Need low Magnetic Permeability welds in 316L Stainless

[MikeMech]

We have an ongoing job that requires welding 316L base metal with 316 alloy. Magnetic permeability of the entire finished weldment must measure less than 2.0mu on a Severn Gage.

Our customer supplies the base metal, we do the welding. We have had infrequent but recurring problems over the years with seemingly random welds failing the 2.0 permeability test. The welding and testing is being performed by a competent, experienced individual. Why is this happening? How can we best ensure low magnetic permeabilty?

My current theory is as follows: Magnetic permeability is proportional to ferrite percentage/ferrite number. The base metal and filler metal chemistry ranges specified could result in an unacceptable ferrite percentage if the Nickel Equivalent is at the low end of the allowable range and the Chromium Equivalent was at the high end of the allowable range. A Schaffer-DeLong type diagram indicates this. Statistically, this is unlikely in any particular instance but probable over the long run. My first attempt at a "silver bullet" fix would be to purchase only filler wire with very low, but still within allowable tolerance, ferrite numbers.

I will be performing some experiments with 316 filler metals having ferrite content throughout the allowable standard range. Any advice on weld procedure improvements to try that may help? Annealing after welding is not an option, nor is changing the material.

Comments appreciated!

Thanks,

MikeMech

 

[MagMike]

This thread: thread1135-117830
dealt with magnetic permeability after mechanical deformation. I realize you are talking about welds, but the key factor seems to be nickel content >12%

Since you can't do any post-weld heat treating, your best option is to control the chemistry.

Good luck. Please let us know how things go with your ferrite content experiment.

 

[MagMike]

Hello MikeMech,

Came across this paper:

http://accelconf.web.cern.ch/accelconf/p91/PDF/PAC1991_2322.PDF

It deals with welding accelerator beam tubes at Los Alamos. They found using 310 welding rods on 316L tubes produced permeabilities in the range of 1.02 to 1.05, without post-weld annealing.

Hope this helps.
Mike

 

[MikeMech]

Thanks for the cold forming thread, interesting reading. I came across the same cern paper in my research.

Here are my preliminary rough test findings:

I tested 3 different 316 filler wires on 316L base metal. Call them A, B, and C, in decreasing order of ferrite number. Three different weld configurations, similar to those used on the weldment, were set up for each filler wire. A total of about 20" of weld deposited for each filler wire.

My Severn Gage has inserts for 2.0, 1.6, and 1.2mu.

1. All base metal passed at 1.2mu.
2. All welds passed at 2.0mu.
3. Filler A failed at 1.6mu in two areas, about 6% of the total weld length.
4. Filler B failed at 1.6mu at only one point location.
5. Filler C passed at 1.6mu completely.

While the data sample is limited and the measurement rather crude, it seems to support the theory of low ferrite = low mu. I'm trying to see if I can get some Severn inserts for 1.4 and 1.8mu to increase the accuracy of my measurement.

I may have to buy a roll of high ferrite wire to round out my data, even though I would probably never use it. While all the samples would have passed the 2.0mu criteria, it still is frustrating that the spots that did fail at 1.6 seem to have no common factor that would explain why they are higher in permeability than the surrounding weld.

Assuming relatively low ferrite in the base and filler metals, I seem to be creating "extra" ferrite during welding, but not consistently. Arg!

 

[EdStainless]

You need to find some 310Mo wire. The real high Ni will tip it in your favor. You might even think of trying 904L filler.

As you move to filler with lower Ferrite number be aware that you are at higher risk for hot tears and shrink defects in teh welds.

= = = = = = = = = = = = = = = = = = = =
Plymouth Tube