Welding Tips and Tricks

Hi,
I have a project where I need to splice weld a 3"x3" solid bar of 4140 steel (basically a butt joint). This is a kelly bar shaft for a caisson drill- subjected to lots of torque as well as axial compression and tension.

What I haven't been able to decide on is what joint geometry would work best. I am hoping I can leave the bar on one side of the joint flat and bevel all 4-sides of the bar on the other side of the joint. However, my hunch is to bevel all four sides of the bar on both sides of the joint. I plan on welding it with the bar unrestrained in a vertical position so I can weld each pass all the way around the bar (even heat input). It would be very desirable to have a finished product that does not require additional straightening. Any experience or input is helpful.

Additionally, I am not sure whether it is quenched and tempered or not. Anyone know a way determine the state of the metal if unknown? I do not believe the manufacturer exists anymore. Do others typically post-weld stress relieve 4140 if it is not quenched and tempered? I am just planning to go with a high preheat and cool it slowly with Kaowool blankets on it overnight. I am thinking I will use 11018 rods.

Any thoughts are appreciated!

Trying to weld a 3" bar is above my pay grade. I would suspect it's similar to most all other 4140 welding. Preheat is needed, 4140 filler is needed, slow cooling a must, etc.

When you say you're not certain of the present condition, is this a repair on an already in-service part or are you making a replacement from new stock? Obviously, if it's a repair the old stuff is likely already 4140HT (means already heat treated). Otherwise, unless you specifically purchased it as pre-hard, you're buying it in the annealed state. New annealed material would be the easiest to work with, but for something that size, might be a challenge if you need to have it heat treated. You'll need a facility that can handle that large part.

I'd need more info. Are you trying to add onto an already existing part with new stuff (1/2 new 1/2 old??)

As to identifying, if you're familiar with any ole mild steel, like 1018 for instance, and you have a feel for that stuff, 4140 in the annealed state will be pretty similar. Hit it with a hammer, it'll dent, corners will scuff easily when hit with other steel, etc. If it's hard, max toughness would be around 42-44 rockwell C, so that's where they'd make a big drill part like you're talking. That would be pretty darn tough stuff. A hammer on an edge would scarcely leave a mark, a file will cut it but not easily. I'm trying to think of an analogy that'll work if you're not familiar with various steels. Let's say 3" square 1018, if your boss said here's a hacksaw, cut that in half for me. You'd grumble and grab the saw and get to it. If it's 4140HT, you're probably gonna quit your job before it's cut.

I would guess that beveling two sides would be best. That way you can have less starts and stops compared to 4 bevels.

I would think a v-groove weld on two parallel opposing faces and potential backgrind of the second face?

Not a pro, but that is what I'd look into.

If you have "Metals and How to Weld Them" by JFLF there is a chapter on welding alloy steels that could be useful to you.

Bughunter- This is a repair of an existing in-service part by adding a new section (1/2 new 1/2 old). You're description of annealed vs hardened 4140 is very helpful.

Bap- I ordered a copy, hasn't come just yet. Doing a V groove definitely sounds easier but I am concerned that it won't work well mainly because I have seen pictures of others who appear to bevel all 4 sides. I am wondering if beveling all 4 sides results in more uniform material strength/properties when resisting torque vs double v-groove. I have a hunch, but I am really trying to find some concrete reasons to use a specific joint geometry.

Ya I understand your concern there. I took a quick skim in "design of welded structures" as well but I didn't see anything that was similar to this application. Maybe try to dig around looking for typical weldments or pre-qualified procedures for torsional loads. Maybe the groove 4 vs 2 doesn't matter as much, but adding stiffener plates an all four faces would help? Good luck on your search, let us know how it goes. Interesting problem.

COwelder wrote:Bughunter- This is a repair of an existing in-service part by adding a new section (1/2 new 1/2 old). You're description of annealed vs hardened 4140 is very helpful.

Bap- I ordered a copy, hasn't come just yet. Doing a V groove definitely sounds easier but I am concerned that it won't work well mainly because I have seen pictures of others who appear to bevel all 4 sides. I am wondering if beveling all 4 sides results in more uniform material strength/properties when resisting torque vs double v-groove. I have a hunch, but I am really trying to find some concrete reasons to use a specific joint geometry.

As to joint geometry, I'm envisioning cutting the two parts on all 4 sides like you would a 4x4 fence post. Maybe a 15* angle going to a point in the center. Maybe less angle would be better because you'd have less to fill. Can't say there.

I think I'd investigate finding a heat-treat shop capable of nomalizing that part once you're finished. You could stabilize it with a torch well enough to get it there. Perhaps ask them how to handle your part best you can. I've never done a field repair on a 4140 part. I do weld 4140, but I have a heat treat oven here, and I make sure i only ever do stuff that'll fit in it. It just eliminates all the hassles of questions like these.

Even if I were to say "I know how to do this, bla bla bla", virtually every metal I ever heat treated has processes that vary depending on the thickness. Times and complete processes change. And I'd say they all start off straight away saying, do xyz time at such and such a temp, and then factor this time, temp and process for every inch of thickness thereafter. Sometimes just add time per inch, sometimes though not often change temp, etc.

