Low pressure sealing technique

[gmak100]

I'm having a bit of a problem sealing at low pressure, 5 to 10 psi difference from atmosphere(could be above or below). It's a static face seal.

I'm currently using two O-rings in two separate grooves, both with 17% compression ratio(0.103" CS o-ring in 0.085 deep groove). The other mating face is teflon coated 1/8" aluminum plate. There is little control over the surface finish of that, and may be the source of the leak. However, I've also tested the setup with 1/8" stainless steel plate with good surface finish and it will still leak at a very low pressure, sometimes even at 1 or 2 psi.

Should I use gaskets instead because of the low pressure? What are some of the steps I can take?

Thanks!

 

[flexibox]

If you follow the standard Oring groove guidlines you should not have any leakage what so ever. You did not mention what the Oring material is I will assume a rubber compound. It would help if you send a sketch of what you have because it may be a deflection problem with the plate in how it is clamped/bolted.
I would not use a gasket because the clamping forces required to achieve a seal are much higher than for an Oring.
Looks like your groove depth is wrong and not sure what your width is but check against attached and you should be fine. surface finish is important.
trust this helps

 

[gmak100]

The O-ring material is Buna-N, 70 durometer
The groove depth is 0.085" and width is 0.125". The o-ring I'm using is 0.103" in diameter, which should yield 17.4% compression (0.103"-0.085")/(0.103")=17.4%

The other mating surface is 1/8" aluminum plate or 1/8" stainless steel plate

 

[israelkk]

Not every one can open the latest Autocad files. Please use older versions such as Autocad 14 format.

How do you tight the plate? Are to sure the plate stays flat when it is compressed against the o-ring?

 

[gmak100]

Autocad 14 format file attached

the plate is tightened to the sealing ring with o-ring grooves by 20 evenly spaced #6-32 flat head screws.

 

[israelkk]

The linear distance between the screws is 2.13". This is way too large for the small #6-32 screws to create an even compression force of the 1/8" plate (too thin) on the O-ring. To my opinion the O-ring is compressed only near the screws but the plate between the screws is not able to evenly compress the O-ring and is raised by the opposite compression force of the O-ring. If you will look at the assembled plate you will probably find that in the mid way between the screws there is a gap between the 1/8" plate and the 0.875" thick ring.

I am not even sure that 20 #6-32 screws can create enough force to compress this large O-ring. You need to calculate the needed force to compress the O-ring 17% of its wire thickness. In addition you probably need a much thicker plate than 1/8" that will not deflect under the compression force on the O-ring. You may continue to use the 1/8" plate if you can add a thick ring above it to keep it flat.

 

[1gibson]

Torque sequence?

 

[gmak100]

We've tried various torque sequences and depending on where we start or which torque sequences, the leak point will shift.


Thanks israelkk! I was afraid that may be the cause. A quick and simple FEA with SolidWorks Simulation Express shows that the deflection between the screws is 0.00106" with the aluminum plate, which isn't much compare to the 0.018" of compression, is it(there should still be 16% compression)? The problem is that even stainless steel plate results in leaks as well, where deflection between screw is reduced to 0.00039".

 

[SamPorter]

Why 2 o-rings? Remove one and see what happens, this will cut your required compression force in half.

A thicker plate would do some good, over that large surface even a relatively low pressure is a large force.


SP

 

[gmak100]

the second o-ring is intended for backup. We've tested with either one alone and the leak still exist.

thanks for all the suggestions thus far!

 

[SamPorter]

Forgive me for the following basic list of suggestions, you have something weird going on and i cant figure it out!

We do low pressure seals all the time, some leaking parts are a result of as little as a hair across the seal surface.

Could there be loose contaminats disrupting the seal?

Maybe the o-ring is cut, damaged from installation maybe? These cuts can be almost impossible to detect. Try a new one.

Is the surface finish in the groove acceptable?

You could try a quad seal, that might work better for you.


Hope something in here helps...

SP

 

[gmak100]

I have my co-worker clean the surface carefully and try a few sets of o-rings as well.

How important is surface finish? Is 16 rms really necessary for air seal as suggested by Parker handbook? The groove is machined and then anodized so the finish on that is easily worse than suggested by the handbook, however, it is still smooth to the eyes.

Will surface finish from powder coating or equivalent surface treatment be not acceptable for use of o-ring seal?

 

[dgallup]

More important than surface finish is the direction of lay. How you produce the surface makes a big difference too. We consistently get by with a 50 Ra on turned parts where the direction of lay is that same direction as the o-ring. If the lay goes across the o-ring a much better surface is needed.

You also don't say how good the sealing has to be. There is no such thing as zero leak. What is your requirement?

 

[Gorpomon]

Reading these threads I don't see a target bolt load, which for an o-ring would be small, but still necessary, probably best would be the load for max compression of o-ring or 75-90% bolting material yield. If that still leaks then increase bolt size or use a rubber gasket. I think if you consider bolt load, bolt size and surface finish you'll find your solution in those three, and also remembering dgallup's point that all seals will leak some.