I tried, very briefly, to do a destructive test on a weld ... but I'm not quite sure what I am looking for.
I made an outside corner weld on some scraps of 1/8" thick mild steel bar. I set the pieces up at a 90 degree angle, so that the inside corners just touched, allowing me to achieve full penetration (maybe?) as I welded the outside corner. I welded a stitch-motion pass using
6011. The weld looked pretty good, as much as my limited experience could tell.
Then I took a hammer and beat the angle down flat. I was disappointed that all of the bending occurred in the weld metal. I was even more disappointed that by the time I beat it flat, the weld metal was starting to crack. When I separated the pieces and examined them, it seemed that the weld had indeed achieved good penetration. Though I did not use a magnifying glass or some such to inspect it, I did not see any signs of porosity or inclusions. There was no sign of cold lapping (no weldment separating from the original metal). But--the weldment itself was far less flexible than the original metal, and broke far sooner than I had expected.
Is this what *should* have happened? Is the weldment always more brittle than the original metal? Or is this a further indication of how much I have yet to learn? If this is not what should have happened, what should I look for, and what might I have done wrong?
I look forward to any help this group has to offer ...
On outside corners and lap joints, there is the tendancy to have a reduced throat problem.
__ /\ / \ / \ / /\ \ instead of / /\ \
For a good sketch of this in a lap joint, see:
formatting link
can then see how this would affect a corner.
Also, if you had 0 root (plates touching), you might not have had a "full" penetration. This is why pipe weldors use a small root gap. A piece of mig wire is useful for spacing the root in little coupons like this.
That is probably going to happen no matter what you do....The weld is not going to have a whole lot of strength in that direction...If you would have bent it the other way (like closing a book) it should stay together..Is this how you bent it or did you open the "book" ALL the way? That is how I am interperting that you did it. I agree that it would be stronger if you had a small root gap.
It is pretty sure that your weld is adequate. For 1/8th material I am trying to imagine a structure that would be expected to withstand such abuse. Generally people will demand very high quality welds and not consider the application. The reference that Rich gave you is one of the best I have read and is worth reading. " Nothin Welds " is aimed generally at structural welding. You might start looking at welds on furniture. That old kitchen chair that you lean back on two legs seems to hold for a long time before it fails. The welds are done on the cheap but they work. Look at the welds on a automobile frame. You will see all kinds of weld faults but they hold for the purpose they were designed for. I remember an engineer acquaintance boasting to me that all the welds on his project were one hundred percent X-rayed. He had the impression that because all the weld joints had been examined that they would all be perfect and have no flaws. No use arguing. His false impression might have been correct if he was building a nuclear power plant. You are concerned about your weld being "brittle" . From what you describe the weld did stretch and deform before failure. It might have been a bit less ductile than the parent metal but then likely it had a higher ultimate tensile strength also. Check your fracture line for porosity or slag inclusions. If the break is clean then you did a good job. If you really are concerned about strength on this particular joint then just add some stitches of weld along the inside of the corner. If you put a continuous weld on the inside you will create a large amount of distortion. It seems to me that you are doing just fine. Randy
"Andy Wakefield" wrote in message news: snipped-for-privacy@posting.google.com...
If you have access to a power saw (an even if you do it with a hacksaw!), cut a cross section thorugh the weld area. I'll bet that you have a nice cold section right at the root, should be easily visible with small magnifing glass. The resulting thin spot is going to be the start of your fracture.
though i'm not as experienced as the other posters, i'd like to add something: your weld bead/pool/puddle will be a mixture of weldrod chemistry + base metal chemistry. before making decisions on the 'success' of a weld, you'll have to keep in mind the limitations of the base metal.
for example, take a piece of angle or flat from a bed frame ... fold it over onto itself and beat it flat with a hammer.. much the way you tested your weld. more often than not the fold will crack on the outside. it is a limitation of your base metal.
