What's new
What's new

Large Metal Additive with Weld Wire

JNieman

Titanium
Joined
Nov 12, 2011
Location
Greater St Louis Area
No, I'm not talking about the dumb shapes and little animals we've probably all mud-dobbered together on the weld table when we're bored. :)

But I've been researching more and more about Wire Arc Additive Manufacturing technology. The state of things seems to be pretty decent. The technical aspects all make sense. The material quality, with a good power supply and good "CAM" planning all appear, according to the tests I see, as-good or better than wrought metals.

So what I'm wondering is... for people who end up with low volume castings and forgings... are we going to start seeing "WAAM" near-net shapes to finish-machine rather than forgings and castings?

I know it'd sure take a hell of a load off the spindle time required to take a 2 ton block of 6061 down to 40 lbs, if it came in near-net at 50 lbs.

The lead time on some of these low volume castings we end up sourcing for our customers can be atrocious. If we can "print" a part per day, even, it seems like turnaround time would be pretty greatly improved. As always, everyone wants everything yesterday...

Has anyone else been aware of this type of thing?

I was at Fabtech last year, and Lincoln/Baker and Fanuc were certainly showing it off.

I'm curious how much sense it makes to bring in-house, or even to outsource to Lincoln, since they offer it as a service. Anyone out there have experience with it, or considering it?
 
The aerospace industry has been doing something similar with LASER and a wire or powder feeder for overhaul and repair for the past ~20 years or so. It is often cheaper to repair and re-machine a worn out part and add that to the existing paperwork than it is to get a new part.
 
Yea, imo, the powder stuff is pretty settled and being used anywhere it makes sense. It's very expensive for metals and the machines even more so. We have some products we've designed for use in our automated tooling, that we 3d print. Those are, iirc, a nylon powder, sintered, so not as resource intensive.

I definitely think if one were to look at purely saving the # of pounds of material to buy, it lends itself more to being used for the exotic stainlesses, tool steels, invars, and inconels of the world. Like TKassoc says, it's tough to care about 6061 or something, if it's so cheap, and you've already got good tools that rip the stuff out so fast. But some of those old aluminum bond jigs and such, if they weren't cheap aluminum I could definitely see wanting to crunch numbers on having a different way. I don't know.

Just checking some 316 SS type wire stock, Grainger sells 30 pound spools for about $11/lb. Quick quote online with a large stock plate on Alro puts it around $3/lb.

So if I was able to get someone to print a part out of 1/4 of the material, I'm already dollars ahead on material cost (obv their cost to print and deliver it is something additional) and I haven't begun thinking of the spindle time saved hogging out stainless steel. I don't know how some of your jobs are, but milling away 75% of the mass of a stock piece is not unusual in my experience. Turning 75% into chips might even be very low for some jobs.

That's just the first material I checked. Maybe I'll have to marinate on this one.
 
The pole endcaps on the rotors of the Dinorwig pumped storage scheme generators were partiall made with wire deposition and separate alloying element addition in 1982. It was one of my jobs as a student to take the CEGB inspectors around the fab shop when it was being done.
 
The pole endcaps on the rotors of the Dinorwig pumped storage scheme generators were partiall made with wire deposition and separate alloying element addition in 1982. It was one of my jobs as a student to take the CEGB inspectors around the fab shop when it was being done.

That is super interesting. Sounds like a pretty cool student gig. Bet you got to learn a good bit seeing one of the extremes of power engineering in action.

I think the addition of alloying metals selectively is a promising concept. Stamping, forming, press dies... all out of cheap low carbon steel, then capped off with equivalent wire to S7, D2, whatever, before finish machine and polishing.

I've never worked with some of those giant molds they use for wind gen
 
Yea, imo, the powder stuff is pretty settled and being used anywhere it makes sense. It's very expensive for metals and the machines even more so. We have some products we've designed for use in our automated tooling, that we 3d print. Those are, iirc, a nylon powder, sintered, so not as resource intensive.

I definitely think if one were to look at purely saving the # of pounds of material to buy, it lends itself more to being used for the exotic stainlesses, tool steels, invars, and inconels of the world. Like TKassoc says, it's tough to care about 6061 or something, if it's so cheap, and you've already got good tools that rip the stuff out so fast. But some of those old aluminum bond jigs and such, if they weren't cheap aluminum I could definitely see wanting to crunch numbers on having a different way. I don't know.

Just checking some 316 SS type wire stock, Grainger sells 30 pound spools for about $11/lb. Quick quote online with a large stock plate on Alro puts it around $3/lb.

So if I was able to get someone to print a part out of 1/4 of the material, I'm already dollars ahead on material cost (obv their cost to print and deliver it is something additional) and I haven't begun thinking of the spindle time saved hogging out stainless steel. I don't know how some of your jobs are, but milling away 75% of the mass of a stock piece is not unusual in my experience. Turning 75% into chips might even be very low for some jobs.

