Mig welding tips & tricks for Sheet Metal

I had a request for some info on planishing Mig welds, so I'd thought I'd post it here for all. I sometimes find it more difficult to explain something by starting with one of the last parts of a process, so we'll start at the beginning.


When looking at your welds from the back side, you want just a little bit of weld proud coming through the back side. This gives you something on both sides to planish (stretch) with hammer and dolly. If you still see the joint between the two sheets on the back side, it is a cold weld and is likely not flowing well into the parent metal but sitting up high on top. I will add that once you weld a cold weld, you can't fix it by continuing to add weld upon weld. I call this mud dauber syndrome. If the weld is too cold, it needs to be ground out and be re-welded.

Some hints to start with..

Welding sheet metal with a Mig involves what I refer to as "dot" welding. One zap at a time.

Before you even get to welding, this starts with panel fitment, you want gaps as absolutely tight as you can possibly manage. Every time you weld a mig dot, it is going to shrink in a circumferential fashion, pulling from all sides. If you leave a gap, there is a better chance than not that the panel is going to shrink together more as the gaps pull together. On low crown panels such as quarters, door skins, HOODS, this means that as they shrink the gap together, it is removing some of the crown that supports that panel. Planishing out welds is enough work as it is without adding having to compensate for panel movement in the mix. This is a sure fire recipe for a nice big oil can in the middle of the panel. So, tight gaps will eliminate some of this work. The panel will still shrink at the weld, but it will be nothing like you'll see with panel movement.

Welder set up... You should see a full penetration weld after each dot, meaning a slight bit on the back side as well, no panel joint visible. Use practice pieces of same material thickness, but insure you don't lay it on a steel welding table as that will become a heat sink. I like to duplicate conditions that will be present on the vehicle, so the practice piece should be in mid-air with no backing, just like most panels on the truck. So I normally tack practice panels together with a couple tacks and then span it across the jaws of an open vise. This should give you as close as possible what you'll experience when you do start in on the good stuff on the truck. Now undoubtedly you will have a good chance that you'll blow some holes. If you have the heat set high enough for full penetration and are blowing holes, then add just a bit more wire feed speed. If there isn't enough filler going in for the heat applied, it will burn away the parent metal. So try more feed speed before jumping on the "too much heat" chicken little. Once you get a happy medium, if you see your welds are too large, both front and back, try to limit weld sizing with shorter zap times. I'll also add that you should trim the little weld blob that forms at the end of the wire before performing each weld. This gives the machine better continuity for the electrical connection, for a more consistent start, each and every time. Consistency in every part of the process, from fit up to final planishing, gives us a more consistent panel in the end.

Once your panel fitment is perfect, and you're ready to weld in the panel, you need to tack one end, skip a few inches, and tack again, skip same distance, and continue from one end of the weld seam to the other, aligning the surface of the two panels as you go. Many people will tell you to skip around to keep heat buildup down, and I have been one of those. But to demonstrate why this process needs to be clarified, if you were welding a lower door patch (for instance) and tacked either end, you have a much better chance of panel misalignment than if you move progressively across the panel. Panel misalignment will result in dips, buckles, etc. that will not come out. So the more we can do for better panel fitment will reduce the need for excessive filler usage. Continue with the same sequence until the full length of the weld has been "tacked" in place. Then go back and planish each weld dot one at a time, in the same sequence as they were put in. Then use a 3" cutoff wheel to grind down the welds to just above panel surface. This gets them out of the way for planishing the next sets of weld dots. After completing the grinding, weld dots in between the first sets, repeat the planishing and grinding phase. When the weld dots are spaced at about 2" apart or less, revert from welding between the previous to overlapping the weld dots somewhere between 1/3 to 1/2. The overlapping will help to prevent missed spots and help to eliminate pin holes. Weld one cycle, planish, grind, repeat. On a weld seam I like to save the sanding flush for the end. If you sand flush after each individual dot, you will have a better chance of inadvertent sanding of the parent metal to the sides of the weld, resulting in thinning the panel.

Here is a video I did on grinding plug welds, which will be the same process, except for where we stop at the end. A plug weld we go ahead and clean up to the finish, for weld dots as part of a weld seam we leave final clean up to the end, as described above. But this should better show the grinding process I use.

https://www.youtube.com/watch?v=V2WHT_zMOE8

Now for planishing, I normally recommend this test, as many people will have different welding wire (some softer and more responsive to stretching than others), some may be more effective at hammer and dolly work, etc.