Randy made a good point about considering the applications and requirement of the joint.
i *think* a better judgement for your test would be made on the initial plastic deformation. that is, after you flattened the Vjoint, where the two original pieces still 'flat'? and the only noticable failure was in the weld area? usually a good weld (joint) will start to bend/fail in the heataffectedzone on both sides of your bead.
how much of a 'failure' the bending implies is a science unto itself.
Many thanks for the responses. I feel better about my welding after reading your posts! Not having done any destructive testing before, I hadn't known what to expect. What I did not expect was the relative brittleness of the weldment -- I thought I would be bending the weld back and forth repeatedly before it broke, much as one would have to do to get the original metal to break.
Well, that's also got to do with the configuration of your joint.
When you bend a piece of flat bar back and forth, there are no stress risers
- corners, cracks, sharp change in direction. These are where cracks and failure should start.
A outside corner joint has one big stress riser right down the center.
Try a butt weld, like a vee-groove, and you will probably be able to beat it back and forth a little before it breaks. Also, if done right, the break should be in the base metal.
You might be interested in the common methods for qualifying a welder to perform critical welds (pressure vessel, structural, etc)
I have to run destructive bends on a regular basis, and the process you are looking at-a bend across the weld- (as specified in ASME for pressure vessel workh) for butt welds requires bending the weld over a mandrel of circular section to form a bend of 180 degrees, the mandrel being sized to give a 20% strain on the weld metal at the outside of the weld (for low carbon steel). This means that for 1/8 material, the diameter of the inside of the bend is 1/2" (I think these are the right numbers... I don't need to make a new rig very often)
bending a corner weld like this produces MUCH greater than a 20% strain by the time the material is flattened at the weld- the inside of the corner effectively starts at length zero, so as the weld area is flatened, the base metal is what stretches. The metal at the inside of the bend won't compress much until you get to high strength material as the compressive strength of mild steels is generally about
95000PSI.
Permissable flaws in the coupon, under ASME, will be visible (I don't nkow what it would be for 1/8 material-for 1/4 and up I think it is an indication less than 1/16 in the weld area) and some number of them are permitted in a given area.
The interesting thing is that the standard is the same for root bend (root on the outside to be stretched) face bend (crown of the weld out) and side bends (where the material is cut across the weld to a width equal to thickness, giving a square cross section, and bent to stretch a cross section of the weld, showing internal flaws) and is substantially the same for all metals and thicknesses, the main changes being the radius of the mandrel, the allowable size of an indication, and the number of indications.
Welds are generally flushed with a grinder before bending to remove stress risers from reinforcement at the cap and root (hard grooves where the cap meets the surface and where the root pushes thru) and to bring the weld thickness down to the thickness of the base metal (remember the 20% strain? If the weld is left thicker, the strain at the outside of the bend will be higher.) The main purposes of hte test are to demonstrate complete fusion, weld metal ductility (usually greater than base metal) and quality of weld metal (no inclusions, no voids, etc) and any tears in the base metal, even the HAZ, usually don't matter for steel (though they may for other metals)
The methods for testing fillet welds are a little different (we don't qualify these-- the groove weld qual covers all of our work) but have the same purpose-stretch the weld metal to open up flaws and show ductility/quality of the metal. The weld you are doing is basically a groove weld, not a fillet, as you are welding on the edges of the material.
A small gap at the root helps insure complete penetration, relieves the sharp inside corner (a stress riser-- even a slight bend out will elongate the metal uch more than the typical 20 to 40% strain that most filer metals will take.), and lets you SEE the root has complete penetration. For thicker material, don't be afraid of multiple passe- a BIG advantage is that each pass anneals the previous and helps refine the grain strucure. This is sometimes specified in a procedure as an annealing pass.
