Welding Up a Tube Frame Table With Minimal Clam Shell Distortion

I want to weld up a modestly heavy steel table to build a 4x8(+) CNC router for sheet goods. I can of course take the attitide that sheets
goods are mostly just cut through so as long as the spindle roughly follows the warp of the table its mostly ok... and it really would be. Well for depth anyway. LOL.
I'd like to do a little better than that. I know I'll have to shim and space the rails to get it as good as I want it, but I'd like to start with the best table I am capable of. Maybe not the best some of the rest of you are capable of.
My thought is to clamp up the tube as best as I can to get it square, tack the 4 corners of each join whether its a t-joint or a brace joint, and then only weld in the longitudinal direction along the main runners and legs.
What else can I do? Yeah I know the tube isn't straight. I'll just have to adjust for that. Cross pieces cut to a stop, and legs cut to a stop may help, but I know they still won't make it perfect.
P.S. I'm strictly a hack welder. I've never taken a welding class in my life. Just hacked it out. I'd have to say most of what I know about welding I learned from his group. I have a flux core only Lincoln MIG, and decent Miller Mig (212 NOT Autoset), and an old Lincoln AC Cracker box.
I seem to weld the best with the little flux core mig, but the tube wall is right at the limit for it, and it has a very low duty cycle.
I mostly use the 212 for aluminum with a spool gun, but its dual stinger dual tank so switching to steel is just a matter of grabbing the other stinger. I have some tri mix on the stinger and I do not like how it welds. I've been told I'd be better off saving that for stainless and get a bottle of C25 for regular daily work. I could of course throw a spool of flux core in it, but I really don't want to run that nasty stuff in my nice machine.
On heavier stock I seem to do the best with the AC cracker box, but it also has a limited duty cycle. It doesn't shut down, but the welds start to get nasty looking. A ten minute cool down and its ready to go again.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On 12/13/2017 11:07 AM, Bob La Londe wrote:

snip


One thing that I have found is that the fillet weld on the inside corner of a square/rectangular structure is the worst one for causing warps. This is with TIG welding which may put more heat into the metal because it is slower, but I have built a bunch of things with square tube and seen it many times
snip


I have read that you are supposed to run pure argon for MIG weldng aluminum. I tried it and it still looked awful, so it's back to TIG welding for aluminum for me.
Good Luck, BobH
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On 12/13/2017 11:45 AM, BobH wrote:


I was thinking that, which is why I planned to not weld the inside corners at all, just the longitudinal seam.


I run pure argon for welding aluminum. My reference to tri-mix and C25 is because that is what is on or I would use on my regular stinger, and I plan to use my regular stinger for welding steel since this is a steel tube project. I might have mentioned my 212 is dual bottle & dual stinger and that all I have to do to switch is grab the other stinger. Literally. I get ok results with my MIG spoolgun/stinger and argon with aluminum from .08 upto .250 thinner can be done, but its miracle welding for me. Thicker can be done with lots of continuous preheat, but its a huge chore, and results vary. Its best in the sun on a HOT summer day. I've mig welded (badly) aluminum as thin as .043 (I think) and as thick as .375. .125 to .250 are pretty easy. Just have to use good prep and good technique.

Thank you. I am going to need it.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On 12/13/2017 1:48 PM, Gunner Asch wrote:

0" It is a frame to rest the bed on later. It will have cross pieces of tube to support an MDF bed or MDF Vacuum bed.

10ga

10ga
Not sure really how the thickness will affect the issue other than by degrees.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload


I don't see how welded tubing will give you enough stiffness to hold up whichever corner is unsupported by its leg, unless the side beams are trusses. You'll have to level it with leg adjustments, in which case you could simply bolt together the corners, plus some diagonal braces or gussets if the MDF isn't enough to hold the table square.
Once you've reduced the necessary welding to small attachment brackets on the legs you could mill their locating surfaces to eliminate distortion.
-jsw
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload


This details the differences between rigid "moment" and flexible "shear" connections. https://engineering.purdue.edu/~jliu/courses/CE470/PPT_PDF/AISC_ConnectionsJL.pdf
A welded moment connection locks in any welding distortion while a bolted shear joint allows the hopefully very stable floor to maintain the structure's geometry after you adjust it.
"Pinned" refers to shear and "fixed" to moment connections when you look up beam deflections in tables or use online calculators. Rolled channel sections look good in theory but their outer flange surfaces may not be square with the web. The sawmill frame I welded from square and rectangular tubing deformed when I welded the inner corners and had to be straightened with a jack.
Optical tables are very rigid and fairly light structures with extremely flat though not level tops, since they float on air cylinders that isolate them from floor vibration and deflection, thus they don't have to be on ground-floor concrete. https://en.wikipedia.org/wiki/Optical_table Notice how thick it is.
The upper and lower surfaces of the one I worked with at Mitre were 1/4" aluminum plates separated by a thick honeycomb which behaved as a 3-dimensional I beam web. Perhaps someone knows a practical way to make one. -jsw
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On 12/13/2017 11:07 AM, Bob La Londe wrote:









Using a deflection calculator I get a little less than .002 deflection on a single span with my planned bracing and leg interval with 100lbs center span. Since their are multiple legs and multiple spans and the likelihood of there being more than a few hundred pounds totalonthe whole table I think I'll be ok. It will more than good enough for sheet goods. Even with the weight of an MDF vacuum table. The worst single load point will be the gantry, and its total weight will be spread over atleast 4 linear bearing. More if (unlikely) if I go with Vtrack roller bearings. I'm planning to use square profile linear bearings.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

The limitations I encountered when making machine parts larger than my mill's capacity were my abilities to measure and correct length, straightness and squareness. For a one-time project length can be relative; like setting both ends of the track to be parallel.
My working standard for straightness is a 48" carpenters' level. It's straight enough as-is to align the sawmill but if necessary it could be checked and improved with a surface plate.
The sawmill's 14 leveling adjusters are arranged so I can turn them while standing up and watching the bubble. It would be much harder to do if I had to turn them at floor level. They are 3/4"-10 threaded rod with a foot on the lower end and a 9/16" hex for a socket milled on the upper, turned by a tee handle and long extension. They run in 1/4" thick angle iron brackets that I rounded and smoothed on a belt sander to protect my boots and pants as I walk past them pushing the saw head. -jsw
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On 12/16/2017 7:00 AM, Jim Wilkins wrote:

I like the idea of leveling feet that can be adjusted from the top. I might do that for some other projects. I'll think about it for this one, but I am considering selected a monolithic uncut section of the shop floor and bolting the machine down to it once its dialed in pretty close. Now a floating top that can be adjusted against the base sounds interesting...
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here. All logos and trade names are the property of their respective owners.