Welding Stainless Steel to Carbon Steel

Date 04-30-2007 22:15

Hello Richard V. Roch, maybe a little bit more information might help with some of the responses. You say the bends failed, where exactly were these failures, in the weld metal, parent metal, at the boundary between the weld and parent metal? Was there any visible lack of fusion or entrapment of slag, or did it just appear to have torn?
Please forgive me for this next question, but I have read some past posts where a discussion took place in regards to the correct thickness and size of the weld coupons prior to testing and also the correct bending radius mandrel for a given coupon type and thickness. I bring this up because there is a definite relationship between coupon thickness and radius size for properly stressing a given weld coupon. As I recall there have been instances where the coupons failed due to being bent around too small of a mandrel or the coupon being too thick for a particular mandrel radius. Another reason that I brought this train of thought about has to do with the fact that you said that your impacts and tensiles were acceptable, I believe all of these items go somewhat hand-in-hand, particularly the tensiles and bend specimens. It doesn't make sense to me that you can meet a tensile requirement and then have an issue with the bend test unless there is possibly something going on with the manner in which the bends are being conducted. Just my $.02. Regards, aevald

By ssbn727 (*****)

Date 04-30-2007 22:53

I have to agree with aevald here!!!

The only other questions I would like to ask Richard is with respect to the bends, let's assume everything was correct as far as the bending of the coupons is concerned... In other words the correct mandrel & coupon dimensions...

1) What type of groove joint, opening, angle etc. was used?
2) Were the failure in the root or face bend(s)?
3) Could you better describe the "strongback" you welded on the root side of the weld and how it's related to the joint itself?
4) Was the spool of FCAW electrode wire properly stored?
5) Did you have the correct mix of shielding gas and was the recommended flow rate used?
6) Would you please describe the specific type of failures?
7) What were the welding power source output parameters set to and the diameter electrode wire being used? Electrode stickout? Position of the test joint?

In other words, even though this is alot to ask but, the more info that's given to us, the greater the possibility that we can give the proper guidance for you so that you can ultimately resolve your problem.

So please do'nt think that I'm pestering you! I'm just echoing Allan's reply in my own words.
I do believe Allan is on to something with him leaning towards the possibility of not performing the bend tests properly however, I'm just attempting to cover all the bases and if I missed something, then someone else will chime in and add another potential route to finding the solution to your problem.

Respectfully,
Henry

By chuck meadows (***)

Date 05-01-2007 03:17

Allan and Henry,
Couldn't the use of FCAW, and the adherent loss of toughness due to the oxidation associated with any flux incorporated welding process, and the use of a "strongback", which can actually be considered welding in a high restraint, possible be part of the problem? I believe this scenario would require a PWHT to relieve the stresses of the highly restrained joint and restore the weld and HAZ to it's maximum strength, especially at this thickness. Like you, Allan, that's my $0.02 worth..

Chuck

By ssbn727 (*****)

Date 05-01-2007 06:00 Edited 05-01-2007 06:09

I completely agree with your points Chuck, and that's exactly why I asked the question about how the strongback was placed in relation to the joint in the hopes that Richard would also give us more details.
the other questions were also included because I wanted to try my best in "covering all of the bases" so to speak... the only thing that I might've forgot to ask was whether or not the E309LT-1 was the only filler metal used in welding the SA240-316L to the SA516 -70 which has alot more carbon than the latter so, in retrospect, I also should've asked whether or not PWHT was performed afterwards prior to performing the destructive tests

For me personally, knowing a little bit how stainless steel tends to behave in such joints that are highly restrained, I was'nt surprised that the results were failures in the bends as opposed to being very surprised with the results as was Allan with respect to the tensiles or the impacts...

In this case we're talking about joining one type of alloy with another that have totally different coefficients of thermal expansion between the three joined alloys combined with the use of a strongback for which as Chuck eloquently pointed out would cause in effect - a significant amount of residual stresses so naturally speaking, there would be too much restraint applied to the joint, therefore causing the failures at various locations on the bend specimens.

Now my question for Chuck would be if the intent was to avoid PWHT as much as possible, then could Richard get away with avoiding PWHT by only using the E309LT-1 as a buttering layer, and then using an E316L FCAW wire in order to minimize dilution combined with tacking or holding in an unrestrained manner the joint together so that there would be enough play or freedom for movement so that the joint could in effect warp or distort in a manner so as to prevent the entrapment of excess residual stresses and therefore minimizing any excess amount of restraint within the joint or would PWHT still be necessary in order to restore all of the mechanical properties that were lessened due to welding - even with all these factors taken into account?

