r3v Welders - Heat Straightening Help?

[bmwman91]

I need some input from r3v's expert welders. A project I am working on (personal project) involves a 3x3x0.25 rectangular mild steel spar ~72" long with some 1x1x0.12 rectangular cross bars MIG'ed in on one side. Since the cross bars are welded in waaay off of the main beam's neutral axis, I now have a nice warped spar lol. So I have done some reading and found guidance on heat straightening in some ASTM and DOT papers, and while I certainly would not trust myself to do this on a structural member, I am happy to give it a shot on this.

So I have two ~13000 Btu/hr MAP/PRO torches and I rigged them up to apply heat symmetrically. I don't have an IR thermometer or temperature crayons, so I am sort of going on color.





Based on what I have read, mild steel of this color (below) should be somewhere in the 1200-1400F range. The first pic is from the inside (looking down the end of the tube) so I know I got decent heat penetration, and the second pic is on the outside. The actual color is dimmer and more reddish than shown since it is hard to get a camera to show this sort of thing well.





The white papers on this seem to indicate that 1250F is the upper temperature limit that one would shoot for so as to not cause deleterious effects to the grain structure and mechanical properties. I did a couple of rounds of heating prior to this, to zero effect, so I figured I would try heating a little more. The above pictures produced zero effect after cooling. Essentially if I lay my 36" straight edge over the center, I have a 0.050" gap, and heating has not changed it at all. Since I have the 3 cross bars, I need to straighten things in 3 places, once I figure out WTF I am doing anyway.

I did do an additional round of heating after these photos, and in that one I heated the sides so that about 2/3 of the height on both sides was a dull orange, along with the whole top side similar to the pics. That also produced zero straightening effect (0.000" improvement after cooling).

So, what am I doing wrong here?

[gte175x]

I'm just thinking....The warp is it's current lowest energy state @ Room temperature, so how would heating help flatten it, without load? So, I think we need to think were the load would be coming from.

I think you should heat the weld - or where the compression is. the resulting internal tension loads should pull it flat.

[ForcedFirebird]

If you slice half way through the material on that line you made of cherry red metal, then weld it back, it should lay out. I found it's not necessarily the heat, but the time in which it was heated then cooled. With torches are slowly drawing temp, then slowly cooling, the metal will tend to go back to (or near) beginning state. Also, many people do not know the mapp gas facility blew up a few years ago and now those yellow bottles contain only propane with an additive and is only slightly hotter than propane alone.

If you are afraid of slicing and welding, you can just run a bead on the surface, or try your heat method again, but cool it more rapidly with a water soaked rag - or even dousing it (but be aware of hardening/tempering by doing so).

[bmwman91]

I'm just thinking....The warp is it's current lowest energy state @ Room temperature, so how would heating help flatten it, without load? So, I think we need to think were the load would be coming from.

I think you should heat the weld - or where the compression is. the resulting internal tension loads should pull it flat.

It is counter-intuitive, but you definitely heat the side opposite the weld. Basically, you heat the metal so that it expands (and yes the expansion ins going to be in the opposite way that you want), and then get it hot enough that its yield strength falls enough such that all the tension in the surrounding cooler metal "squishes" the hottest part. Then when it cools back below its critical temperature and has the same strength as the surrounding metal, it will actually be shorter in that specific area which counteracts the tension in the weld. So basically you are shrinking the metal on the side opposite the weld in the same amount that the weld shrank the part. And yes, the net effect will be a straight beam which has even more residual internal stress and is ever so slightly shorter in length overall. Applying this to the weld would make it worse, UNLESS you were mechanically loading the part to force it into a specific shape, in which case it would probably work to heat the weld.

If you slice half way through the material on that line you made of cherry red metal, then weld it back, it should lay out. I found it's not necessarily the heat, but the time in which it was heated then cooled. With torches are slowly drawing temp, then slowly cooling, the metal will tend to go back to (or near) beginning state. Also, many people do not know the mapp gas facility blew up a few years ago and now those yellow bottles contain only propane with an additive and is only slightly hotter than propane alone.

If you are afraid of slicing and welding, you can just run a bead on the surface, or try your heat method again, but cool it more rapidly with a water soaked rag - or even dousing it (but be aware of hardening/tempering by doing so).

Thanks. Based on the color rings that are all the way around the beam, I suspect that the heat is being applied too slowly and in too wide of a pattern, because it looks like the whole center section was 800F+ lol. So if all sides are heated and expanding a similar amount, this won't work (as I am discovering).

I am fine with slitting the beam in 1 or 2 places (maybe 2 since there are 2 seams on the original weld) and MIGing them. Either that or I can rent a basic O/A setup since that would give me the heat output and tighter nozzle that this seems to demand. Worst case, there are local welding companies and they could probably straighten this for me, although I am sure the cost would be high.

