How can this (perhaps narrow) region of the HAZ not be nearly-
> or fully-hardened, and therefore more brittle? Is there some-
> thing about the fact that the excursion through the critical
> zone and back was so quick that limits the amount of hardening? >
> And how much does it really matter? If I make a chain out of
> 99 links of normalized 4130, with a tensile strength of 97kpsi
> and 25.5% elongation, and a single link which is made from full
> hardened 4130, with a tensile strength of 250kpsi (?) and 4%
> elongation (?) and subject it to a very brief 96kpsi load,
> won't I just wind up with a chain with 99 stretched links and
> one that isn't? And doesn't the single hardened link reduce
> the total elongation capacity of the chain by a negligible
Even if you believe you have the correct process, materials, and skill, test pieces and sample welds put through destructive testing are the only way to know what will fail or pass. I may have missed some of the earlier post you mention as the "4130 Debate" and I'm not sure what your project is. You did mention aircraft, and that's the area of welding I am most familiar. Back in 1972 when I was in aviation school, welding for the first time with OA on 4130, the final test was the instructor getting out a large hammer to test our work. I still use this method anytime I have new materials by making a few test runs. The biggest fear I have in these times, is someone sells me something other then what it should be. This has been a major problem in the aviation industry and the FAA brings it up every chance they can at meetings. There have been counterfeit bolts and other aircraft related hardware. How can you be sure that the ER80S-D2 filler you just bought really is? Just because it is stamped as such doesn't mean it is if someone labeled it wrong. I know this sounds a little paranoid, but when you are dealing with someone else's well being I think it's important to trust your work and materials.
My many point, use test pieces and test before committing to the real thing.
FWIW, two years ago my welding instructor gave me some certification samples he had to weld for his Air Force re-certification, and I had him weld a couple more. They were 3/4"-dia, 0.065"-wall, normalized 4130, butt-welded. He welded them with TIG, using 4130 filler rod, as required by his AF certification test.
Then I welded two more samples with O/A from the same material, using high-quality mild-steel filler.
I then beat the hell out of four samples with a big hammer on an anvil. This was as close as I could get to a scientific test for weld ductility. None of the welds cracked. I beat the tube really flat for roughly 2" in both directions from the weld. Until they went dead flat and I was really creasing the edges, no sign of cracking appeared, except for what I note below. They never did crack all the way through.
I detected no difference between the TIG-welded samples and my O/A welded samples, except for two minor things. First, my samples were slightly softer from a half-inch to an inch or so from the weld than the TIG-welded samples were. Second, the O/A weld beads never did crack, even when they were pounded flat. The TIG welds did crack a bit when I really flattened them. But those welds were narrower and had slightly more build-up than my O/A welded samples. The cracking appeared only where the build-up was thickest.
It appeared to me that the concerns about the HAZ in thin-wall 4130 are probably about what Finch et al. say they are: much ado about nothing much.
Since you've read Finch, you may be interested that I did a pretty thorough check for confirmation of his negative remarks about bronze-brazing 4130, and was unable to confirm it. I even asked the author of the old industry standard, _The Brazing Book_, who was involved with aircraft brazing of 4130 tube during WWII. The custom bicycle people, who are pretty sophisticated in their knowledge of brazing 4130, say they've done it for years and have done many tests of their brazed joints without a hint of what Finch was talking about.
Impact strength of normalized 4130 is more than twice that of plain-carbon steels at a similar ultimate tensile strength, Bob. That's why race car builders started using it in the first place. It's also why it was developed in the 1920s, for aircraft use.
Race car builders started using CHMO because they could get similar tensile strengths with thinner guages - thus lighter weight....at least that's why we used it 35 years ago.
In those days, speeds were much lower, and impact energy absorption was unheard of.
With most of today's sanctioning bodies such as NASCAR dictating sizes and guages, the fabricators have gone back to mild steel because it will bend and distort - absorbing impact energy - in situations where CHMO will fracture.
Impact strength is misleading. Yield points of similar size and guage mild steel and normalized CHMO are really quite similar. It is only after heat-treating that the CHMO yield point raises - but heat treating CHMO also increases brittleness as hardness increases...to the point that it will shatter like glass when the yield point is exceeded.
I agree with Carroll Smith when he says using normalized CHMO without post-fabrication heat-treating is both a waste of money and an unsafe practice that results in material that has the strength of mild steel, but brittle welds.