I'll preface this by saying that this "test subject" serves as a guideline only, and is intended to help you see the effects of shrinking and how the planishing counteracts those effects, more so than establishing a hard and fast measured amount. But it should get you in the ballpark of a measured amount to keep the panel in relatively good shape with minimal warping effects during the initial planishing efforts.

For your planishing test subject, you need two sheet metal strips about 1" wide by about 15" long. These will be tacked together on the long edge, and works best if you can use a shear or find someone with a shear, as you can cut a piece 2" x 15 and then shear it through the middle for a perfect seam. For you to see the effects of the shrinking and then the planishing, it needs to be a perfect cut through the middle for best results. So if you can find someone with a shear to help you out, it will help tremendously.

Next, the process and specifically amount of planishing needed is going to be directly related to weld dot size and/or wire type/softness, etc.

Now that that test piece has been cut on a shear, take your two pieces and align together TIGHTLY along the long sheared cuts and tack the seam at about one inch in from one end.

http://i5.photobucket.com/albums/y167/rmccartney/Metalworking/Picture102-1.jpg


DO NOT PLANISH at this point. Go another 3/4" and add another tack. Is there any change in panels positioning? Go another 3/4" and add another tack. What we are trying to do with this process is to monitor how much shrink is occurring. The first tack should "anchor" the two panels together at the end. With the panels tight together, each subsequent tack, and the shrinking effects realized once cooling takes place, will start to pull the panels even closer together, and also start to shrink the overall length along the weld seam, where they try to overlap each other. So If you haven't seen this happen, keep welding dots at 3/4" spacing until it does happen.

Now that you see these panels overlapping, the next phase is to see how much planishing it takes to "undo" the overlap. Start at your anchor tack, and hammer and dolly once. Go to each subsequent dot and apply the same hammer and dolly in the same approximate force. The flat should be similar in size, but I'd gauge your effort more on hammer force than size of the flats.

http://i5.photobucket.com/albums/y167/rmccartney/Metalworking/Picture105-1.jpg



When you get to the end, check the overlap to see if it still interferes with adjacent panel fitment at the un-tacked end. If so, start at the beginning, repeating one dot at a time, monitoring overlap. When your panel overlap issue has been resolved, your weld dot planishing effort should be the number of hammer strikes per weld dot as it took to resolve the overlap, using approximately same striking force. This assumes your weld dots don't mysteriously grow in size to add the need for more planishing, so again the importance of OCD consistency. This planishing effort will not be the end of the metal bumping to your panel ie: once you get welding in your patch panel/hood scoop/etc. It is the minimum needed to relieve the shrinking effects so the differing forces will relax a bit. So when you get to a this process on the "real" panel, the planishing will go ahead at the number of XX hammer strikes to get it done, then move to the next weld dot. What we're doing here in striking one dot once and then the next is only for test purposes to identify the number of strikes you need. After this initial planishing, any remaining planishing needed after the entire weld has been completed will be based on what the panel looks and feels like, high spots, low spots, etc after welding, initial planishing, weld dot grinding, and panel reading is completed.


Now that you have completed this, just for the heck of it, go to the first anchor dot, and start planishing it and it alone. Keep repeating until you see the adjacent ends start to separate as the weld dot is being stretched. Look at how wide the gap is. This approximates the amount of extra effort needed to overcome the shrinking and panel movement that happens when you leave a gap that size in the panel. It also demonstrates the differing planishing efforts that will be needed for inaccurate and inconsistent gaps. Keeping track of what, where, and how much is the tricky part, and again stresses the importance of consistency in all the processes, starting with tight gaps at fit up. Everything that you can do to keep consistency throughout all the processes only makes the planishing efforts more consistent throughout, lessening the need to keep track of the errant what, where, and how much.

Many people will weld the entire weld and grind and planish afterward. Some will weld dots as we have discussed and grind them down before planishing. Keep in mind, each weld dot as it cools, shrinks in a circumferential fashion, pulling from all directions. By planishing these before grinding, we are more readily able to isolate that weld dot by it's lonesome, and the planishing effort is more effective in providing stretch to negate the circumferential shrinking.

The primary focus on planishing of Mig welded "dots" is to remove the shrink effects. The primary focus of grinding to just above panel surface is to get that weld proud out of the way for planishing the next set, as well as returning the panel thickness to as close to original as we can. Leaving large weld dots will increase the panel thickness by 3 to 5X, which acts as a heat sink, and may cause cold weld joints. Our welder is set up for panel thickness, any repeat weld dots should see that same panel thickness.

