20 ga seems awful light for a metal chair, is that what came on it? Tack welding and filling in between with JB weld will give you a repair that is very susceptible to future damage from separation and/or cracking. (not recommended)
The following was in response to someone asking about patch panels on a car, but much of the same logic follows no matter what the sheet metal object is. Perhaps it will help...
Any welding process, be it O/A gas welding, arc, mig, or tig, is going to cause shrinking. Sheet metal, due to the limited thickness, is especially prone to this shrinking and the side effect of warpage.....no matter which process you used, gas, mig, or otherwise. With any sheet metal repair involving a long weld, you will be better served to first of all, insure a full penetration weld, and secondly, promote consistency in your methods. A joint fitment that is tight in spots and wide in others will require more filler (which means more heat) in the wider gaps, yielding more shrinking side effects for more warpage in those wider gap areas. You may not think panel fitment as a welding issue, but when it comes to how much filler is required, it most certainly is.
Welding in certain areas of sheet metal, specifically high crown vs. low crown, will tend to have different results in the way the panel reacts to the heat. Welding through a low crown panel will tend to shrink and further flatten out the panel, for an immediate loss of the shape. Welding in a high crown area (such as radius at the top of the quarter) still shrinks the area, but has a less noticeable effect on the shape of the panel. So in addition to consistency and improving skills and methods, if more than one option is available, you need to look at the area being welded and choose the weld location that will yield the least amount of adverse effects based on panel location and how the panel will react to the heat.
I'd suggest to eliminate some of this learning curve on scraps (same gauge thickness as what you're working on), and once you feel comfortable with consistent results, move on to the actual repair.
Fixing the shrinking side effects (warpage) from welding involves stretching, normally using a hammer and dolly. So that last bit to look at when planning where to put your weld should address access from the back side for hammer /dolly work. Not all circumstances will provide this access, but you can minimize some of the problem side effects by planning ahead. Any attempts at cooling an area with water will only serve to temper the weld, making it more difficult to planish, and risks cracking.
As far as a choice in the process, Gas or Tig is definitely the cleaner choice in welding in that they leave no splatter or slag all over the place. The amount of warpage is relative to the amount of heat you put into the panel via how long you sit there, so typically, MIG vs. TIG (when applying filler rod with the TIG), the MIG would have less of a HAZ. However, if you can trim your joints to zero gaps such that a no-filler weld can be performed (fusion welding) using the TIG or O/A, then you should have about as small a HAZ as possible, and as consistent in width as possible, for less distortion. To explain this further, various starting and stopping in your weld will cause inconsistent width of the HAZ, along with the shrinking effects that come with it. This may lead to a "wavy" distortion. Consistent heat, consistent speed, without stopping, in performing this weld will help to keep the HAZ, and all the other conditional reactions consistent, for less distortion. Having said all that, those conditions are hard to maintain when welding sheet metal with the MIG. Most enthusiasts will be using a MIG welder (as that's what they have) which is not very compatible with a full pass weld when dealing with sheetmetal. But you can still apply the same principles of consistency in using a MIG and "dot" welding. Single dots, skip around, use same overlap, same size dots (elapsed time of trigger pull) etc. This won't make the weld as nice as the fusion weld described above, but it will help with consistency if the MIG is what you're using. The biggest thing on consistency with the MIG is to practice on some scrap pieces the same guage as you'll be working with to insure your welder is set up correctly for a full penetration weld. It's hard to be consistent if you're putting in the patches and still fiddling with welder settings, (practicing on the good stuff).
Next, if you aren't very good at trimming the panels to fit tightly together, the O/A or TIG is going to be more likely to blow holes (depending on your proficiency to keep up in adding filler rod where needed); this is an area where the MIG is more forgiving as it is automatically feeding filler. The O/A and TIG processes prefer a tight joint for less chance of blowout. Lastly, TIG can sometimes be awkward to use in some remote locations (i.e.: under a car, any hard to reach location where using a foot pedal is cumbersome) O/A will be a bit easier as no foot pedal, and the MIG is more of a point and shoot type deal. The downside to the MIG over the other two processes is the additional grinding and consumables required to get rid of the excess weld bead. So I would consider all of the above in making your choice.