I know that this may be a stretch and maybe I'm asking an awful lot from you Chuck but, am I headed in the right direction here? Also, what's your opinion as to why the tensiles and impacts were acceptable if the all of the bends failed? Is it because both tests actually relieve the specimens of residual stresses that are entrapped simply by the nature of the physical & mechanical aspects of each of the 2 tests?
I look foward to your response as will others since I'm confident that you'll lead us in the right direction towards resolving this issue ;)

Respectfully,
Henry

By Richard V. Roch (**)

Date 05-01-2007 11:39

Gentlemen
Thanks for the replies and here is the information you requested. I should have included in the first posting.
1. The failure occurred in different areas of the weld, including weld centerline at cap and fill area, and fusion line at C.S. Also one of the specimens had total failure it broke completely in two.
2. A single vee type groove was used with 35 deg bevel angles. The mandrel and everything associated with the bending was the proper radius in accordance with ASME Section IX.
3. The strong backs were two 1" thick C.S. bars welded to the back of the coupon transverse to the weld direction. They work good because there was no distortion at all.
4. The coupon was stress releived at 1150 Deg. F for 4 hours.
5. The spool of wire was properly stored in a dry room where we keep all the wire.
6. The shielding gas was a tri mix of 90% He, 7.5% Ar and 2.5% CO2 with the flow at 40 CFH.
7. As far as welding power source output parameters, our machines are calibrated and the amperage that was recorded was from 149 to 182, the voltage was a steady 25.5 to 27.5 and the travel speed was from 8 to 12 I.P.M.

Gents, hope I answered all your questions. If there is any other information I may have left out please let me know. We kept a good record of all the parameters during the welding part of the test.

One question that I would like to ask Mr. Meadows, in an earlier query I asked about using tri mix when welding C.S. to S.S. and you did not recommeded this gas. Why?

Gentlemen thanks and hope to hear from everyone.

By Milton Gravitt (***)

Date 05-01-2007 12:45

Richard the only thing that you might try is to change from the tri-mix gas to a 75/25 argon and C02 mixture. I've aways used tri-mix gas on solid welding wire and 75/25 on fluxcore wire,when a tri-mix gas sets for a while sometimes the gases will separate and you need to roll the cylinder on the floor to mix it up. Just my .02 cents worth maybe it will help you.

By Richard V. Roch (**)

Date 05-01-2007 14:18

Thanks Milton, all responses are greatly appreciated.

By 803056

Date 05-01-2007 14:59

I would suggest using a wrap-around bending fixture.

When bending dissimilar base metal combinations such as austenitic stainless steel and high strength low alloy steels together, one component (the HSLA) will be "stiffer" than the other (ASS). Using a standard plunger and die type bending machine will typically result in the majority of the bend being located in the more ductile material. The sample may even shift to one side in the bending machine. In the absence of weld defects, the bend may not be uniform throughout and some localized areas where the bending stresses are greatest will be stretched beyond their limits of elongation.

Examine the fractured surfaces carefully. Verify there are no weld defects present. If you would like, send the prepared bend samples to me and I'll bend them in my machine. They may still break, but I've had very good luck with metals and aluminum alloys that are difficult to bend using the plunger and die machines.

Best regards - Al

By Richard V. Roch (**)

Date 05-01-2007 15:05

The laboratory did use a wrap-around bending fixture.
Thanks

By aevald

Date 05-01-2007 14:50

Hello Milton Gravitt, I believe that your observation on the shielding gas that was being used might be an issue here. I often run into individuals and companies that do not pay as close attention as they should to the type of wire that they are using and the shielding gas they are using on it. It was explained to me that a specific shielding gas in combination with a specific wire being used under a certain range of parameters will yield results that the wire is designed for. If an individual makes the choice to alter this combination of parameters, or is unaware of their significance, you can run into trouble. It was explained to me, I hope correctly, that the fluxes and stabilizers in the wire are designed to dilute, burn out, provide shielding, alloy, and other duties based on the welding parameters and shielding gas used. Change the shielding gas and you can significantly effect the duties that the fluxes perform and thus end up with problems in the finished weld metallurgy. Regards, aevald

By Milton Gravitt (***)

Date 05-01-2007 18:59

Hello Aevald you might be right on the gas but where I have work they wanted to use the 75/25 on the SS and CS fluxcore wire and when mig welding with solid SS wire we would use the tri-mix to bring the color out on the SS.

