Can you not bond it in place?
 

grease ?

are you counter-boring (ie a recess 0.442" dia) or creating a groove for the O-ring ?  how deep ? (0.05" ?, ie will the diameter of the O-ring be inside the groove/recess ?)

an oring groove at the bottom of the cbore is probably a better solution. The oring should stay in place even when the parts are disassembled and should lead to a better seal.
Just a cbore alone might lead to pinching/damaging the oring.

But if you want just make it slightly smaller than the OD.. taking any tolerances into account.   

It could be bonded, but we would like to not need any additional parts/steps for assembly.

The title of this thread is deceiving, it is a c'bore, not a groove. diameter is TBD.
We would like 20% compression, so the depth for this .070" thick o-ring would be .070" * .8 = .056"

I've never come across an interference fit O-ring before.  Is the assembly going to experience significant temperature changes?  If so you will need to consider thermal expansion.

I don't know if thermal expansion will be an issue if it is already assembled...
My first thought as to a solution was to use grease, as suggested by rb1957.
I have not seen a seal retained by squeeze from the OD, but have from the ID.

Technically, the glass is always  full.

If the only time there is a risk of the O ring falling out is when it's being assembled, can you not just change the assembly method?
 

Most, if not all of the servo/solenold valves I've worked with the OD is slightly less than the o-ring for a very slight interferance fit.  sugest looking at the Parker Oring catoluge.  

See the actual size of the o-ring and size the c'bore 10 to 15% smaller to put a bit of compression on the o-ring.  It should stay.

Ted

As byrdj alluded to, you could simply use the gland dimensions found in the standard Parker book to determine your c-bore dimensions.  A dovetail gland is common for retention applications.  How did you arrive at 20% compression as your "pre-set"?

A word of caution about excessive grease used during assembly, and that is that additional problems can be propagated by it.  A very light coating of lubricant that is compatible to whatever material the o-ring is made from is all that is recommended, and that even depends on the material that is being sealed against (acids, oxygen, etc)

O-ring sealing is almost mundane in that dimensions and tolerances for nearly all gland configurations are already well established.  No need to re-engineer what has been time tested and proven.  An o-ring will simply not fall out of a correctly sized gland.

It is better to have enough ideas for some of them to be wrong, than to be always right by having no ideas at all.

Also look at the gland dimensions for rotary o-ring application.  The OD is sized to apply a squeeze to the o-ring.

Ted

From memory (I haven't looked at the Parker handbook for over a week now), and from practical experience, o-rings are best installed with a slight stretch on the i.d., and then compressed (squeezed) through the cross-section.  If you squeeze an o-ring such that the ring i.d. is reduced, you put compressive stresses on that side of the o-ring, and it may buckle or squirm, and create a leak path.  The buckling can (especially when the cross section is large compared to the ring i.d., which is true for a -011 ring)  be very local, creating small ripples along the surface.  We call that sphincterizing when we see it.

The dovetail groove is really the best option for retention, unless you can find a compatible grease sticky enough to hold the ring, and as ornery says, you need to be sure that the grease that inevitable leaks/extrudes out, and the dirt and debris the mess attracts, won't mess up other parts of your system.  
 

Face seals for internal pressure applications have slight compression fit on the OD.

Ted

What is truly amazing is that thirteen responses were pasted after Tetwin11 question, not one single answer. So make the assumption that most do not know how to solve the problem, or just don't get it!

The o-ring is 0.441 OD and is required to form a quasi face seal. It is a bastardized gland requirement sine no inner wall retains the element. Hence the obvious need to grip by OD for retention under it's own weight. I would begin by presenting the argument that the circumference of the gland is 1/128 inch shorter than the circumference of the rubber. The reason for this is that the maximum allowable stretch of any elastomer is two to three percent, typically. Do the mathematics, this is 0.028 inches.

But we need almost full contact between the elastomer OD and wall on the counterbore. The two percent is far too large, my experience draws on one-fifth the allowable stretch, say 1/128 inch as mentioned. This means the new diameter is (0.441X3.14156-1/128)/3.14156 inch=0.43851 inch, say 0.438 inch.

I would maintain the typical face seal depth for the counterbore. Face seal gland depth on a 011 o-ring is 0.050/0.052 inch, Parker specification ORD5700 catalog.

So your counterbore specification is 0.438/0.437 OD X 0.050/0.052 inch deep.

Regards,
Cockroach

I believe my first reaponse was to create a gland OD 10% to 15% smaller than the o-ring actual OD.  The math was left to the OP.

Ted

The OP was looking for the methodology behind the sizing of the elastomer identified by size.  The suggestion to make the C'BORE 10/15 percent smaller, and the reasoning behind this estimate as high as it is, lacks.  This would lead to the elastomer "rippling" and loosing wall contact, hence no seal.

I'm surprised nobody suggested stretching the o-ring prior to installation to a C'BORE of equal circumference.  The OP can still do this on the reduced diameter geometry as suggested.  Again, allowable stretches are 2% t0 3%.

10% to 15% is not realistic.

Regards,
Cockroach

tetwin11,

If it is possible, a face groove would work better than a counterbore.  That way the o-ring could be a slightly tight fit around the ID of the groove, rather than the OD.  It would seem to me that stretching the o-ring slightly would give a more uniform and consistent retention force than compressing the o-ring.

Regards,
Terry

Cockroach, the OP said nothing about sealing, just retaining an o-ring in a counterbore.  What do stretch limits have to do with compression or extra squeeze on the OD?  Your gland depth puts 28% compression on the .070 cross-section o-ring.

Ted

netwin

Althougth having a interference between hardware an the o-ring is possitive to reach your target, it's even also possitive for the o-ring to work

As said, a "dove" housing is a good point to start on this case, mostly cases I see is just adding some grease.

This means, by the way, an "extra" step on pounting, but is you need an o-ring on your hardware it will be most likely you'll need some grease on any step on your installation.

Other ideas, maybe will be througth some "extra" steps.

In some cases we use to provide with a "trapped" o-ring, being an o-ring together with a washer (can be made on plastic or metal) in order toavoid this problem.  But, basically, this is a solution we offer on this cases.

I don't know if you are prepared to get this solution, in this case, please, let me have a little sketch/drawing about how your idea on this housing will be.

 

http://www.hidromar.es

depending on how deep the cbore is, just make the hole .015" under the od of oring size. The oring will stay. Take a peice of scrap alum and drill a hole and try it. keep it simple.

I presume the purpose for the o-ring is for a seal, HyTool. Why else the elastomer?

What I'm saying is that you could theoretically drill a c'bore of diameter 0.441, thus matching the elastomer OD, then stretch the o-ring. You will find this solution would work, in addition to my post.

The c'bore depth is the recommended value from Parker. I suggest following that for this application.

Regards,
Cockroach

Cockroach,
Seal was be my first assumption, too, but that was not specified.
But could be for retention of something, or cushion for something.
We don't know for sure only that the o-ring was to be retained when the part is turned over.

Ted

Yeah, fair enough. Buddy went MIA, presumably he got his answer.

Regards,
Cockroach

Thanks everyone for your input. This is all good info and it will certainly help me come to a conclusion.