Possible to have non-circular radial seal for o-rings?

[Faustie]

Hi everyone,

Is it possible to have a non-circular radial piston-type seal using o-rings? The Parker guide specifies minimum corner radii if I want to use a non-circular axial seal, but I don't see similar information for radial piston-type seals. If I were to take the minimum corner radii and try to implement similar, would the design seem feasible (so long as I follow compression criteria etc.) or is this likely to present problems? If so, any other solutions I should go for? Unfortunately we don't have the space for an axial seal, but could consider other seal types.

O-ring profile:

Side shot of groove (unfinished, just an example):

This lid would then fit into a flush box, with the radial o-ring providing a tight seal.

Thanks!

 

[lucky-guesser]

Putting an O-ring or similar into a non-round application is nothing new, but all of the examples I can think of a where you clamping the O-ring to the "top" of your box, not sliding it down inside of it.

1) How much room do you have between your lid and the outer box?
2) Life cycle count? Place it once and good forever or will it require regular access?
3) What is the pressure you need to seal against?

 

[Faustie]

Yeah, I looked at face seals initially but the radial seal fits our volume constraints slightly better as no flange is required, so would be preferable if it can be reliable.

It's a relatively static application. To answer your questions:
- The room between lid and box is TBD but I'll size it to ensure compression of the seal following Parker's guidelines. It'll be a fairly tight fit, roughly .1mm clearance (sizing equivalent to H8/f7 as the Parker guide recommends for standard piston applications, with groove depth sized to achieved desired compression). I'll add a chamfer to the top of the box to make assembly easier.
- Ideally place once and good forever. Maybe a few assembly cycles for testing beforehand at most.
- ~1 bar nominally, ~4 bar max. Sealing against vacuum.

My main concerns are assembly and reliability due to uneven stretching of the o-ring. Considering alternative seal types.

 

[3DDave]

The o-ring will eventually equalize in stretch so that should not be a reliability issue; use a small amount of a compatible lube to help with that.

The larger problem is that o-rings tend to be molded round so the long straight sections will not have a radial force to overcome the tendency to try to exit the groove before/during installation.

I would include a small lead-in chamfer on the mating part and make up a couple of right angle thin plastic angles, like rafter squares, to push the straight sections firmly into the grooves during installation. If not, there is a good chance of cutting the o-ring because one cannot look at all four faces at the same time.

 

[Faustie]

That makes sense, cheers - I'll trial the installation a few times and see if it is reliable.

Presumably I could also manufacture the groove using dovetails for a really secure fit, or consider other types of seals like D-ring or X-ring though I'm not too familiar with these, and unsure if they would offer improved assembly process.

 

[lucky-guesser]

Having things tight enough to hold 4 bar will also probly be tight enough to make installation "fun" as Dave pointed out. A generous chamfer, some sort of lube and patience will likely all be required to keep from getting cut on the straight walls. Only having to be installed a couple times is good, both for wear of the ring and fewer chanced of operator error.

Hard to say if it will work for sure, would just have to give it a shot.