Help w/ Welding Distortion in Sheet Metal

This may be a pain to do, but what if you switch sides after each weld. If you do one side first you could be locking in heat stress that way.

Or, a welder on both sides to simultaneously weld the seams?

A really talented auto body restorer might be able to straighten them with a torch. $

Well, I think at least part of the trouble lies in the fact that one side is done first, but it's not really practical to keep switching sides. It's somewhat of a production product, and that would take far too long.

The seam welder only has one welding head, so can't do both sides together... which would be nice.

As far as straightening afterwards.... again, it's a production product, so I can't add individual craftsmanship and skill into each part. Trying to automate for a continuous and consistant run of parts.

Regarding one side vs. the other side though... does anyone have an opinion if it makes a big difference as to the direction that the two sides are welded? Would I get better results if both sides start at the same end? Currently, they are welding in opposite directions. I'm not comfortable with that method.

Should all 4 corners be tack welded first? Or would that give me worse results?

I don't think you should tack weld the ends. My advice is to start in the middle and go to one end. Then, it's anybodies guess what to do next.

1. Flip over to other side, weld the opposite way.

2. Flip and weld the same way as first weld.

Even making the passes from the same direction might help.

You'll need to experiment.

Maybe Milo will have some ideas.

Yeah, your comments are good sense suggestions. Achieving symmetry seems like one of the key issues here. There are some limitations because of the extremely simple program that runs the welding head. Currently, it only runs in one direction.. not backwards. I'm working to convince my employer to go thru the trouble to make it able to weld in both directions... if I can back up my theories with enough fact.

In fact, I don't believe it's even capable of stitch welding... only continuous welding. But I may be able to augment its capabilities by combining manual hand welding (first) with its automatic welding, if I can come up with the right process.

I'm hoping one of the gang will have particular experience in this kind of thing. I forget who is the master welder here.

One other glitch, that effects the practicality of flipping... the legs of the formed channels are not quite equal... therefore the weld is off center high on one edge, and off center low on the other edge. And height adjustment of the welding head is done manually, and it's cumbersome.

They're not makin it easy for me over here.

How much time elapses after completion of one side of welding and start of the second?

One approach would be to weld one side, let it cool back to ambient, then weld the other side. Ugly for WIP but maybe better result?

Well, the other side is welded as soon as the first one is done. Maybe a minute lapse. But currently, we are spraying water on the first side to cool it after the weld passes.

One more thought. It will add cost, but maybe tack weld both sides every foot starting in the middle of the part going outward. It's still important to weld each side alternately and not complete one side first.

Your goal is to apply the same amount of heat to each side as equally as possible.

Good luck.

There is one more fact I should add. What I'm forming is essentially a 2" thick slab. Inside this slab is a piece of polystyrene that fills it completely. It's acting as an insulator of course. But I'm not sure how that is affecting the heat transference during the welding, in regard to my warpage problem.

If you are creating a sealed container, the internal gauge vacuum created may be your issue.

No, at this point, the ends are open.

Then I would suggest you need to liquid cool before and after the wheel.

I assume lowering the amps or increasing IPM has not worked for you.

Yeah, IPM is maxed while still keepin it continuous. I'll have to check on monday, regarding how low the amps are set.

The top and bottom beams that clamp it in place are chilled. I wonder if leaving it in the fixture longer before welding, would help. Perhaps the chilled clamp would cool it down from ambient before we start welding. You think that would make any difference?

Probably not ... you are not going to chill it down that much relative to how much you are heating it up. You need to be thinking of taking the heat away ... or if possible ... reducing the heat in.

If only you could use the TIG process ;o)

My most educated guess, is that the machine you are using is designed for thicker gauge stock. If you are trying to make the best of what you have ... then go with active cooling.

I agree. The chilled clamps are likely not doing very much at all, that makes a difference. They already had this welder built before I got there, and are wondering why it's not working well. They built it themselves, probably without a lot of knowledge of the intricasies of this type of welding. I was hired to fix their quality problems. I've fine tuned a lot of their issues, but this one is a bit more challenging.

The cooling I spoke of, is a squirt bottle that a guy follows the welding head with. Not high tech. There is no type of steady stream.

