Minimum micron finish for a good metal to metal fluid seal

[USAeng]

As the title said -

What is the minimum surface finish in your opinion that would be needed to obtain a good metal on metal seal for a fluid like a #2 fuel oil? Which is basically like water....

Right now I think one part is being machined at 16 microns and the other at 4 microns... but some drawing show 8 microns instead or 16 and yet another leaves out the 4 microns and may leave it up to the machinist to put on a "lapped finish" of his discretion as long as its 63 microns or better...

Thanks for any input

 

[JJPellin]

I work in a large oil refinery and am not aware of any applications requiring a metal-to-metal seal without the use of some sealant. Are you asking for a pure, dry, metal-to-metal seal with no sealant compound applied? It might be best if you describe the application in order to get replies from people familiar with that service.

With that said, we seal metal-to-metal joints with no gasket in water (up to 3,600 psi), steam (up to 600 psi) and lube oil (generally low pressure). But, all of these use a sealant and often use a joint design that makes the sealant less necessary. For our highest pressure applications, (water at 3,600 psi), the joint consists of a cast carbon steel head bolted to a cast carbon steel case that has the sealing surface overlaid with 309 SS. The bolting on the joint is designed to result in stress beyond yield in the SS overlay. Even then, we coat the surface with triple boiled linseed oil or a non-hardening sealant like Permatex or Copaltite. Some would argue that the sealant in this application is really used as a lubricant. But, we have no requirements for surface finish in this joint and it has never leaked a drop despite being disassembled over 20 times during the past 60 years.

Steam turbines often use joints with no gaskets on the case joints, governor valve to case, nozzle ring mounting, etc. But, once again, a sealant is applied. This is typically a hardening sealant, but some use non-hardening compounds like triple boiled linseed oil.

The case split line on a large multi-stage centrifugal compressor is also metal-to-metal with no gasket. But, they use a hardening sealant like RTV and deliberately leave a slight cross-hatch pattern on the faces to grip the sealant. Even then, it is common to perform a contact pattern check to make sure of full contact between the mating surfaces. Contact is more important than surface finish in some applications.

So, basically, the design of the joint and the use of an appropriate sealant are probably more important than the surface finish in the range you are describing. Additional information will yield more useful responses.

This forum might not be the best place for your question.

Johnny Pellin

 

[USAeng]

yes, it would be a metal on metal seal.... no liquid gasket or other sealant... it is basically a seal inside a unit that sprays oil. Two parts of the nozzle seal together and no gasket materials can come loose and clog the tip.... which I think was the thinking behind using the lapped parts to seal.

thank you

 

[USAeng]

oh and the pressure is about 60 psi +/-

 

[dgallup]

Depends on what you call a leak. We use 0.1 microns Ra, machine lap. That is for gasoline but we measure leak dry with nitrogen.