Controlling distortion during welding

[Welderpaul]

A few people have recently posted questions about controlling distortion during welding; a few have asked me questions by PM, so I'll try to help out a bit.

Whenever you weld something, the weld causes the metal to firstly expand, then as the weld cools, the metal shrinks, but it shrinks slightly more than it expanded. For some jobs, it's not always an issue, but sometimes it can cause problems...such as this job.

These are 200 x 75 channel floor beams. They have tags welded on every 400mm, all the tags are on one side and each have 260mm of weld round them. If we just weld these without pre-setting/pre-bending, or making some allowance for the shrinkage, the channels will bend inwards like bananas making life hard for the lads trying to erect it.

So, the channel is laid out on a spare beam and a packing plate is put in the middle - here the tags point up, and a piece of 12mm plate is used to pack. Clamped down tight each end.

Can see a bit better here, this is only a short beam, 2.4m long, but will still cause problems if it bends. Longer beams need a thicker packer - 25mm for the longer beams.
Clamped up and ready to weld (The trick is to get your clamp on first so the other bloke has to do all the work in dogging it down...)

Weld it up, leave the clamps on so it can cool a bit whilst you clean the welds, then clamps off and plate out.

Tadaaaa....perfectly straight!
Just another 23.9 tonnes to go.

I'll add to this as other jobs come up.

 

[eSCHEn]

Thank you for that Paul, very informative

So is the plate thickness and size determined by trial and error or do you know there'll be x deflection in a y shaped cross section over z length?

 

[Welderpaul]

Thank you for that Paul, very informative

So is the plate thickness and size determined by trial and error or do you know there'll be x deflection in a y shaped cross section over z length?

Very much trial and error; usually have a rough idea where to start, then do one and adjust to suit.
We have put quite a bit more packing in because there are two of us banging the welds in, not stopping, and un-clamping as soon as they are cleaned up.

If one man only is welding a beam, less packing is needed as the job is cooling quite a bit more and the weld isn't bending the channel so much.

 

[mike os]

thanks Paul

 

[balmerchris]

Very helpful, and a simple solution thanks Paul.

 

[blenkie1]

We face a similar problem with these

http://www.mig-welding.co.uk/forum/threads/bending-88-9mm-8mm-wall-pipe.57625/

All tacked up on a jig and all working as it should. Opening and closing ok. There is a locating pin with a 1mm clearance to help join the 2 sides together. Come to weld it together and try it and it won't close! Loosen the bolts at the feet at you can feel it move. What I do is heat one side of the pipe with the bottles and then let it cool and it usually moves to enable it to fit. You can sometimes actually see the pipe moving.

 

[blenkie1]

Like Paul said it's very much trial and error to how much.

 

[Welderpaul]

You can sometimes actually see the pipe moving.

Something I show lads when they ask - cut a piece of 20mm bar or whatever one meter long - hold it horizontally against one of the building columns so the end is touching it. Hold it loosely and put a big tack on the top of the bar and let go of it - you can clearly see the far end of the 20mm bar drop down, then rise up again.

 

[Spark plug]

Thanks Paul, you learn something new everyday. It's amazing how mutch something can distort, especially when your welding away under your lid quite happily then lift up the lid and the obscenities start.

 

[blenkie1]

Thing is you can correct things with heat but then it's knowing how much, you can start going the othe way! Best way like Paul says is to pack it. It get easier with experience to know how much but it's not an exact science.

 

[Robotstar5]

I did some work at a company that made sidesteps for 4x4's, the treadplates were welded to the top of the tube then some more welds run on the back just to take the curve out created by the top welds.

 

[GW8IZR]

Sorry if this is a stupid question but..

if you find by trial and error that 25mm packing obtain s flat result thats good.

If you did the same job without the packing would you see 25mm of bow or something less?

I think something less is probably the answer ?

 

[Turbo]

These are 200 x 75 channel floor beams. They have tags welded on every 400mm, all the tags are on one side and each have 260mm of weld round them.

How do you ensure the tags remain at 90 to the beam?