Excuse my memory here if I'm mistaken... Lots of stuff has leaked out. I want to say I recall something like 85% toughness increase when using a best practice heat treat on 4140 vs the as-welded state. So, almost double the strength. And there's quite a bit of precaution to take when welding the already hardened material. Again, guessing here because I don't know, but I'd be inclined to wrap the ends of it in insulation, pre-heat the parts, weld, wrap the weld real good, and let it sit overnight to cool, then take it somewhere they could fully treat the thing so it was a best case process.

(EDIT)
I should also mention, large cross section repairs CAN be done and work well for a long time. I can't find the articles now, but the Alcoa 50,000 Ton forging press in Cincinnati OH was repaired back in 08-12. The bolster plates cracked and needed welded. Again, going from memory, but they were something like 3' thick by 12' high with 3ea laminated together with bolts, I forget how long, but really long, so BIG heat sinks. Then, there were 3 set of those, so 9 pcs total. The article said how many needed fixed but I don't recall, it wasn't all of em. They had to grind out the cracks and re-weld, at a cost of like $100M and expected to last a couple-three years. It's still in production use today so obviously somebody did some quality welding!

It was all robot welding, and obviously not something you take to the local welding shop.
I think it was a sister to the Wyman-Gordon presses. If you can find the articles on how they did it, it's fascinating. One big-ass press!
https://web.archive.org/web/20150201213 ... s/5662.pdf

Bap- really appreciate the help. My structural engineering friend recommended the Blodgett book but I have only recently committed to learning the theory behind welding, as opposed to just burning rods, and I have a lot of other books in-progress . Stiffener plates won't work because the shaft actually has to be able to pass through rubber hydraulic seals. I will definitely post an update with any progress.

Bughunter- Thanks for the input and interesting background. As of right now, I plan to bevel all four sides on both pieces, which may be overkill but seems slightly more foolproof. I do know that I want an included angle of atleast 45 deg. It might be more welding but I have difficulty getting a proper root pass in on deep joints that are any more narrow than that-plus I might try dual shield to speed things up which would require a little more access room. These kelly bar repairs are not frequent but they are also not rare. I may be mistaken, but I am pretty sure repairs are left in the as-welded state. Maybe regular use results in work hardening? Although I would guess that stress relief might be the more applicable goal of heat treating. As far as thick repairs working well, this is actually the smallest size kelly bar for caissons that I know of. Kelly bars twice as thick are typical (and have similar repairs performed on them).

COwelder wrote: I may be mistaken, but I am pretty sure repairs are left in the as-welded state. Maybe regular use results in work hardening? Although I would guess that stress relief might be the more applicable goal of heat treating. As far as thick repairs working well, this is actually the smallest size kelly bar for caissons that I know of. Kelly bars twice as thick are typical (and have similar repairs performed on them).

No, use does not work harden it. You're not trying, nor am I advocating attmpting to harden the final product. Quite the opposite, you want to stress relieve the part which will lower the hardness. It'll increase the toughness, but the hardness will reduce.

The welded metal, especially the pre-hard part, will be very hard when finished. It's tough to guess how hard, but i'd guess 54C Rockwell. You most likely want 40-44C. That harder area will be brittle, not strong like you want. Is it the end of the world? I don't know. It really depends on the application. If you have the opportunity, I'd ask someone who's done this repair successfully, how they did it. I bet they wrapped the finished welded part to allow it to cool as slow as they could. Possibly used a rosebud to heat it and stress relieve it afterward. It's hard to say. If they just leave it as welded, I'd ask what filler they used.

From what you said, obviously this won't be in a drill rig when you do it. And I assume you're going to have it finish machined to make it work with the seals for the hydraulics?

I disagree with Bughunters statement earlier that you need to use "4140 filler" or something like that. I have welding procedures that use regular ER70S-6 on up to 100ksi fillers.

If this repair is a concern where is your procedure? This will give you everything you need to know. Perform a PQR and you will be good.
Jonathan Lewis

I was making the assumption that his Original Part to be repaired was going to be north of 140 KSI and that under matching the filler metal was not going to be acceptable. He made it sound like this thing takes a tremendous stress in multiple directions. And that seems reasonable.

I was also thinking ahead, perhaps incorrectly, that if they didn't require the mechanical properties of 4140 they would have used a lower alloy Steel. Something easier to repair. But you just never can tell. It's not as if 4140 costs any more than anything else.

https://www.hobartbrothers.com/news/286 ... steel.html

Superior- The situation is a little more informal than PQRs and engineers, etc. No chance I would get the job if I opened up that can of worms.

I did manage to get a hold of a very reputable manufacturer who makes similar products and they had an informal procedure they shared with me:
-Material is 4140 HRHT QT
-double U-groove prep on both halves, oriented for vertical up
-weld with 110 KSI LoHy
-preheat 500-600F, maintain interpass temp above 500F, Post-heat at 500-600F (they do not have any specified hold time, they said basically just try to keep above 500 while welding and then heat it after to be extra sure)
-weld alternating sides, checking after each pass for distortion, adjust heat input/bead sequence to even out distortion.
-wrap in insulation and let cool overnight. Next day hand grind flush.

I thought the post-heat temp was interesting. I know tempering in that range is supposed to cause decreased hardness and toughness. I still do not fully understand the metallurgical differences of post-heat vs post-weld stress relief vs tempering...