Aluminum fab project-advice / tips please

[therrera]

Hi fellas (and ladies lurking),

I am in the process of building two aluminum heater covers for a client. Since I didn't have access to a brake, I cut and fit all the parts and one is ready to weld out. The parts were cut on a shear at a friend's stainless shop but he couldn't bend it for me as it was too thick for his dies.

My experience with aluminum has primarily been on the repair end of things (tools, boats, engine blocks, tranny cases, motorcycles, bicycles, etc.) and not much on the fabrication end. The fab jobs I have taken on were larger projects where tube steel or angle was welded to form various parts for clients. The welds were never longer than the pieces being welded. Distortion was next to none.

This project is different because I have to weld long seams on some 3/16" plate and I am concerned how the aluminum will behave (twist, distort, etc.) when its welded out.

I would like if others that have already been here could tell me what to expect and how I might minimize distortion. For example I am thinking of clamping this thing to my steel table (5/8" thick) while I weld the long seams.

It also occurred to me to put some braces inside (2 x 4's cut to size) with some bar clamps to hold the box square, etc. However I might be over doing it as this is not a large box or a welding procedure might make it unnecessary to take all those precautions (welding it in small steps, say 3" at a time and let it cool?).

The box shaped deal is 12"x 24" x 3" so it is kind of a stout, stubby shape. Being 3/16" plate, it occurred to me that it may not pull that much or very little. I plan to fusion weld the corners. On the flange surfaces I "veed" it out for penetration so that when I grind it flush, I won't grind the metal to nothing.

So what do you say? Any ideas for me? I am fitting the second one up and will be checking any replies. I will probably be a couple of hours before I'm ready to weld them both out.

Thanking you in advance,

Tony

[DSW]

Alum is one of those materials that almost always needs filler. My guess is that if you try and weld this up with no filler, you will get a lot of hot cracking in the beads.

You will probably want to keep the heat input as low as possible with this. Move around to different areas and take a few breaks to allow the material to cool as you go. A big factor will be using enough amps. 3/16" and tig means you will probably be running 200+ amps and going as fast as you possibly can to keep the total heat input as low as possible. For many it seems counter intuitive to run more amps faster to keep the heat down, but the idea is to get the bead started and done before the alum soaks up all the heat away from the bead.

I can't tell from the picts exactly how you prepped all those joints. Some look like tight but joints and those might give you some issues. I'd have set all the outside corners up as a V with just the inside edges touching, so all I had to do was lay the bead in the ready made groove. The flat but joints I'd have probably beveled so I had a similar joint for ease of welding.

[Coupebuilder]

DSW is pretty much on the money, my addition would be to weld from one direction to minimize twist. Typically you would weld one seam, rotate the "box" and weld the other. This puts the start of the weld on opposite sides and may induce a twist.
Try to weld from the same start point on each side..i.e. front bottom, do both welds. then go to back side and work from top to bottom for both welds. you may get a little warpage, but not a twist. Hopefully this makes sense.

[Kevin Morin]

therrera
I'd recommend against fusing any aluminum welds- its likely to crack as DSW mentions.
I'd weld the top side and vertical edges where there are V's formed between the sheared plates only after gouging all the tacks to a uniform hollow U and with minimal bulge.

Then I'd use 2" - 3" stitches and use carbide to gouge the starts and stops depending on your pedal finesse at starts and stops?

I'd suggest you consider looking at all the welds as contracting 'fist' squeezing inward on the piece- if you balance all the contractions top-bottom, left-right, back and forth then you will minimize the distortion. The longer the seam the tighter the forces or more contraction from the welds along that axis of direction.

My idea of welding this out would be to put a set of evenly spaced welds on all sides - then divide the spaces between them, and do the next pass, and so on until the entire shape is finished. Looking at the short ends, they would be done in the first pass so the remaining welds would all be along the long axis.

If the box distorts then use the welding table to clamp it to a jack frame and reverse the shape by bending against that distortion.

DSW also advises the welds be at high travel speed and I second this. By moving fast the net heat is reduced even if you weld at higher amperage to get the lead edge to wet out. If you have pulsation this is good place to use a high pulsed setting to increase the wetting function of the upsweep/rising pulse.

If the piece will be finished and needs to be painted or brushed then I consider a router to bull nose the seams on the top outside corners then belt sanding with buffing material to get a final smooth brush metal finish.

I would not suggest grinding off the grooved weld beads unless the grove is 90% of the adjoining material depth, and shaped with a U so there is fully adequate penetration of the remaining root face.

Hope this 2nd vote for DSW's post helps? I think I'm saying the same as Coupe and DSW but using different terms of reference, maybe reading all of us will make a picture that is helpful?

Cheers,
Kevin Morin

[dave powelson]

Water spray cooling with air blow-off, Tony....on and on and on.
It takes very little heat transfer into the large flat portions
to make them expand, then bow.

[therrera]

Hey fellas,

I came in to check the replies and I think I get the drift. Weld high amperage, space out the beads, weld in one direction, apply cooling.

I will start with all the short welds first, then clamp the part to the table and weld the long seams with skipped beads several inches long.

I am going to put water soaked rags on the part to soak up heat. I have been doing this ever since suggested to me on a cast iron part I was working on last year by one of the guys on this board.

All butt joints have been "veed" to get as close to full penetration as possible. I will have to grind the flange areas down as they will mate against another surface with a gasket sandwiched in between.

I will add filler material and not rely on fusion welds. I'll post pictures when done,

Thanks,

Tony

[therrera]

Hi all,

Well I finished the covers without incident. Thanks for your advice. The only part that bent slightly was the long flange on the underside of the thing and all it took was a crescent wrench to bend it slightly back so the flange would sit true, all of 30 seconds to do.

I did each one slightly differently to see how they reacted. The first one I clamped to the table and welded the short seams first and then the long ones using water soaked rags set on top of it to act as a heat sink. When I paused I would move the rags around and press them against the metal to cool it quicker.

Then I flipped it over, clamped it down and welded all the bottom flange area (which were beveled for penetration), skipping around first the long weld, then the short flange pieces, then the other side long weld, then the other flange and so on until it was welded out. Finally the long seam on the front bottom flange area. All the while I had stuffed the soaking rags inside the cover right under where I was welding to soak up the heat.

The other cover I varied the procedure, welding first the bottom side (with the rags stuffed in it like the first time). I clamped it down also. Then I flipped it up and did the back side small flange area and finally the top side last (I did this first in the earlier one). It behaved the same though and did not distort. I was adventurous and wanted to see what would happen.

I think the thickness of the material made this possible. I wonder how a lighter gauge material would have reacted. I imagine it would have been more critical to keep the heat down and weld it in little increments then let it cool.

Attached are the pictures.

Thanks again,

Tony