When you look at the back side of the panel, it has a surface as well, so the weld dots there should be ground to just above surface also. Attempting to planish around varying height weld dots will tend to distort the panel based on how it deflects at the time of hammer impact. So grind BOTH SIDES to free up the space for planishing the next set of dots, when planishing in this method you want hammer and dolly to touch one dot only.

And lastly, when you planish at the edge of the panel, it does not have all that metal surrounding all sides, so any planishing effort is going to move/stretch more as there is less panel to move. So ease up in those spots.


Quite a bit of this may have been jibberish without having pictorials for further explanation. If you feel pictures would help out, go to the metalshaping tutorial link in my signature, there are numerous pages that explain the processes in pretty good detail.

Nice! Even I could understand what you were doing and why. Thanks

This is great information. It goes to show you, you don't just haul off and weld stuff. It takes good advice like this to make a really good weld.

I am confused on the need to do any hammer work (planishing?). I take my time, keep things tight, weld exactly like Robert says, but have no need to hammer anything. My goal is to make two pieces of metal look exactly like it was only one in the first place. Hammering unless done VERY VERY exacting, would draw attention to the weld and require more bondo. Talking very thin sheet metal as in cabs, fenders, etc. I mean no disrespect, and maybe I am just not getting the need yet?

In any case, thanks Robert for taking the time to really explain this very intricate process.

If you've ever welded a vertical seam through a low crown panel such as a quarter panel, then you've noticed how the weld will pull into a valley. I think some may be under the misconception that because a panel used to be somewhat flat, and now that the metal has moved away from its flatness to such an extent, well obviously the metal has stretched. NOT! There are very few, if any cases where a welding exercise will result in the permanent stretching of metal. Sure the metal will stretch while exposed to the heat from welding, but as the area cools, the molecular structure re-aligns and becomes tighter, closer than they were previously- resulting in shrinking.

Looking at the cross-section of the quarter, it appears much as an arc. When you add the heat from a welded seam, in many cases as it cools and shrinks, it will lose some of the crown, (become more flat), and possibly start to oil can as the metal loses its shape. In the picture below, the red arc would represent the original shape, and the blue arcs shows what occurs as the weld shrinks; the arc loses some of its length and pulls the adjoining metal into a valley.

http://i5.photobucket.com/albums/y167/rmccartney/Metalworking/crown.png

To correct this, the hammer and dolly principles as explained previously are used. Use caution, no matter what type of panel you are working on, to not get trapped with "tunnel vision" and only focus on the immediate weld area. We should always keep an eye on the surrounding area as well, and check your progress with some long metal strips/rulers laid across the area to help see where the shape needs to move. On a quarter, I like to lay a long ruler across the seam to read the dip, and then also drag it along the panel parallel with the seam, to see any change in the crown from the unshrunk area and into the welded area. Hopefully this will give a better understanding to what happens when welding so we can use more metalworking and less filler.

Dang Robert, this is awesome. I'm a long way from metal work but will definitely save this thread until I get there. Would be nice if we could put this in the tech tip section as a reference for folks. I don't fully understand how to do that yet. Thanks for taking the time for a detailed discussion.

In learning this process years ago I made many mistakes that I have had to go back in and fix properly. The biggest one was using lap joints. If you really want your cab sheet metal, etc, to be clean looking when you are done, butt weld everything. Its harder but Roberts instructions above will help with that.

All a lap joint does is trap moisture and make spreading bondo over it to clean up the area almost impossible. Easier to weld, not so good looking. Welding is a lot of fun and really gives a person a sense of pride (or not), but anything you can say you did yourself is very honorable. We are lucky to have Robert around to ask questions.

Robert, what do you use for an anvil...new good ones are pretty expensive.

Allen, an anvil as in a blacksmiths type anvil or are you referring to dolly for planishing welds?

The former. I bought the complete series of DVD's from Hot Rod Mag that takes a car from the salvage yard to complete frame off restoration including replacing complete floor panel, quarter panels, a couple patch panels, and other patches here and there. Those guys use a blacksmith type anvil an awful lot for flattening, rounding, punching, etc. Curious what you're using.

Mine is shown in the link below. I bought it used at an auction. I will say, it does contain some defects that would not be ideal for trying to keep sheet metal flat, so I typically will use the bed of my jump shear as a flat anvil for sheet metal.. one of those use whatever works..


http://www.garagejournal.com/forum/showpost.php?p=2832557&postcount=18

Can't remember for sure, but think the good ones are forged and the cheap ones are cast. Northern sells the cheaper (which I think means softer) ones for $100, is that a waste of money? The more expensive ones are $400 but still don't think they are forged. What do you think?

http://www.northerntool.com/shop/to...132&gclid=CNfz_uTX-r0CFZJj7AodGCIAEQ

Personally I'd prefer a vintage piece. Try some searches on ebay, CL, auctionzip, local antique shops?