By aevald

Date 05-01-2007 22:14

Hello again Milton Gravitt, I would totally agree with the gas choices that you have included in your last post to me. I would however check to see that the gas selections match those that the wire manufacturer specifies for their particular wire. GMAW SS or CS will allow you a much wider leeway when making a gas selection. FCAW SS and CS steel will generally specify a given preference for gases, some of the CS FCAW wires should only be run with CO2, others only with 75/25 Argon CO2, and still others will call out for either gas, now there are wires in the CS FCAW class that specify 90/10 Argon CO2. Chuck Meadows made mention in his post that SS FCAW wire will accept either 75/25 Argon CO2 or straight CO2, he also mentioned that AWS has a statement that allows other gases as long as they don't adversely affect the completed weld metallurgy.
There is a manufacturer in my area who uses a special tri-mix on their CS FCAW, apparently they felt that they wanted a better shielding gas for their GMAW applications and didn't want to have to carry two different types of gases on their welding carts when they switched from the FCAW to the GMAW wires. I can only assume that they haven't had any particular problems from doing this since they have used this mix for over two years now. The problem that I forsee from something like this would have to do with any liability issues that might ever arise from weld failures using an unapproved gas. Some food for thought. Regards, aevald

By Richard V. Roch (**)

Date 05-01-2007 15:55

Chuck on a previous query you recommended not using the Tri mix gas with the FCAW and 309L filler. Can you follow up on that?
Thanks

By chuck meadows (***)

Date 05-01-2007 20:09

Richard,
Tri-mix isn't recommended for FCAW because there should be at least approximately 20% CO2 for maximum properties derived from the flux. Now, tri-mix is used for GMAW, but my personal preference is 98-2 Argon/oxygen.
Back to FCAW..The only 2 gases used to test the wire at the factory is 75/25 argon/CO2, and straight CO2. But, ANSI/AWS also says that any gas mixture where optimum properties can be achieved is OK.

By Richard V. Roch (**)

Date 05-01-2007 21:18

Thank you Mr. Meadows and to all who responded I really appreciate it.
Thanks

By GRoberts (***)

Date 05-02-2007 01:33

Richard,
I assume that this is the weld that you refer to in this previous post?

http://www.aws.org/cgi-bin/mwf/topic_show.pl?tid=11304

Have you been able to check the weld for intermetallic formation? What ferrite level was your 309L filler metal.

I definately agree with Chuck. Either CO2 or 75/25 gasses are more than acceptable to pass bend tests with with FCAW sst, you shouldn't have a problem because of that.

I have passes hundreds of bend tests with strongbacks. If the high restraint doesn't cause cracking, the bend tests won't be affected.

What sort of heat input did you use. Too high can cause large grains to form, and have a negative affect on bend tests.

By Shane Feder (****)

Date 05-02-2007 02:07

Richard / Greg,
I am just wondering why you would use strongbacks on tests.
I was always taught that a weld should always be able to expand and contract without restraint in order to obtain optimum mechanical properties.
If the joint distorts that much through welding would it not be because the joint design is incorrect and a more appropriate design is required. Or it could be the run sequence that is causing the distortion
If you get a large amount of distortion on the test plate there is a strong possibility that the production joint will distort as well.
Are you going to weld strongbacks on the production welds ? That has always been one of my pet hates, temporary attachment welds on stainless steels.
Your thoughts would be appreciated,
Regards,
Shane

By GRoberts (***)

Date 05-03-2007 00:22

I use strongbacks on tests in order to limit angular distortion. On joints welded from one side for test plates, the angular distortion can get out of hand so as to produce a test plate that is very difficult to use &/or not usable. Some codes even limit angular distortion to 5 degrees max for tests, or they are rejectable. I do not think that restraint affects the properties of the joint, but mostly the propensity to crack, which can most of the time be offset by higher preheat when necessary (for low alloy steels). I believe that is actually beneficial to apply strogbacks in testing so that the joint is more representative of what happens in production. That way you know if the restraint will cause a problem when you get beyond the testing stage. Most production joints are much higher restraint than test plates just because the members are so much bigger even if strongbacks are not used in production.

