metal to metal sealing

[Andy330hp]

A project I am working on:

A polished piece of metal is to be pressed against another piece of polished metal. The second piece (the seat) is actually a cylinder, and fluid passes through the cylinder. Fluid pressure, if too high, lifts the first piece off the seat and fluid exits through the center of the seat. When pressure drops, a spring exerts pressure on the first piece, pressing it against the the seat and sealing the system.

The question I have is: what surface finish would be required to acheive a metal to metal seal like this? Pressures and hence seat pressures are low (10-20 psi fluid pressure, seating area of approximately .020 in^2 results in a rough est of 500 psi seat pressure). 4130 steel.

Thanks

 

[Onno]

The question about the quality of polishing can be answered a little with some product exmaples.
VAT Swizerland and Swagelok America produce vacuum valves with considerable leak tightness. In general they have do not polished their sealing surfaces. Surface roughness is low at about 1 micrometer(10^-6m) to 3*10^-6 m. And flat of course.
The valve block is generally stainless steel and the valve seat can be different materials but copper is a good option.
The balance of forces you show in your calculation I think is not right. 20 psi (1.3bar) is fluid pressure. The seat has an area of 0.02 in^2 so a FORCE is excerted on the seat of 20*0.02 psi*in^2. The PRESSURE on the seat is the same as the fluid pressure 20 psi.

 

[EdDanzer]

Actually the surface finish should not make a difference if the seal material is deformed when clamped. We have face milled mild steel surfaces (about 40 rms) and clamped 1100 series aluminum rings and have sealed nitrogen until the parts deflected enough to open and allow the gas to escape. Part deflection will lots of sealing problems.

 

[profengmen]

Another note, don't forget to pay attention the the flatness of the faces. They both must be as perfectly flat and paralled with each other.

Also, when specifying your surface finish, you must look at the lay of the machining marks to be sure they correspond with your intentions. If you are polishing the plates with abrasive slurries, the lay wont be a factor when you hit "mirror" type finishes.

Now, also, the rheology of the fluid might be a factor, it probably won't; but, I don't know what you are using. Very flat polished plates, when wet, will have a fairly strong bond. This may affect your "opening" performance.

 

[scooter911]

Check with a couple of the companies that make metal seated relief valves. Surface finish will be important but also depends on the geometry of the seat design. They can be a hastle to repair as seats must be in good shape, line up as designed, and completely parallel as pointed out before.