Dollys can be made from a bit of rail scrap. Yep the stuff that trains run on. Cut into the shape you want on the band saw and smooth as needed. With a bit of thinking you can get a lot of shapes out of one piece of metal. Call it re-purposing and save the environment. ")

I have a friend who insists on Martin hammers. He searches high and low for that brand and wont use anything else. Hes a friend, and I appreciate his expertise and friendship.

A ball peen hammer, baseball bat, shovel, whatever that makes the proper profile will do. I have this cheapo HF 7pc hammer and dolly set. 4 hammers and 3 dollys. I might get that set out, but sometimes its a ball peen hammer and a long punch. What I like about this hobby is when you need to put a shape in metal, you just grab whatever is closest that fits the need. Sometimes its a Craftsman screwdriver. With the lifetime warranty, who cares if it survives. LOL!

Seriously, I have a new friend. Its a REAL anvil with something called a Hardy Hole. This 7/8x7/8 hole in the center of the anvil will accept any tool you put there. I have many very custom forms that I use for just about anything. You just get some 3/4" square solid stock and weld all sorts of rounded, squared, whatever shapes on it, and next thing you know you have a collection of really useful forms. I agree that the more antique hammers are way more useful, but in absence of them, you aren't totally SOL. The honest truth be told, the best restorations are not necessarily coming out of a high end restoration shop. Many of the top ones are coming out of single car garages all over the world.

I am hoping that someday, the vendors will get it, and stop using lamo, crappy composite handles on their hammers festooned with plastic grips that fall off and really get that we are happier as a whole with hardwood. Doesn't matter if its better or not.. more people feel better when they swing a wooden hammer!

Unless its at a loved one... anyhoo... I have a really nice piece of rail that works well too.

Thanks to Jim (Fourbrads) for making this a sticky. Hopefully it will be more easily found by any members new to MIG welding that it would help in their success.

I'm glad I found this thread. Now I know what I was doing wrong!
What can be done about "pin-holes" in sheet metal? I know if I try to do anything, I'm just going to blow a bigger hole in my 41 rear fenders. In some places the metal around them is pretty thin.
Robert, thanks for all you do!

If the pinholes are clean of rust, you can easily weld them up using a piece of copper (flattened copper pipe works well) as a backer. Turn the heat up higher than you would for just the metal thickness and give each pinhole a blip with the MIG gun.

Victor, The fender will normally show you better what is in the inside as far as more extensive damage, over an enclosed panel, so you lucked out there. Personal experience with pin holes and severe pitting has shown a hard rust scale inside the deep pits, and although I did not use this tailgate skin in the below pics, I did test some of the pits for removal methods of the scale. I could not remove the scale completely with either a wire brush or a crud thug. Only when using media blasting was I able to completely clean out the pits. Why is this important? The scale left inside the pits tends to re-activate in the presence of welding, such as someone fixing a couple holes from the outside, not knowing what lies in wait nearby. So although the holes showing from the outside appear to be manageable with a quick zap of the MIG, the inside view shows a completely different story, and demonstrates that this needs to be fixed properly BEFORE the paint goes on, lest you see more holes in a year or two. The next best part about media blasting, it will do a better job at finding weaker metal, ie: it was pitted and you're not sure whether to band aid it or not. If you have a couple holes showing, rest assured more are on the way.


[img]https://i5.photobucket.com/albums/y...20Restoration%20Album%203/Picture046.jpg[/img]



[img]https://i5.photobucket.com/albums/y...20Restoration%20Album%202/Picture546.jpg[/img]



My preference is to media blast any pits to clean metal, then make any repairs needed, up to and including replacement sheet metal. My experience has shown media blasting the best method to insure the pits are clean, as well as the best abrading method for either epoxy or powder coat adhesion. Now with the panel clean, you can decide whether this is a weld the pin holes closed or is it extensive enough to go ahead and weld in new metal. If we're still in the pin hole stage, as Kevin mentioned above a flat piece of copper does wonders to help control the heat and keep the weld in place without melting away more of the panel. If the area you're welding isn't perfectly flat, take a small section of copper pipe and flatten it with a hammer. This pipe is typically ductile enough that you can bend it to conform to the panel's shape...

Thanks Kevin and Robert for sharing your knowledge and experiences.
I'll keep working at it!