different.

Sean

-------- Original Message --------
On Jun. 9, 2021, 17:18, Jon Wallace via Personal_Submersibles < personal_submersibles at psubs.org> wrote:

 Sean, thanks for that analysis.  See attached image for reference to K350.
I screwed up some of the K600 numbers based upon the plans which turn out not to be accurate.  Specifically, the actual K600 hatch seal configuration I see does not match either the K600 plans nor the K350 plans, so somewhere along the way the details were changed without documentation (which Lloyds didn't catch) or I don't have the final set of K600 drawings.
K600 groove - 23.5 ID, 24 OD, 23.75 mean diameter
K350 groove - 23.25 ID, 23.75 OD, 23.5 mean diameter, according to the plans.
BOTH have a groove width of 1/4 inch as shown in K350 image (including dovetail on ID side) and a depth of .215 inches.
Cross section diameter of both 2-471 and 2-472 is .271 inches.  On the K600 that's a 2.1% stretch which would be within the limits you wrote about, but now I'm wondering if Kittredge used the 2-471 on the K350 to reduce the cross-sectional diameter enough to counter the relatively hefty .271 o-ring inside a .25 x .215 channel.  Too big of a diameter would cause extrusion at some point, wouldn't it?
Jon

    On Wednesday, June 9, 2021, 05:25:03 PM EDT, Sean T. Stevenson via Personal_Submersibles <personal_submersibles at psubs.org> wrote:  

 Per ORD5700, stretch should be limited generally to no more than 5%. A 2-471 O-ring (21.955" ID) on a 23.25" gland is 5.90% stretch. A 2-472 O-ring (22.940" ID) on that same 23.25" gland is only 1.35% stretch.

Sometimes it is advantageous to specify an O-ring which must be stretched for installation in order to create friction against the gland in order to prevent the O-ring from slipping out of the groove if it isn't retained by other means (e.g. dovetail).

At 5% stretch however, the cross-sectional diameter of the O-ring is reduced by about 2.5% (and at 5.9% squeeze, almost 4%) so the gland depth may need to be reduced accordingly in order to achieve the design squeeze on the O-ring. Per ORD5700, for face seal glands in sizes 425 through 475, that design squeeze is 21% to 29% with no extrusion gap.

For face seals subject to external pressure, the gland is usually sized with its inside diameter equal to the mean diameter of the O-ring, with tolerance of +1% of that mean ID, but not more than 0.060". The reason for this is that by ensuring that the O-ring is situated on the correct side of the groove for the anticipated pressure, it won't experience premature failure as a result or being shifted across the groove on every pressurization cycle. For a 2-471, that mean diameter would be 22.230", and for a 2-472, 23.215", which would correspond to 1.25% and 1.20% installation stretch respectively, but these are recommended design minimums.

My guess is that the 23.25" gland ID just provides enough stretch to keep the O-ring from slipping out, and performance in that embodiment was good enough. Provided the cross-sectional diameter reduction and associated gland depth changes are accommodated, the only downside to greater installation stretch is premature aging of the O-ring. (See ORD5700, 3.6).

Sean

  _______________________________________________
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/20210610/46c4a7bc/attachment.html>