A guy I used to work with served his time in the shipyards. He was a burner by trade and one of his jobs was applying heat in the correct place to correct warping caused when welding the plate steel on ships. He freely admitted it was a black art, impossible to teach and only learned through experience.

 

[The flasher]

Good information there Paul. What make are the clamps in the picture. Cheers.josh.

 

[Brad93]

How do you ensure the tags remain at 90 to the beam?

A guy I used to work with served his time in the shipyards. He was a burner by trade and one of his jobs was applying heat in the correct place to correct warping caused when welding the plate steel on ships. He freely admitted it was a black art, impossible to teach and only learned through experience.

Tabs can be straightened a lot easier than a beam. And when they're torqued up to the beams they connect with or the bracing bars whichever it may be they will pull straight. Remember on structural jobs you usually spec a hole 2mm bigger than the bolt intended to go through it.

If you want @Welderpaul I could probably add some bits to this, sadly i am no longer working at my old TIG place making ridiculous tolerance stuff to military and aerospace spec I'm now back on the structural MIG stuff but the principles remain the same I guess.

The image attached shows a UB to which a shelf plate has been added, the common use for this is to run a course of blocks off. I forget the specs of the material now. What some of you may not realise is when a UB comes off the mill the two flanges are rarely 1.) Parallel to each other
2.) Perpendicular to the web
3.) The same width

The steel stockholders call this the roll tolerance and it can be quite large. For example today I worked on some 203x102 UB26 Beams. The flanges were actually 105 and he height was larger too.

 

[Brad93]

Good information there Paul. What make are the clamps in the picture. Cheers.josh.

Those are piher if I am not mistaken. I have a few

 

[Brad93]

To continue on from my post above. Beams often have a twist or curve to them, whether this is present when they come off the mill or happens when they are transported on the lorry to the stockholders with no support I dont know, what you need to do is work with the beam to give you the best chance of straightening it with your welding or being mindful and planning ahead with the use of presetting to ensure the beam comes out straight post welding and cooling.

You can see in the first picture attached the difference in curvature between two identical beams. The beam is set square at the other end.

In this second photo you can see the position of the shelf prior to further clamping and tacking.

 

[Brad93]

So that's enough about beams what about other more complex but smaller assemblies with a large about of welding?
This is a part for a car transporter, common design this particular one goes on one of the biggest manufacturers of transporters in the UK.
This is a ramp that is hydraulically actuated at the rear of the transporter (if my memory serves correctly) Here you can see it on my jig table, set up inbetween two piece of channel, one welded and one clamped to the bench.

Here I am welding in these pressed tread plates which the drawing spec'ed a full length weld. Problem, this weld goes onto a piece of 12mm flat stock stitch welded to piece of round bar, so it doesnt have enough material to, 1.) Dissipate the heat 2.) reinforce it against bending post-weld.
So you get your bracing bar, which consists of some bits of box spaced so you can still get the mig gun in there enough, to which you get your dirty great clamp and do it up till 'it feels about right'.
For some reason we found the material tended to do the opposite of what you would expect and bend up wards, it tended to sag and sink down. Don't ask me why, perhaps this is where welderpaul can advise.

After welding it usually came out pretty damn straight. Any bend could be sorted out with correctly positioned heating with the Oxy-acetylene.

This second part, also for the same transporter, was a right sod to get square, level and parallel.
This we used to weld up as square as we could, then level post welding with spacers, clamps and O/A. As you can see we had nice level surfaces to work off of... NOT! Cobbled together from two spare trestles and whatever scrap i could get my hands on.

 

[Brad93]

This one was not fun, in fact, I hated every minute of making two of these little frames.
The box section was either 20 or 25mm I don't recall, the gussets are stainless, 6mm requiring full weld. The legs are unsupported but the whole assembly must be flat, parallel, square and to within +/- 0.2mm so less than half a mil to play about with.
This one was sorted with the application of setting the legs out of square, welding opposite to where the frame was moving to, clamping to the bench, some presetting and lastly a lot of beating with a nylon mallet and checking on a surface plate for squareness and a 600mm Vernier for size.

 

[Woody.v8]