In a race car these days, you actually want something that will crush, distort, and absorb the energy of the crash while doing it.
CHMO's impact strength will transmit more of the crash energy to the driver's compartment until it simply breaks due to it being a harder material than mild steel...or due to brittle HAZ areas.
I also wonder if the HAZ is decarburized during welding due to insufficient coverage of the shielding gas. If the material is thin enough, it could be decarburized through-and-through, and be less hardenable. I suppose this hypothesis could be easily tested.
4130 offers 20-50% more yield strength, which must help resist buckling in bending and compression, and probably 50% more UTS, which has got to be an advantage in tension and shear.
Aircraft Spruce sells 4130N (normalized) tube for structural applications. My understanding that this is standard aviation practice. Normalized 4130 has a UTS of 97kpsi; annealed 4130 has a UTS of 81kpsi (source: Machinery's Handbook).
It goes back 'way more than 35 years, Bob. Dick Kraft was using it in roadsters around 1950. He was an engineer and he was building semi-triangulated frames even then, although it isn't clear whether he was using 4130 for any reason other than tensile or bending strength.
Tom Beatty's 1951 lakester had a chrome-moly frame. Some Indy cars built right after the war (WWII) also used chome-moly. I have heard, but I can't confirm, that Miller built some Indy cars on chrome-moly frames in the
Many of those were twin-tube or ladder frames and the 4130 was intended to give a better strength-weight ratio, as you suggest. But the basic idea behind 4130 is *toughness* with strength, combined with reliable welding. When I was materials editor at American Machinist I read the technical papers on 4130 that went back to the 1920s, including reports to the Army (who then had the Air Corps) on the "desirable" impact strength and the exceptional elongation (bendability) of the then-new material. The same properties that applied to aircraft frames, for which the steel was developed originally, were sought by some early race car builders.
I wasn't talking about safety here. The toughness of 4130 was specifically intended to increase the ability of a structure to withstand shock loads without breaking. It was formulated to have three desired properties: relatively high strength with good ductility; high impact strength and overall toughness; and reliable weldability, using both electric and gas welding. BTW, many aircraft in the '20s and '30s were built with 4130 frames that were *stick* welded. No kidding.
Yes, there is much better technology for safety today, and the parameters of chassis design have changed accordingly, as you're well aware.
But you were suggesting a high-strength plain-carbon steel as a substitute for 4130. Firstly, that would require more carbon than the 1018 and 1020 that have traditionally been used for space frames and some other frames, which means more problematic welding. Secondly, you would have less impact strength and general toughness even if you had comparable yield strength. You even would have less bendability before the material broke.
I think you have some misconceptions here, Bob. The lowest-carbon
1000-Series steel in the as-drawn condition that has comparable yield strength (440 Mpa) to normalized 4130 is 1030. But the cold-drawn 1030 has less than half the elongation of 4130 (12% vs. 25.5%). I don't have the impact strength for 1030 handy but it's quite low compared to 4130 (around
55J Izod), which is very high.
Again, that's part of the reason for 4130's existence. BTW, I doubt if 1030 is commonly available in drawn tube, but the ones that are, such as 1020 and
1040/45, have less yield strength (1020) and both have lower elongation. Even 1020 in the cold-rolled condition has only 15% elongation, versus 25.5% for normalized 4130. If you want to shop around and look at normalized 1040, to make another comparison, yield strength falls to 370 MPa and elongation only goes up to 28%. And then you have to weld a 0.40%-carbon steel, which is just over the threshhold of "easy weldability." 4130, of course, has only
The race-car field is notorious for myths and misconceptions about materials properties. I used to race as an amateur and I remember -- it was like alchemy. The engineers at the top end of racing, in F1, CART and so on, are first-class and they know these things very well. But most of the field is full of old-wive's tales and misunderstandings.
4130 won't "shatter" at strength levels comparable to those of plain-carbon steels. In fact, just the opposite. 4130 is much more bendable and much tougher, in terms of impact strength, than carbon steels, even while it exhibits higher strength. Again, those are keys reasons they created the alloy in the first place. It will crumple up just like 1020 -- in fact, it will do a lot more crumpling, due to its higher elongation, than 1020. It just occurs at much higher levels of force and impact.
And that indeed can be a safety issue. I'm not questioning the practicalities of designing crush zones into a race car chassis. Only the properties of the two materials, as you've described them. It can't happen. It doesn't happen. And I was once pulled out of a thoroughly crumpled Schweitzer 233 sailplane, made of 4130, that made me very thankful it didn't happen.