[mrsleeve]

Looks like you are attempting a primitive stress relief post weld heat treat. I will have to look in my book of many things and see if I can find the information to help you out, that said doing this from home is going to be tricky

[bmwman91]

Looks like you are attempting a primitive stress relief post weld heat treat. I will have to look in my book of many things and see if I can find the information to help you out, that said doing this from home is going to be tricky

Yup. DIY weekend warrior shenanigans.

It seems like I am going to need an O/A torch with a focusing nozzle to actually make this work. The MAP/PRO torches I have have too wide of a flame and they can't heat things locally fast enough to achieve the kind of shrink I need.

Or, I can try FF's suggestion to slit and weld some strips, which I am 100% certain will achieve the same effect.

[mrsleeve]

Welp, I went digging into my "book of many things" and I dont have any resources for this type of work with me this trip (all that reference material is at home in my shop in my BIG book of lots of things) and its been a very very long time since I have done any heat treat / stress reduction work. But IIRC there are other ways to skin a cat on this one, Particularly a LONG soak time at temp, and rapid cooling to get the warp back out and then another long soak at a lower temp with incremental step down cooling times over the course of a few hours back to ambient to get the temper back out of it so you dont have a brittle spot.

Or heat and whack with a hammer or like Firebird said a whizz wheel slice and then weld it back up, likely the most time and cost effective solution at this point


Alternately when you try and weld something like this again, PREHEAT the work piece to about 250-300 and keep it above 200 while working it. Granted this is mostly to keep temper close to mill specs in higher grades of steel during the welding process, it will help with the warping provided you dont overheat it locally. Also only weld short runs of a 1/4 inch or less at a time and keep moving to keep the heating even and not over heat 1 area giving cooling time if need be between passes.

[bmwman91]

Good info, thanks. If I can figure out how to pre-heat my workpieces before welding, I will definitely do so. Also, yeah, I got impatient and started running long beads rather than small sections at a time. Whoops!

[gte175x]

BTW....what on earth are you doing that you need < 50 thou tolerance over 3 feet? Haha

[bmwman91]

BFc OT had the best guess so far.

This is probably a structural spar of an Alton Brown Turkey Derrick, used to precisely lower a turkey into boiling hot peanut oil.

If it's off by more than .005" the turkey could explode.

Close, but no cigar. This is a frame that will support a hardwood bar countertop that I am installing in my house. The beam is 6ft long, originally straight to 0.010" over its length, and it is now out by 0.140" thanks to my novice welding skills. While I could just shim the countertop, or simply bolt it down, I am pretty sure that I can straighten this and have everything be close enough to perfect. It is not that big of a deflection, but wood glues are not known for their long-term resistance to creep in pre-loaded applications, so I do not want anything coming apart a few years down the road because I bolted the top onto a bent frame. Shimming would eliminate the issue, but then my OCD would be pestered every time I looked at the thing lol.

When I am done with everything, I'll post up a project log thread showing the woodworking and metalworking shenanigans behind all of this.

[mrsleeve]

really your worried about less than 5/32's of an inch out of flat over 6 feet for a wooden countertop.... WOW thats a bit past OCD when like you say you can shim it and be done with it.....

[bmwman91]

Ha. Yes, I know. But at this point it has become a personal challenge to MacGyver this sucker into straightness.

[Bavarian_Propellers]

What about using some mechanical persuasion to overcome the warp - i.e. a slight tweak in a press (assuming you have access to one or the $150 HF one). Just a thought!

[Kershaw]

really your worried about less than 5/32's of an inch out of flat over 6 feet for a wooden countertop.... WOW thats a bit past OCD when like you say you can shim it and be done with it.....

Seriously.

Unless it's cabinetry, woodworking is generally not this precise. This is exactly why trim and shims exist.

[skids]

When heat treating or tempering steel what you are attempting to do is change the molecular structure. Changing the molecular structure after metal has been worked will allow you if done correctly to have a very strong metal that is capable of holding an edge while still maintaining ductility, in other words not being brittle. Think of Samurai swords.

This only works when you have done the engineering calculations for you given alloy, all steel is an alloy of some sort a recipe of ingredients.

Heat treating or annealing or tempering will not in and of itself purposely bend steel. Quenching may produce desired effect, but again trial and much error or much calculation will be required to produce desired effect. Can you math?

A proper heat treating regimen after bending the steel will produce the desired effect I think you are after, set up a jig to bend it straight, then heat to about 1450 for 1 hour to relieve stress/memory and your good to go.