With a stick like 7024, this isn't a big issue (the slag puffs up and insulates the weld during cooling so it will cool slowly) but sticks that leave a thin slag, like 6010, and slagless processes like MIG, the cooling rate can be high enough to harden the surface of the weld. An unprotected root may also cool fast enough to harden. This is one of hte reasons that (though certainly not the only one), in pressure work, 6010 (with a thin, non-insulaing slag)generally needs to be capped with 7018 (with a thicker, more insulating slag). Ran Brinell hardnesses on some welds recently where this was a major issue in a repair- the 6010 root was way over 200 (off scale for the test), the unground 7018 caps were in the 170 range, and the ground 7018 was about 150. Greater hardness will be more brittle. The ductility and toughness go down as hardness goes up. Don't be afraid of your grinder in finishing the weld. Remove those stress risers and the top few thou where the hardening is worst.
OK-- can you tell that I'v been doing a lot of reading for inspection certifications lately? (hopefully will have NDE level 2 and CWI before next fall)
A very helpful response -- many thanks for clarifying what destructive testing is all about. All your reading has already paid off -- at least for me :)
How does one insure a weld that will pass the sort of tests that you describe? Obviously, avoiding porosity and slag inclusions ... you talked about grinding down flush and avoiding any stress risers ... what else would be keys to success? For example, is it crucial to have exactly the optimal amperage dialed in? Or as long as one has achieved good penetration, no inclusions, etc., could one be running a little too hot, but still be okay?
I am not entirely sure I understood what you meant by allowable indications -- is this some sort of slight defect in the weld? Perhaps a variation in thickness?
One last question (just so I can *really* display my ignorance!) concerns the hardening -- I was thinking that mild steel (and the corresponding mild steel electrodes) had too little carbon in them to really respond to heat treatment, and therefore wouldn't harden appreciably in the welding process. I take it I was wrong about this?
Many thanks again for the helpful responses from everyone on this ng
-- I don't know whether I could say that reading this ng is better than taking classes, because of course one doesn't get the hands-on training ... but I would definitely say that reading this ng has done just as much or more to help me improve as a weldor than just about any class I could imagine!
Good grief man, if I had a nickel for every high pressure weld that's been made, and continues to be made, with rod other than 7018, I would be flying my welders around in private jets. If I had a nickel for just the ones I've made myself, I could drive them around in Hummer stretch limo's. It's dangerous to make blanket statements like that. It may be true in your experience, but keep in mind it's a big old welding world out there.
JTMcC, happily making high pressure pipe and pressure vessel welds with 5P,
Checked the book at work: Details for those interested (relevant to ASME controlled work on pressure vessels and piping; other applications are similar)
Bend tests: See 2001 ASME section IX, QW466.1, test jig dimentions
for mild steel (P1 material) the radius of bend is specified as
2t, where t is the material thickness at the bend (for 3/8 or less material) so the inside DIAMETER of the bend for 1/8 material should be 1/2". This is 20% strain at the outside of the bend.
QW452.1 gives the requirements for transverse (across teh weld) bends for various thicknesses and positions
QW163 gives he acceptance criteria for a bend: There shal be no OPEN (emphasis added) discontinuities in the weld or HAZ exceeding
1/8" measured in any direction on the convex surface. Openings at the corners don't count unless they show lack of fusion, slag, etc.
The guided bend is considered the most reliable test, as it shows what the weld metal actually performs like after the werder lays it in. X-ray (for welder qualification) is also used, but several of the inspectors I work with don't like it as well, as it shows only porosity, inclusions, etc, but not the strength of the metal. I believe the bend test is a tougher test to pass, but it will miss anything not in the coupon (x-ray is usually an entire weld) and may miss things not near the surface.
If you are interested, I highly recommend getting to a library and looking in Section IX. (Don't even think about buying it unless ytou find an outdated copy or you are wealthy)
I don't really know about ASME, but AWS calls them 'discontunities'. pinholes, cracks on the edges, that sort of thing. You are allowed a certain amount and size of discontunities. If they are big enough, show up in tight groups, or contain any slag, then it is a 'defect'.
Hardness and embrittlement can be caused by things other than carbon, hydrogen namely. This is why lo-hi rods are used, and one of the reasons preheat is used. Hydrogen can also cause cracking.
PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.