By chuck meadows (***)

Date 05-02-2007 04:03

Greg,
You know, it could possibly be due to an intermetallic phase, especially when Richard said the failures were not always in the same vicinity. The thickness of the weldment (1.5 in.) is certainly indicitive with the possibility of sigma if heat inputs are not adjusted and monitored closely.
I can almost guarantee the FCAW (309L) is up in the 12-18 FN range. I'd like to know the final outcome and cause of this problem since these scenarios are accepted all day long with no problems as Richard is experiencing. For some unknown reason, I just feel it has something to do with the restraint of the steel at that thickness. But sigma is also a possibility.

By Richard V. Roch (**)

Date 05-02-2007 11:20

Gentlemen
The strong backs are just to keep the plates from bowing to much. As for the heat input, it was approx. 23.0 kj/in. We have not checked for intermetallic formation or the ferrite level. We are going to eliminate the tri mix in favor of the 75/25 Ar/CO2, we are going to bring the PWHT soak times down to 2 hours, we are going to offset the plates during fit-up to elimintate a strong back and we are going to use a brand new roll of FCAW wire. If this does not work we will probably need to start looking at the FN on the filler and doing a more in depth investigation. I appreciate all the feed back.
Thanks

By Lawrence (*****)

Date 05-02-2007 12:55

Richard

Ive been watching this thread with interest

I don't know how much time/resources you have to dedicate to this project. But if you change all those controls at once you may never know what actually caused the cracking. (which from a production standpoint may not be all that important) I realize that making a good weld is what you are actually after, but it would be interesting to change one control at a time... ie. change to 75/25 but keep the strong back and heat treat the same and see what happens.

Do keep us posted on your progress!

By Richard V. Roch (**)

Date 05-02-2007 14:33

Yes sir I am dealing with time constraints and don't have the luxury of doing the R & D work to pin point the problem to the original failure. We are going to weld this coupon this Friday or Monday and should have testing results by mid-week next week. I will definitely post the results.
Thanks everyone

By Milton Gravitt (***)

Date 05-02-2007 14:51 Edited 05-02-2007 16:18

Yes let us know the results. I'll keep my fingers cossed. Good Luck!

By js55

Date 05-02-2007 14:51

I have read the other posts and don't really see anything inherent in gases (Helium)or the alloy combination that would cause such a cracking phenomena, so I decided I would think outside the box a bit.
Stainless steel FCAW wires are sometimes manufactured with additions of Bi2O3 to assist easy slag removal and to a lesser extent produce that pretty blue color on the weld surface. These wires are known to embrittle at high temp. The Japanese were the first to write papers on this phenomena.
To my knowledge most manufacturers have either quit producing these wires or produce non Bismuth wires for high temp service. It is possible that you have a Bi2O3 enhanced wire (perfectly acceptable for most applications) and in combination with your heat treat (though the embrittlement to my knowledge takes longer than heat treat regimes-it is possible) it is embrittling the weld.
Now, this phenomena is related more to the low melt film produced by the Bi2O3 and hot ductility issues, but its possible that under the extreme strain of 180deg bend (20% outer fiber elongation) the embrittlement is manifesting itself.

By Richard V. Roch (**)

Date 05-02-2007 15:53

js, is there a website I could access for this information? It could be something I may need to look into.
Thanks

By js55

Date 05-02-2007 16:25

You might give the fellas at Euroweld a call. Roger and Bill probably know as much about the Bismuth issue as anybody.
You might try googling them, or various combinations of key words around bismuth, flux core, welding, etc.

By GRoberts (***)

Date 05-03-2007 00:38

I believe I have run into a bismuth problem one. What I figured out from my experience plus a very good paper from Metrode was that my problem was due to re-heating after solution annealing, so it would obviously not be related to this issue. There wasn't a specific problem listed in the Metrode paper of low ductility at room temperature due to PWHT, but with service temperatures above either 1250F or 1020F depending on if you listed to API or the Japanese. I agree with JS55 that test temperature and time at temp in this case don't fit the classic bismuth problem profile because I don't know of a specific problem with ductility of bismuth containing wires after PWHT, but if problems are being experienced, using a non-bismuth containing wire could certainly eliminate another variable.

Another thing I found out in talking to 3 or 4 FCAW wire manufacturers, is that they all said that they add bismuth to their 316LT0-1 wire, but not thier 316LT1-1 wire. So if you use an all-position wire, more than likely, it did not have bismuth additions. However, it is best to contact the manufacturer to be sure.

