metal to metal valve seal

[LearningThings]

I am putting together a new manifold. The sealing is a metal to metal seal. The surfaces have a high finish - but I cannot get them to seal up. What I am thinking is - I are threading in a valve and the sealing is happening between a surface on the valve and a stationary surface in the manifold body. Is this even possible? Does the relative motion between the valve and the stationary body cause galling at high surface forces (valve being torqued into body)? Will this always be a low yield process? I are getting less than 50% yield (less than 50% of the units I build pass leak testing on the first torquing). When I go back and retighten - they sometimes seal and sometimes they do not. My question is twofold:
1. Is this just a bad sealing arrangement?
2. Is there something I should be putting on the valve before installing it - like a lubricant or sealant?

 

[gerhardl]

Metal to metal sealing is dependent on accuracy and precision both in machining, mounting and positioning/pressure when closing. Even small inaccuracies, hairline scratches or small impurities can give leakages, possibly increasing leakages over short time.

A better answer can only be given if you give full technical data and figures/pictures.

 

[LearningThings]

I wish I could give more details, but for legal reasons, I cannot. I can deal with surface finishes and such - my main concern was - rotating together two dry metal surfaces while torquing - will this ever work. Basically the threads are pushing the body down onto a sealing surface.

 

[Danlap]

Concept:
[ul]
[li]Almost all mechanical systems, artificial or natural, involve the relative motion of solid components. Wherever two surfaces slide or roll against each other, there will be frictional resistance, and wear will occur. Source: "Springer handbook of Metrology and Testing"[/li]
[li]one time Sealant or lubricant for metal to metal contact is normally temporary solutions (within 2-10 cycle/months/or years if lucky) since it will be extruded inside or outside/"scrapped"/phased out. At the end one will end up with metal to metal contact. It is nice to have in the beginning especially when the "secret component" as you mention will be in storage for some period (lubricant act as rust prevention, etc.). Source: logic and experiences[/li]
[/ul]
Your question in general:
Is this even possible? Yes, it is possible to achieve metal to metal sealing
1. Is this just a bad sealing arrangement? Yes, most likely or premature wear
2. Is there something I should be putting on the valve before installing it - like a lubricant or sealant? See concept above

Your question is not really clear (aside from the secret design). You mention galling however after re-tightened it works perfect again until some point. Galling is more plastic deformation (if its not wear), and re-tightened again and "works" meaning that it was elastic. This galling phenomenon normally happened between two similar/identical material with same hardness level mating each other.
How to manipulate? make one metal component as interchangeable item, this part normally have lower hardness level. When to change it (to write in your IOM)? suggest to do empirical test/observation on what cycle range it start to fail.

PS: it is not suggestible to put metal to metal contact as primary and only sealing component to the environment
Your name is discreet already, it would be helpful if you can make some analogy of your case. for instance if this is more of Orbit valve kind of type, etc.
I would not buy product that fail before even 100 cycles ever achieved.

Regards,
MR



All valves will last for years, except the ones that were poorly manufactured; are still wrongly operated and or were wrongly selected

http://www.eng-tips.com/faqs.cfm?&rat1=2&fid=376

 

[moltenmetal]

Yes, it is possible to have a rotating part seal against a stationary part in a thread-loaded sealing arrangement.

No, it is NOT optimal. All high pressure, high integrity sealing systems draw components against one another without having one of them rotate while the joint is being sealed. Take a look at a "cone and thread" connection- these are used to 100,000 psig. The parts are drawn together and loaded by a thread, but the cone on the end of the tube and the seat in the fitting do not rotate relative to one another as they are being drawn together.

Yes, a thread-loaded metal to metal seal where one part rotates as it is loaded, can be made to work much better with:
a) dissimilar materials, or at least material hardness, to resist galling
b) proper lubrication
c) a properly designed surface contact point and
d) the right surface finishes on the parts

An example is the needle and seat arrangement in a typical needle valve- they can be made to seal, though they don't last long typically. Note that better quality throttling valves have non-rotating stem tips.

 

[monkeydog]

What leakage rate is acceptable?
Helium leak rates at a MSLD level will not be achieved.
Liquid leakage of clean goo no contaminates is achievable.
You need to define your requirements for leakage to determine if metal-to-metal seal is a reasonable approach.

Also is this a one time actuation for sealing? If so, then even helium leakage is achievable.

 

[GAL02]

Good morning all !
I have never seen complete sealing in metal to metal contact of disc & seat ring,
always there will be considerable leak, these types of leakage are of Decent leak
and other is of allowable leak in drops/Min. or Ml. etc.

To avoid leakage or to achieve complete sealing at least one of the mating part must of soft material
so that it will act as a cushion to other but it will be of short life time & need frequent
replacement of it.

With regards