(BTW, you mentioned using thinner-walled 4130 to decrease weight in a race-car chassis. I'm sure you're aware that the torsional stiffness of a chassis is directly related to wall thickness, all else being equal, and that, since at least 1960, torsional stiffness has been the primary design criterion for race car chassis. That's why the Brits used mild steel in their lightweight, space-frame race cars; it gave the exact same performance, disregarding safety, as the more-expensive 4130. But they had a national misconception about the difficulty of welding it, if you read Costin & Phipps classic book on chassis design, published about that time. They had their own myths and old-wives' tales.)
4130 was developed for aircraft use in the 1920s. It was supplied in both the annealed and the normalized condition. For military use, it was fully heat-treated in large ovens, fully assembled, by the early 1930s. Those frames often were welded from annealed tube.
For civilian use, it is almost never heat treated. It's used the normalized condition for almost all tubing applications. For components made of sheet, plate, and bar, it's sometimes heat treated.
No disrespect meant here, Ed, but I'll stick with the late Carroll Smith's take on the matter.
While your credentials are certainly impressive, Smith's engineering background is based entirely in the motorsports arena with its own set of parameters and peculiarities, and it includes the design of many successful race and consumer-oriented, high-performance cars - including Ford's LeMans-winning GT-40 and the Shelby Cobra - in concert with another male named Carroll, Carroll Shelby.
CHMO - annealed, normalized, or hardened - really doesn't have a place in the work that I do, and, from Smith's standpoint - with which I agree - CHMO that has not been heat-treated post-fabrication is, essentially, no stronger than mild steel in racing applications, and has the added disadvantage of having brittle welds/HAZ.
I have seen countless CHMO tubes on Supermodifieds, Sprints, and Midgets broken within a half-inch of the weld - which supports Smith's postulations in my own mind.
My copies of his " ** to Win" series of books are dogeared from use, and he hasn't failed me yet.
Bob Paulin - R.A.C.E. Race Car Chassis Analysis & Dial-in Services
And with no disrespect to Carroll Smith, I did materials for a living, and the figures I reported and the testing done by ASM etc. are undeniable. You can find them anywhere, and, if you do some digging, you'll find the results of structure tests that confirm the lab tests.
Rather than accept Smith's conclusions second-hand, you'd do well to run some tests yourself, even simple ones like those I did two years ago. Among the people who live with 4130 tube structures day in and day out, particularly the welding instructors at the EAA and Lincoln Electric, you won't find much support for Smith's ideas about 4130. And the FAA failure reports say the same thing.
If you want to duplicate your near-weld failures, try MIG welding it. Failures have shown up there. If you want to see some failures in a plain-carbon steel with strength comparable to that of normalized 4130, do some welds and tests with 1040.
BTW, I sent a short list of errors I found in Smith's "Design to Win" to the publisher in the mid-'80s. I don't know if they were incorporated in the revised editions or not.
Excuse me, but if I understand your work history correctly, you didn't do materials for a living, you wrote magazine articles about what others did with materials for a living. Not saying you're necessarily wrong on this issue, but Carroll Smith probably had lots more first hand experience than you. Bob probably does too.
NHRA rules ban 4130 for safety systems, that's experience based too. Post crash inspections turned up too many brittle breaks next to weld joints. Perhaps that was due to improper welding techniques rather than poor materials choice, but it wasn't getting caught in pre-race inspections, and eliminating use of the material eliminated the problem.
If you understand correctly, Gary, it includes a lot more than magazine articles. It includes work at ASM, materials research for MITI (including the machining of test models) and for US companies (aerospace models, mostly), and interviews with most of the people who actually study and test
4130, among other materials.
What that kind of research does is pick out the flyers -- the people who know a lot about something but who have a blind spot or two. For example, Richard Finch, who certainly knows welding and brazing, has some ideas about brazing 4130 that no other expert seems to share. And I've checked with the top experts in the field. As you know, I sometimes spend weeks of full days tracking these people down. It's a lot more than reading what Carroll Smith says in a book.
The pros who train for the EAA have probably welded more 4130 than Carroll Smith has ever seen. Some of them do it all day long, every day, and have examined hundreds of crashes. Every one I've talked to disagrees with what he says about the effects of the HAZ in welded 4130 of 0.065" wall thickness or less. The old-timers still do post-weld "stress relief," but the people who have actually tested it say that it has no effect on thin sections. And that includes people who have put the joints on Instron machines and Izod test machines to measure it.