By js55

Date 05-03-2007 13:24

Thanks Greg,
I'm not even myself convinced that its a Bismuth thing. But its worth a look. I didn't have time to review the articles I have on it, and its been about 6 years since I read them, but I remember the Japanese running some Varestraint tests that may indicate something similar. Its been awhile. And I know that most of the research on this issue had to do with high temp services and hot ductility reduction, but in service there isn't the extreme deformation you would see with a bend test.
I just don't see anything in the gas mix, the parameters, or the alloy combination that would necssarily lead to such cracking.
I'm guessing its an intergranular cracking, which often means liquid films with high alloys.

By Richard V. Roch (**)

Date 05-08-2007 17:59

Gentlemen I am convinced that a thin layer of oxide remained in-between passes. The welder only wire brushed the flux off of the weld beads, but I think a thin layer of oxide built-up remained between all those weld passes. I think this is what caused the bends to break. What do you guys think?
thanks

By js55

Date 05-08-2007 18:20

I think for us to take the analysis to the next level, or concur with your evaluation a more precise description of the failure needs to be provided. In other words, does the failure appear as though it was generated from a planar defect, which would indicate a lack of fusion consistent with your idea-does it appear as more of a rounded defect such as possibly a slag inclusion or porosity-or does it appear to generate from a much more localized region as a fracture which may be more consistent with Chucks idea of residual stresses from the strongback.
Having said all this, if you get your welder to brush his welds more thoroughly and the poroblem goes away, this is moot. And you may have your answer.
It really is important to segregate your procedure elements/variable from your perfomance elements/variables.

By Richard V. Roch (**)

Date 05-08-2007 19:41

1-1/2" thick plates, a single vee joint. Side bends with failure on three (3) specimens at approximately 3/8" down from the cap and approximately 1/2" over from the Stainless Steel side. One specimen failed at the carbon steel fusion line. I was just looking at the coupon we welded and noticed quite a bit of oxide built-up on the stainless side of the coupon. I know from past experience that oxides can cause problems with hardness issues, I'm not so sure about stainless steel.
Thanks

By Richard V. Roch (**)

Date 05-15-2007 18:48

Gentlemen
Sadly I've been informed by the lab that 3 out of 4 bends failed. I am at my wits end. I now firmly believe the heat treat is having a significant affect on the weld. I would appreciate any additional information on heat treating stainless steels.
Thanks

By Milton Gravitt (***)

Date 05-16-2007 18:04

Richard did you change the way that you where welding the test plates.

By Richard V. Roch (**)

Date 05-17-2007 12:36

As I replied in the 5-2-07 posting, I changed gas from tri-mix to 75/25 Ar/CO2. I also did not restraint the coupons during welding and we limited the heat treat to a two (2) hour instead of four (4) soak. The metallurgist look at a polished sample and told me the weld had a dendritic structure, which is causing the weld to separate during bending. I gave the welding engineer all the parameters and she cannot understand why we are having this problem. She thinks the weld may be cooling to fast due to the small (thin) weld beads. I explained to her that this is another thing I changed from the original procedure that failed, I ask the welder to keep the beads a little smaller, thinking this might help.
This is the latest folks, the weld has a dendritic structure HELP.
Thanks everyone

By sean32

Date 05-16-2007 20:34

I think the PWHT is having an effect on the bends. You may try swithching filler metal from the 309 to the Flux cored version of Inco 82 (ENiCr3T0-4). We join stainless to carbon quite often and have better luck with the Inconel wire when the weld is subjected to higher temperatures either in PWHT or operating conditions. This wire is best ran with a 75% 25% gas mixture.

By Richard V. Roch (**)

Date 05-17-2007 12:37

So far most of the experts have told me that the heat treat should not have an effect on the weld coupon. This is the first time I have attempted this.
Thanks

By js55

Date 05-17-2007 13:29

Richard,
Keep in mind that we offer advice based upon our experience and our study, and discussions with others of expereince and study. But when it gets right down to it, sometimes it just happens some way without any 'obvious', and I emphasize obvious, metallurgical reasoning behind it.
Were I in your shoes (not intending to necessarily dispute my own position); a frustrating result-a procedure to qual-a deadline to make; I would leave no stone unturned until those dayum bends pulled. If sean's advice on nickels and heat treat works after nothing else has, all I would ask is share it with us.

By Richard V. Roch (**)

Date 05-17-2007 19:19

I totally understand, and let me say this is one hell of a site where folks can get helpful information. Everyone who replies to this posts is contributing very useful information. Please don't misunderstand some of my comments. I will continue looking into this problem and hopefully something good will come from this.
Thanks everyone :-)