That's just the first material I checked. Maybe I'll have to marinate on this one.

I have a hard time imagining that the wire feed printers will deposit material faster than a mill can remove it, and en equally hard time imagining that the machines will cost less per hour than a mill.
 
I have a hard time imagining that the wire feed printers will deposit material faster than a mill can remove it, and en equally hard time imagining that the machines will cost less per hour than a mill.

I don't think it's helpful to compare the mill MMR to the deposition rate of a printer. Print speeds look like 10kg/hr and higher (only), so obviously, that's not a metric.

But they aren't the same thing. So it's not a comparison worth making. The time saved in roughing is just one variable in the mix here.

What I'm looking at is lead times and costs for castings/forgings vs having them made near-net via a WAAM process, and then making sure any other secondary benefits and drawbacks are accounted for.

If you're considering throughput, as in, making sure the printer can keep up with the mill, to maintain part production without stall, that's a different management problem that's not new, and easily solved. If you have a secondary mill op for a part that starts on a lathe, and the mill op takes 2 minutes cycle, whereas the lathe takes 20 minutes, obviously you aren't going to set up and run both at the same time.

I don't know what else comparing mill MMR to wire deposition rate would tell me.
 
Check out what Slade Garder is doing at Big Metal Additive (find the company on Instagram) for some large metal parts. He uses WAAM mostly with aluminum, because of the particular machines he uses.

Your own example of starting with a large block and removing 90% of it, is one of the drivers for the technology, especially with expensive exotic alloys. Also repair work, as someone else pointed out. The other, as you guessed, for low volume (onesies and twosies) parts that people don't want to keep on the shelf, you'll see more and more companies investigating additive manufacturing for that particular use case.
 
Check out what Slade Garder is doing at Big Metal Additive (find the company on Instagram) for some large metal parts. He uses WAAM mostly with aluminum, because of the particular machines he uses.

Your own example of starting with a large block and removing 90% of it, is one of the drivers for the technology, especially with expensive exotic alloys. Also repair work, as someone else pointed out. The other, as you guessed, for low volume (onesies and twosies) parts that people don't want to keep on the shelf, you'll see more and more companies investigating additive manufacturing for that particular use case.

I've been a fan of Big Metal Additive for a while. I was put in touch with another engineer there a while back. Very cool stuff.

There are case studies about being able to print train parts in Europe, where they had trouble getting reasonable quotes on forgings because they needed 2 parts, right? When you have to maintain a fleet of anything for 60 years... yea, you can't exactly keep a shelf full of anything that might break.

I think the current buzzword the hot-air-balloon speakers are using is "digital warehouse".

Or you know, what you and I would call CAD files.
 
There's a good intro-discussion article here about one of the really interesting points of an wire-fed additive process:
Wire-arc additive manufacturing can be used to build very large parts to near-net shape

Which is the ability to use different materials fully fused. Imagine laying down 90% of a mold with low carbon steel, and just applying the more expensive alloyed tool steel where necessary.

I think them material opportunities are pretty interesting - not just from a mass/cost savings, but also from being able to optimize a "hybrid-material" part. Takes the old cladding and cold-spray abilities to another level.
 
Well, I started thinking about getting going on this back in October. A couple months ago I started actively working on it here and there. We have a Universal Robots UR10 in a setup we do R&D stuff with, that was available... so I started with that, and making parts. Nothing I'd want to go for hours and hours on end, getting hot from constant arc-on time, maybe, but fantastic for the variety we use it for.

There's a lot of trial and error here. I managed to lay beads straight as an arrow, 3 ft tall, for a 'layer height calibration' run, sort of.

Managed to get weld settings and parameters pretty decently set for solid parts. This was a Monday morning special where I forgot to check the gas bottle before running... ran out of gas in an obvious location. Swapped the bottle and kept trucking. Other than the snafu, it was at least a valuable experiment.

Business viability is still big questions, but I think if this is a technology cheap enough, and easily-enough to use, there's going to be some uses. I've already had a couple customers express interest and amazement at a couple of the test pieces I've made. As I predicted before, one of them was specifically interested in using it to eliminate the lead times he has for short run castings, or to eliminate timely and expensive machining-from-solid before he goes into full production casting operations.

attachment.php
 
With the cost of high power fiber lasers falling I would think laser fusing the wire would be much better. No spatter to deal with.
 
With the cost of high power fiber lasers falling I would think laser fusing the wire would be much better. No spatter to deal with.

Laser melting is definitely a thing. There's not really a big difference in deposition rates that I can see, in my reading.