My best suggestion would be to stitch weld the center on each side in a few places, by hand before the automated welding begins. Good luck.

That is kind of what I was going to suggest to them. Trying to get some additional opinions.

Although I have learned some of the theories of welding, I am not great welder. Welding in my job is only a rare occurrence. But, there are some REALLY smart welders on the site. Give the post some time to develop, and I would guess that there will be some very exact recommendations.

You should consider the stitch welding.

To do it properly you should have the control circuit modified. If you are using heavy duty water cooled SCRs it is not to great a problem to retrofit a variable pulse control to these, any good power control techs should be able to do this for you.

Then you could spot stitch both sides of your section with a final run for a full seam weld, it does increase your number of operations but you already have provisions for loading and reloading your machine so it would be an inrcease in production time, it depends whether you think it is worth while.

Another alternative is a prep bench with a manual spot welder to stitch spot your sections prior to your seam welder. It would have to be a fairly substantial unit probably gantry mounted and water cooled to keep up with your seam welder with the provision to move along a pre clamped section.

How far apart would you suggest I put the spotwelds (on both sides) prior to running the seam weld? It's about 6' long. I could have that manually done.

Also, keep in mind, the only cooling we currently have is the internally chilled clamp, full length, top and bottom (is this even doing me any good?), and a guy with a spray bottle wetting it down after the seam weld on each side. It's a very crude set-up.

They are open to making modifications (within budgetary consideratons of course), but I want to make sure I get it right the first time.

I'm hoping to spend about four minutes per part, on this welding operation. But I could increase that somewhat, if improved quality (flatness) is the result.

Why don't you set up a few units and try for best result? You dont have to invest in a new spot welder for this a small hand held one will do for a test.

There will be some rippling between spots this will determine how many spots you have to make to keep distortion to a minimum.

I'm surprised the guy with the squirt bottle isn't ruining your welds, cracking them with sudden cooling. You say the plastic is inside the box while the welding is being done? That's gotta make a mess. Can't it be injected later?

I welded for fifteen years, always tacked the job together before welding. On items where warping could happen, stitch welding was used, alternating sides to keep the stresses equal.

If you have the item in a strong jig to hold it in shape while welding, it may not warp too much, but it would have to cool slowly in the jig. This quick cooling makes warping worse. Some weldments are oven tempered to relax them, but you have plastic in there.

It's not plastic, it's styrofoam. Like a cheap disposable cooler.

Now as far as what you say about the water cracking the weld... I don't know enough about welding to comment. But I would like to hear others opinions on this. It's my impression that the cooling water is helpful.

I want to clear something up. The warping I'm talking about is very small. Like 1/4" over 6'. But for our purposes, this isn't acceptable. So it's really just a fine tuning that I need to find. It's not massive deformation that I'm trying to turn into 1/4". It's 1/4" that I'm trying to turn into less than 1/8". I'd like to find a way for my welding to be as cold as possible, so the part will appear as it would, if I had glued the two parts together.

Evidently, it is possible. There are door companies that weld doors together in a very similar way, and turn out perfectly flat 7' x 3' doors on a regular basis.

Hello Out of Box Experience.

I read your thread last week, and had little to offer so I kept my big mouth shut (I learned that from lynlynch!). As I read again today, about halfway through it occured to me: You are making a baby steel door! You confirm this in your post #23. Your comment "...perfectly flat 7' x 3' doors..." is not completely accurate: tolerances for steel doors for twist, crown, perimeter flatness, squareness, thickness are all detailed at SDI 117-09. Perfectly flat is sort of a relative term.

Hollow metal door manufacturers use a number of different techniques for pan assembly. Rather than describe them all, let me ask you a few Q's:

These two 'clamshell' halves, are they identical? If no, then explain.

Are you in love with whatever current design is or can the pan designs be altered?

Continuous seam resistance welded today... again, are you in love with that technique? You had said somewhere watertight... How much pressure? How long until oxidation is an issue?

I would be happy to discuss further, if anyone is still watching.

