Help Required Solving Brazing Application For Leak Proof Joint

[TorsionalStress]

I am currently trying to solve an application that requires a leak proof joint for air testing. The application is brazing a steel insert into a groove of a steel work piece.

The joints aren’t always leak proof with the current specs;

>Brazing Process: Furnace, Automatic with vacuum
>Joint Design: Combination of Lap & Butt
>Joint Clearance: 0.014” minimum
>Cleaning: Steamed
>Filler Metal: BNi-2
>Brazing Temperature: 1980 ?F
>Brazing Flux: None
>Atmosphere : Vacuum
>Flow Position: Vertical Down

My personal opinion is that the joint clearance is too big and maybe the fact that no flux is being used.

Any response will be greatly appreciated!

 

[flash3780]

Check the AWS Brazing Handbook for proper braze gaps. I have experience with nickel-based alloys, in that case the gap should be .002"-.005". With that in mind, your gap seems to be too large.
A proper gap is required for capillary action to draw the braze alloy into the joint.
As for flux - since you're vacuum furnace brazing, flux should not be required. If you were torch brazing it would be necessary.

 

[TorsionalStress]

After reading this article, it could be the filler metal type used.

• BNi-2, similar to BNi-1, allows good flow properties at lower brazing temperatures. *(being used)*

• BNi-3 flows well in less-than-perfect vacuum conditions. Ideal for tight fitting joints and for wide-area joints.

 

[tomwalz]

Agree with the above and would question the steam cleaning.

I really like a caustic cleaning that saponifies any oils or greases.

Thomas J. Walz
Carbide Processors, Inc.

http://www.carbideprocessors.com

Good engineering starts with a Grainger Catalog.

 

[TorsionalStress]

tomwalz, you agree with which above (flash3780) or (TorsionalStress)?

 

[tomwalz]

Both,

I would check the alloy and the gap.

In order, based on my experience, braze problems are caused by 1. cleanliness (including general surface preparation) 2. Wrong alloy 3. improper joint design.

Don't know what experiences the other two have had or what they have seen but their answers are in my top three.

I start with cleaning because it usually solves the problem and it is generally the easiest and cheapest. Steel is often coated to prevent rust. Steam cleaning can remove a lot of it but generally doesn't get all of it though it could dilute it considerably.

Try some Easy Off Oven Cleaner on the parts. It will saponify all oils, greases and protectants. They saponifed material will rinse easily.

Try a water break test on it with an eye dropper. If three drops form a ball then your surface needs cleaning. If they spread out into a flat puddle it is probably clean.

We have a video of this test on Youtube under Carbide Processors.

ANSI/AWS A5.8-92 gives 1780 - 1830F as the solidus and liquidus for BNi-2. Why take it to 1980F? Wouldn't a soak at 1830 - 1850 be better?


Thomas J. Walz
Carbide Processors, Inc.

http://www.carbideprocessors.com

Good engineering starts with a Grainger Catalog.

 

[TorsionalStress]

Would the recommended gap clearance be different for brazing unique shaped parts?

 

[tomwalz]

Dear TorsionalStress,

You truly do ask excellent questions. I've never run across this one before phrased exactly this way.

Braze joints get stronger as they get thinner. This is usually considered valid down to about 0.0005 inches. Below this they get weaker. This is a really rough rule of thumb and it doesn't get tested much because most braze joints in the common, industrial world are a few thousandths thick.

The traditional view of braze alloy was that it held the parts together. The current view of braze alloy is that it does several things. It holds parts together. It helps relieve stress on the part caused by differences in the coefficients of expansion. It also serves as a third phase in a laminated structure. In the case of my primary area, cutting tools, the braze alloy serves as an impact cushion. It is a soft cushion between the tungsten carbide saw tip and the steel saw body. This helps prevent the tungsten carbide saw tip from being damaged.

Over the years I have seen incidences where the steel surface of the saw body and the carbide surface were not perfectly smooth. The braze alloy joint may be thick enough to prevent breakage over most of its area. But if, due to the irregularities of the two parts being joined, you have the two materials touching through the braze or extremely close to touching you will most likely see fracturing at that point.

You would want to ensure that you had a proper minimum thickness of braze alloy between the two parts throughout the braze area.

Depending on how you introduce the braze alloy, you may encounter flow problems as the parts exhibit certain geometries. You are dealing with the flow of molten metal much as you would in metal casting. You will need to make sure that you introduce the braze alloy in such a manner that it respects variations in the geometry of the two parts.

What comes to mind is braze alloy flowing down a channel that then has to fill a channel running perpendicularly off to the right of the original channel. Making that right hand turn and flowing down the right-hand channel is, I would think, likely to be less successful than just having the braze alloy flow directly down the channel.

In brazing, the parts move. If you have irregularly shaped parts then they will move in a manner consistent with those irregular shapes. A square bar is much easier to fixture for brazing than a coil, for example. The square bar moves in a pretty dull, predictable manner. The coil, particularly an irregular coil, moves in much more interesting ways.

Depending on the shape of the parts and the difficulty fixturing the parts for brazing, you might have to accept an irregular braze joint thickness to ensure that you don't exceed either a minimum or maximum tolerance.

Having said all that, I would just go ahead and braise the two parts together in the simplest, easiest manner. Brazing is, by its very nature, a very forgiving process and the simple, easy method just might work entirely well enough.

Thanks again for the question. It has helped me look at a lot of old information in a rather new manner.


Thomas J. Walz
Carbide Processors, Inc.

http://www.carbideprocessors.com

Good engineering starts with a Grainger Catalog.