The other issue is machine complexity and consumables, in my opinion. The laser systems I'm aware of all require an inert chamber. I haven't looked at the laser stuff in a while, so I might be outdated. Places like MWES has their "ADDere" systems ADDere I | ADDere that are like that. That's what usually comes to mind. I saw them at Fabtech 2018. Good looking parts for sure.

As for spatter in an arc welding process, that is not nearly as much of a concern as my part might make it seem. Good robotic welding, when really dialed in, can exhibit super minimal spatter. Some welding processes are better than others, and as I prove, some people are better than others at welding troubleshooting. I am someone who just picked up robotic welding. I'm more from the engineering side with a machinist' background. We have some welding experts in house who said they could show me how to fix that stuff.

I know a lot of the automated welding I've seen with our customer parts has been super clean.
 
Well, I started thinking about getting going on this back in October. A couple months ago I started actively working on it here and there. We have a Universal Robots UR10 in a setup we do R&D stuff with, that was available... so I started with that, and making parts. Nothing I'd want to go for hours and hours on end, getting hot from constant arc-on time, maybe, but fantastic for the variety we use it for.

There's a lot of trial and error here. I managed to lay beads straight as an arrow, 3 ft tall, for a 'layer height calibration' run, sort of.

Managed to get weld settings and parameters pretty decently set for solid parts. This was a Monday morning special where I forgot to check the gas bottle before running... ran out of gas in an obvious location. Swapped the bottle and kept trucking. Other than the snafu, it was at least a valuable experiment.

Business viability is still big questions, but I think if this is a technology cheap enough, and easily-enough to use, there's going to be some uses. I've already had a couple customers express interest and amazement at a couple of the test pieces I've made. As I predicted before, one of them was specifically interested in using it to eliminate the lead times he has for short run castings, or to eliminate timely and expensive machining-from-solid before he goes into full production casting operations.

Is there any way you could post a larger image? I can't tell much from your picture, but it sure sounds neat.
 
Yea, I couldn't figure out how to attach a decent sized picture.

Here's the original size of the one I uploaded before. You can see the line of pores where I ran out of gas real well.
https://i.imgur.com/KS5F29J.jpg

Here's another part I recently did, and had half-machined. Still work to be done, but it's a good show-and-tell part for folks who are curious, and it helps internal conversations in the workplace a lot, when talking about my progress. Plus I learned a fair bit in the process.
https://i.imgur.com/rXZK5Ge.jpg

The next steps are just going to be getting the "feeds and speeds" right for this operation. It's a bit of a chore because while I have to dial in the "feeds and speeds" I am also still determining appropriate layer height and stepover of the toolpath, so it's a lot of variables to tackle at once. Slow and steady controlled experiments, though, make for good progress. Fail fast, persist, and eventually I'll get there.

The more I work with it, the more of a believer I am that this can be a good economical 'tool in the toolbox' for people. Not just the Airbus and GE of the world.
 
Though I'm pretty new to using Fusion like this, I had been working with a couple Autodesk techs when seeking information about DED projects, and was asked if I could put together a video showing what I'd accomplished.

It was spliced into the end of a Fusion demo on hybrid manufacturing workflow, where I just show the machine side of things. If you're curious, here it is:
IMTS spark Exhibitor Directory | IMTS spark

I think you have to sign up to view it, but it's free. That's all we get for IMTS this year :(

There's some video about how to teach the programmed zero position on the Univeral Robot pendant, as well as some timelapse of the first half or so of the part deposition. Then you can see the half-machined part toward the end.

Side note: I can talk to people all day, just fine... talking to a camera was weird and awkward.
 
I think it comes down to quantity, cost, and speed needed. For small quantities, especially when needed ASAP these processes will probably see increasing use, but for larger quantities it's hard to beat the economics of casting.

Metal deposition in its various forms is an energy intensive deal. Add that to the cost of the machines and prices tend to run quite a bit higher.
 
So you are using a mig welder on a plasma table to do this? My only concern is porosity in the weld.
 
So you are using a mig welder on a plasma table to do this? My only concern is porosity in the weld.

Using a MIG welder (Miller's RMD process right now) on a robot.

Porosity of the weld isn't an issue so far. The only issue so far is geometry that results in a complete fusion/overlap in all directions. I still end up with voids from lack of fusion here and there. I'm off this project for the time being, but when I have time to get back onto it, I'm basically planning to increase wire feed and bead size, and play with the stepover amount to find a good zone.

But so far porosity hasn't reared its head. I think it cools enough by the time the gas coverage moves away, that it's less reactive and less prone to oxide buildup that'd remelt into the next pass. Just a suspicion, not something I've measured yet.

I'd rather not resort to this, but a backup possibility is programming all the non-wall paths with a weave-motion.
 








 
Back
Top