Yes, in a sense, I am making a baby steel door. Main difference is no hinges, and all 4 edges will look identical, and the size of course. Constuction is virtually identical. Except I'm not a door manufacturer. The process is basically what you would get if a non-door maker watched a door being made... and tried to copy the process to make a styrofoam filled slab that is very similar to a door, just different dimensions... shorter and narrower

It doesn't actually need to be a pressurized container. I was referring to the weld needing to be continuous, the entire gap needs to be filled, and it needs to be liquid tight from the outside, so it doesn't fill with water, in case it was in water or on water.

The shape of the two halves were also copied from a door. In a cross section of the profile, one side has a dog-leg, and the other side fits around that dog leg. Sort of like a lid on a tin box. The edge on the "lid" side is shorter than the edge on the other side.

Yes, I'm stuck with the resistence seam welder. As far as I know. I'm open to hearing alternatives, to give me a better sense of what a better method might be, but most likely this is what I have to work with.

And yes... of course "perfectly flat" was poor word choice for an engineer.

I thought your warpage was worse than that. Anything welded is going to warp a little, unless you go to extreme measures.

I welded for years, it was standard practice to straighten out the weldments by whatever means necessary. For your little warp, an experienced weldor can figure out how much to elevate one corner with a block, and stomp on the item to straighten it out. Sometimes I had to use a hydraulic press.

Well, I can't really be using corrective action on each individual slab, as it comes off the welder. I have to achieve some sort of consistency for a production product.

And regarding stomping on one corner, etc... these slabs are not necessarily twisted as a whole, where as I can simply twist in the opposite direction and regain proper tolerances. That might be the case if this was a single sheet of sheetmetal. But since this is a 3 dimensional assembly, made of two seperate sheets that are warping individually, and differently, during welding... the deformations are not always symetrical between the two faces. One face may have a bump in the center or high corners, for instance... while the other face does not.

I need to limit the deformations as much as possible during welding, rather than correct the deformations afterwards.

So I suppose that it is indeed the "extreme measures" that you mentioned, that I am trying to figure out. And at the same time use more or less, my existing equipment. I suspect that it is a process issue more than anything else.

Would you tell us how many of these you use? Hundred a year? Hundred a week?

Regarding process, yes I believe you are correct. Door manufacturers go through more than a dozen steps for vertical edge assembly. It sounds as if you are trying to reduce this to two, perhaps three?

"I need to limit the deformations as much as possible during welding, rather than correct the deformations afterwards." 100% correct. Very difficult to check quality in, quality should be built in.

Roughly 50-100 a week, for the next 6-8 months.

Well, there's more than 3 steps. Shear top half, shear bottom half, break both edges on top half, dog-leg both edges on bottom half, break both edges on bottom half, insert polystyrene, assemble halves, clamp and weld one edge, clamp and weld other edge.

Oh, my!

Perhaps 100,000 pounds of steel is scheduled through 2Q 2011.

Is making these panels your core business? If no, have you considered subcontracting this item out?

It sounds as if this type of suggestion may meet some resistance from others within your firm. I will point out that most steel door manufacturers do not even make all of their own parts...if there is a specialty company that does a faster/better/less expensive part or component, they will get the order.

We are the subcontractor. My company took the job, invested in making this seam welder, is now making parts with a very high scrap rate because of many parts not being flat enough and other quality issues... and now hired me to help make it better.

I have already made many improvements in material suppliers, shearing and breaking, calibration of these other machines, foam supplier issues, handling methods, and other assembly areas and steps of making these parts, etc... This welding issue is the last part of the puzzle that I need to address and improve, with the tools that I have at hand. Welding with a seam welder is the area that I am the least familar with... hence, my coming here and describing my situation.

Inserting polystyrene before welding doesn't look like a good idea to me.

Welding without calipers doesn't give you full control of the process.

At least if you define the process parameters, you can predict and control the outcome.

IMO, it takes more than just some free setup to meet standards.

Consider molds and calipers (preferably cooled) if you want to stay with your manufacturing by hand.

If possible, you should automate the process.

Isn't it possible to spot weld the frame, fit between a mold and extrude foam inside? Once filled with foam it is water tight, can get a caulk finish and paint. We used to make panels this way. No melting styrofoam involved also.