Smith is an icon but, like many experts, he has some very strong ideas and they don't agree with those of other people who have equal or greater experience with, for example, 4130 steel. And some of what he says is flat wrong. For example, if he truly says that 4130 that's been welded has no greater yield strength or tensile strength than mild steel, as Bob says, then he's just wrong.
Fully annealed under controlled conditions, 4130 has a yield of 360 MPa.
1020, fully annealed in the same conditions, runs around 295. Tensile strength runs 560 MPa for annealed 4130 and 395 for 1020. (Note the greater spread between yeild and tensile in 4130; that's a sign that 1020 actually is more prone to brittle failure, which is proven in other tests). Welding isn't "controlled conditions," and the 4130, like other slow-quenching steels, will run somewhat higher than those values while the 1020 generally will not.
I wouldn't question it. Practical safety issues are something else. Bad welds, design for crush and so on are not the same thing as impact strength and ductility. But don't try to tell us that 4130 "shatters," or that it has no more tensile strength than 1020 after welding. There is a mountain of carefully measured results that prove both suggestions are wrong -- unless your welds are no good.
Are you making a case against building a chassis from 4130 without post-weld heat-treatment, or against building a chassis from 4130, period? What do you think of using a O/A torch to temper the HAZ? Carroll Smith insists on it; Richard Finch says it's crazy.
Perhaps Smith and Finch disagree because racing cars and airplanes have different design requirements. Everybody knows that to handle, a car chassis must be stiff, especially in torsion. And if the chassis is sufficiently stiff, it will never approach the yield strength of even the weakest carbon steels except in a crash.
Strength, then, is for crashes, and in crashes, impact strength, ductility and fracture toughness are as important as ultimate strength. Is it possible that the aviation crowd (Finch) regard airplane crashes as largely unsurvivable, so they design for strength alone -- to, say, keep the wings on during a dive? I don't know. How close to yield do airplanes get in normal operation? How stiff do airplanes need to be in order to "handle"?
O/A has a bigger HAZ, of course. Maybe decarburization was also a factor.
Perhaps an argument for pre-heating 4130 joints prior to TIG welding.
Maybe. I wonder if beating the hell out a 4130 weldment with a hammer is not a good simulation of what occurs in a major impact. Are fifty n-joule hammer blows over the course of five minutes really the equiv- alent of a single 50n-joule blow that lasts one-thousanth of a second?
I guess I should take Eric D's advice and do some real tests.
I don't believe Smith said exactly that, I don't think Bob did either, but he can speak for himself.
I do know that NHRA specifies a greater wall thickness for mild steel tubing (with a hole drilled into the tube so the inspector can confirm it). That provides the necessary strength with none of the complications.
OK so to sum up here. (feel free to correct me, this discussion has gotten really long, and I lost track a while ago)
Bob is saying that there is no pint in using 4130 if you don't do a full post-weld heat treat of the frame. And you might as well use mild steel tube to build your frames as the welding is simpler and the actual strength is almost identical to normalized 4130.
Ed is saying that a 4130 is stronger than a mild steel frame of the same weight, and that post weld heat treating isn't really necessary iof the welding is done correctly.
Gary pointed out that the car frame builder use slightly heavier wall thickness mild steel tube to get te same strength as 4130 with only a little more weight.
I teach welding, and I have a lot of guys come in to learn to weld bike, car, motorcycle and air frames. Many of these guys assume that they have to use 4130 because that is what everybody else uses, especially the air frame guys.
The bike frame guys are actually more interested in titanium and aluminum than 4130.
The Motorcycle frame and car frame guys seem to have accepted mild steel tube as OK.
The air frame guys are still locked into 4130 as the only materiel they can use.
I have gone along with the school of thought that if you use ER80S-D2 filler when TIG welding 4130, that no post weld heat treat is necessary.
I have yet to be proven wrong.
I don't feel that gas welding 4130 is a practical solution for modern fabrication. Yes it works, and in theory you can stick weld aluminum too. I still wouldn't recommend it to anybody.
I think Bob has a valid point about the whole assumption about having to use 4130 in the first place, but that is not my decision.
People come to me and ask, "How do you weld this". I teach them.
