Justification to not weld 316L and 4140? 2

[geesamand]

Hello,

We have an application with a rotating horizontal shaft with overhung weight. We must weld a corrosion-resistant collar of 316L combination bevel/fillet welded to the 1018 shaft and due to the overhung load these welds are machined and polished. The collar is long enough vs. it's thickness that it is not full penetration welded.

Now we have an esteemed customer who requested 4140 instead of the 1018 carbon steel and our spec review did not catch this stipulation. Unfortunately there are ways to make 4140 and 316L stick together with welding tools (note choice of words) with no discussion of joint strength and I'm seriously concerned that doing this in this application will result in high notch sensitivity, corrosion sensitivity, and fatigue failure. While stress levels are low the stress is pure alternating fatigue.

Is there a welding code that could provide guidance (a categorical "no!" is acceptable) about the proposed 4140/316L design? I see a lot of information out there about how to weld 4140 but not related to 316L and highly stressed joints.

Thanks,

David

 

[stanweld]

You can weld the two dissimilar metals together with E/ER 309 or a Nickel base filler. Preheat at 400 F will be required. You may also wish to use a temper bead pass adjacent to the 4140. If you must stress relieve after welding, use the Nickel base filler metal; one can be selected to better match the strength of the 4140 alloy steel.

Whether the design with the intended manufacturing methods will be effective is another matter.

 

[geesamand]

That's the issue - I'm not cozy with the notch sensitivity of the 4140, and welding that's even slightly improper could make it much worse.

Above all the extra strength of the 4140 is of no value in this application so the 1018 is preferred for it's familiarity.

 

[sjrfc2]

We wound up with cracking problems when we mistakenly welded 4140 to carbon steel. Maybe the preheat will help but the higher carbon content in the 4140 seemed as though it was a problem

 

[stanweld]

Weld joint strength will be increased; however, increased stress intensity at the interface of the dissimilar weld metal and high hardness 4140 HAZ may prove problematic. A solution to the problem might be to butter the 4140 bar with Nickel base filler metal, stress relieve the buttered bar, near final machine, weld the 316 to the butter layer with matching nickel base filler, then final machine and balance.

It is understood that these manufacturing methods are more expensive than those used in your current design. You could inform your customer that whereas you have not manufactured equipment to their design, service life cannot be predicted. Service life of your currently designee equipment is predictive and you will extend your warranty accordingly if they will relent.

 

[EdStainless]

Stan's last suggestion is the way to do this.
Weld onto the 4140 with a general Ni filler (Inco 82 or similar). Remember your preheat. Then you could even re-temper the shaft at that point for stress relief.
After that you should be able to make your attachment welds for 316 to the Ni patches without damage to the 4140.

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

 

[geesamand]

Thanks everyone.

If for some reason we need to develop and qualify a 4140/316L weld procedure you've provided good starting point for that process.

It's the second part of the question (strength and toughness) that became the determining factor - lacking any data on the structural performance of the 4140/316L welds, we would need to develop a weld process and the joint would have to pass performance tests in order to be found suitable for the design. That is not commercially justifiable for this case. I'm sure I'll be asked again and this information will greatly help frame expectations of the customer.

In many cases we build our shafting out of whatever material the customer prefers for their process, which means switching between low carbon steel / austenitic stainless / duplex stainless / monel/hastelloys. The welding of those materials is trusted and well known here and the materials are fairly tough even after welding.

Rarely are we asked for high carbon alloys like 4140 to be used in a welded application which is why I was hoping to find data that gave me trust in the 4140/316L process. It is dangerous to assume that our product testing and development using low carbon materials can roll over and assure the 4140 will work equally well; I had to tell them no.

 

[metengr]

One critical bit of information in your post is missing, what is the heat treatment condition of the 4140? This was never asked nor mentioned in this thread. If the 4140 is quenched and tempered, one can obtain desirable strength and toughness.

 

[geesamand]

Our agreement with the customer did not specify heat treatment, only "4140". I assumed annealed because it's best for welding and brittleness or notch sensitivity are critical. There is no benefit in this application for the increased strength, whether it's 4140 annealed or Q&T. The literature I found stated that while toughness of the heat treated bar may be good, all bets are off in the HAZ.

David

 

[metengr]

geesamand;
Weel, just to help you in the future, you can temper bead a butter layer on the 4140 side of the weld joint as mentioned above and through grain refinement (using this controlled bead deposition process) one can obtain toughness in the BM HAZ that can even exceed the Q&T base material notch toughness. There is a time and cost factor involved.

 

[bprcwi112]

AWS D14.6 is the spec or code (not sure off the top of my head) for welding of orbital elements. Ive never used it but my company used it for a job before i got there so ive seen papers with it written on it. thats all I know

 

[racookpe1978]

What about using a three-part joint: 316L collar to a short (2-3 inch) 1018 section using your usual welding methods, then a transition weld joint of this 1018 to the client's final 4018 steel?