In case you really want to weld it off: you can make a contact wheel weld (again in a mould) but need to do a double one, starting from the same side, roll welding ends together - inside AND outside at the same time.

That is how we kept sections of telescopic cranes straight when welding. (sections up to 50 feet long, within 1/8") double side in and out, simultaneously.

Another thought. If the distortion is consistent on every piece, it might be possible to pre-curve the section, so the welding process straightens out the bend.

If the welding is done by an automated welder, it could work. If it is a human doing the welding, there might be too much variations to their welding to have that work correctly each time. Good luck.

Have you thought of using RF induction welding.A focused beam on both sides, you will be able to weld both at the same time and localise the heat to the area welded then instantly cool with a water cooled external mandrill, the part will remain flat if you make the mandrill long enough

It's something that I can look into. Thank you. But first I need to try to find a way to properly utilize the investment we've already made in the seam welder that was made just for this purpose. I think we are just not using it properly, regarding proper tacking or not tacking, proper clamping, proper cooling, proper type and order of welds, etc.

Here is our process:

~Currently, we are not tacking

~the part is clamped between two internally chilled clamps (long beams top & bottom, controlled by pneumatic cylinders)

~one long seam weld is applied on one side, by the seam welder

~the weld is sprayed with water from a squirt bottle

~the part is unclamped and turned 180 degrees

~reclamped

~one long seam weld is made on the other side

~again wetted down with some squirts of water.

~that's it

I'm not a highly skilled welder, but this simplistic process just does not seem to have the finesse or technique that is required, to achieve the high degree of flatness that I need.

I suspect that my chilled clamps really are somewhat useless, since they are clamping and chilling the top and bottom horizontal surfaces... while the weld is actually being applied to the vertical 2" wide surface of the edge of the part. And the part is only held by the clamp for 30 seconds or so.

I suspect that we'd be better off tack welding with a hand welder, one or both sides, prior to welding the long continuous weld. But I'm not sure of the extent of, or the positions of, those tacks.

I'm suspect of the effectiveness of the squirt bottle applied to the fresh weld, probably 30 seconds after the weld is done.

The lack of symmetry in the process troubles me.

All that I have to work with are a home-made seam welder that currently only welds in one direction, and is currently incapable of making stitch welds. And I also have a human with a hand-held welder that can tack, or make pre-welds.

I hate to sound like an idiot, because several suggestions have already been made... but I'm having trouble turning them into a specific by the numbers process, given the tools that I have to work with, that I just mentioned. I can suggest changes (to my employer) and some investment to make those changes, but I need to know what I'm talking about first, and have a process plan that is likely to work.

We are the subcontractor. My company took the job, invested in making this seam welder, is now making parts with a very high scrap rate because of many parts not being flat enough and other quality issues... and now hired me to help make it better.

I have already made many improvements in material suppliers, shearing and breaking, calibration of these other machines, foam supplier issues, handling methods, and other assembly areas and steps of making these parts, etc... This welding issue is the last part of the puzzle that I need to address and improve, with the tools that I have at hand. Welding with a seam welder is the area that I am the least familar with... hence, my coming here and describing my situation

Ahh!...I understand your situation, I think you might look at a Steff power feeder (or similar) for a consistent feed rate, do not use a hand held water spray after the welds as you will get localised heat displacement causing a warp instead use a pressurised water atomiser like they use for outdoor cooling in crowds in the mid west the ones that produce a fine mist that covers the whole surface they have ones that produce a fan shape that you might orientate along the weld.. again run into a external mandrill to maintain consistent shape..I could do a drawing for you if you need further explination...

Interesting. I sent you a photo of the current set-up in a message.

When all else fails, look at it like you should have in the first place, I am not a welder, physics is my specialty. if you have two different gauges of metal, one will be fine and the other will buckle, when you weld one side and go to the other you have created two different gauges because of the seam, which is why it buckles(on thinner metals). you need a heat sink, if this is production then making two large solid arcs from blocks of aluminium clamped on the outside when you finsh welding outside, then weld the inside.

Have a great day

Archie

Enerventure