Carroll Smith's advice is kind of like keeping Kosher. they started out as guidelines for doing things safely, became rules of life, and have transformed into religious dogma. Sure, a long long time ago eating pork was taking your life in you hands, but not today. Similarly, if you follow Carrolls advice religously you probably won't make stupid mistakes and hurt yourself - but he isn't always right, today... His books were written a long time ago, by a racer, giving other racers advice that wouldn't get them hurt.
First, a "mea culpa" on my part. I should have looked up the tensile properties of 4130 before spouting off. Everyone else was right - normalized 4130 typical properties are 75 ksi yield strength and 95 ksi tensile strength. This is a significant increase over most carbon steels. However, annealed 4130 (which is also available) may be significantly lower than this.
I don't want to get into all the detailed arguments going back and forth but(whoops, I am still getting involved)...
impact toughness: I would doubt, based on general metalurgical principles, that normalized 4130 had significantly better impact toughness (as measured by Charpy impact absorbed energy or, more properly, KIC or COD fracture toughness). 4130 can have excellent fracture toughness when in the quenched and tempered condition. Conversely, 4130 in the fully quenched condition had extremely poor fracture toughness. Normalized
4130 should have only fair fracture toughness (note: fracture toughness of normalized carbon and low alloy steels is usually directly but inversely correlated to the carbon content - as the carbon content goes up the toughness goes down. That is one way, among many, that modern steels have improved their fracture toughness, i.e. by decreasing the carbon content).
HAZ decarburization: It is unlikely that any significant decarburization is taking place either due to shielding gas or other metallurgical effects. The carbon content in the weld metal is lower than the base metal. Although this would indicate some possibility of decarburization of the HAZ, the HAZ (i.e. base metal) has the higher chrome content which has a high affinity for carbon. This inhibits any carbon migration from the HAZ to the weld metal. In fact, for HAZ decarburization to happen during welding, the weld metal generally has to have a lower carbon content and a higher alloy (chrome) content compared to the HAZ. Even then, the effect is seen intermittently and for a VERY short distance (one grain size). However, note that this effect can become quite evident if a post weld heat treatment is used. However, we are debating about 4130 welds without PWHT.
Final comment I have NOT been involved with thin-wall 4130 tubing structures, so I am not qualified to speak. However, it is interesting that most auto racing rules strongly recommend carbon steel tubing rather than 4130 (as mentioned by Bob Paulin). The desire for lighter weight tubes for "normal" static loads would obviously lead one to the use of thinner wall 4130. However, dynamic loading is a different case. Again, I cannot speak with any authority but I can think of an interesting analogy, for those of you that are old enough. We all are familiar with beer and soda cans made from very thin aluminum. We can crush them easily, even in the longitudiunal direction (by stomping on them or, if you are in the movies, by crushing them against your forehead). Anyone remember doing this with the thicker wall steel cans many years ago??? Section modulus (diameter and thickness) are important for this. Anyone remember "Euler buckling" of thin columns in statics class? You don't change the material or strength to prevent this, you change the section dimensions. (I think this was how I got started on this in the first place)...
I guess I will have to get a couple of short lengths of 4130 to test.
I now surrender on this topic. :-) However, my thanks to those of you that are keeping me honest on the facts and with supplying additional information. Alternative theories are a good incentive to check my own facts and knowledge and to think about things in order to provide improved understanding.
Perfectly Satisfactory airplanes have been made of wood.
I think in general the best structures result when their design has been made taking good account of the material's properties, rather than throwing exotic materials in to survive in the problem spots.
Welded tubing space frames need to be triangulated in 3 dimensions to be stiff. Then the joints could be simply "pinned" and 'most all our fears about weld material strength, bead shape, and HAZs disappear, as buckling at the tube's un-tainted uniform mid-span becomes the main faiure mode.
If thru ignorance, some unfortunate engineering necessity, or to intentionally add some degree of deformability a tube frame uses curved members or open sections then the tubing and the welded joints will be asked to handle bending loads. The metallurgy in and around the joint, while interesting, can be fatally undermined by the mechanical details of bead size and shape, undercutting, gussets, and the overall joint design.
The 4 link front section of a basic diamond bicycle frame is a dreadful strutural shape that seemed to need near fanatic attention to details like brazed lugged joints and butted tubing to survive. My suspicion is the move toward big bike frame tubes has brought bigger joints that automatically are capable of handling larger bending loads without stress concentrations, etc.