We have run into a curious situation that I'd like some insight on.
We cut test plates (#1) from some FB 3/8 x 7 material, A36/A529-50 (dual cert). The material works OK for our E71T-1 and E7018 testing in 2G, 3G, & 4G positions. Face and root bend tests are no problem (at least for most experienced welders).
However, when we run our stainless qualifications using the same plates, A36 base metal with 309L filler FCAW, the root bends all crack out. That is, the plates break along the stainless/carbon interface. The face bends seem OK.
The (#1) flat bar chemical and physical analyses are within spec for A36. Tensile is 66K, yield is 47K, elongation in 8" is 26%. The Mn is pretty low at 0.65, C is 0.15, and other elements are mid range for A36.
We tried material (#2) from another shop doing the same type tests as we are and had success. That material is Canadian spec G40.21-96 50W with Mn = 1.24, C = 0.17, tensile =83K, yield =61K, elongation is 21.5%. No problems with the root bends when using 309L filler.
So then we tried some of our own material (#3) , different heat. A529-50 with tensile= 88K, yield = 66K, elongation = 28%, Mn = 0.88, C = 0.16. It would seem to be comparable to the #2 batch with even better ductility, and Mn roughly midway between the percentages for #1 and #2. No dice, same problem with root bend cracking at the stainless/carbon interface.
Before I go any further, these are all FB 3/8 x 7 materials an cut so that grain direction is unquestionably transverse to the weld, as it should be. Also, we tried some different brands of 309L FCAW with no improvement. Backing bars are the same material and heat in all cases.
We will try a piece of plate with Mn at 1.45, C at 0.11, and roughly equal physical analyses to see if manganese is the dominant factor. However I was wondering if anyone has had similar experiences or anything that might explain the root bend cracking?
My main concern is: we can shop around for material that will produce good results in order to pass the tests. Admittedly, the bend test puts stresses on the filler and base metal that wouldn't normaly be encountered. But with no practical control over the P1 base metals being welded in production or in the field, how can anyone tell if those welds will have problems or not?
Chet Guilford
My guess is that the main problem is the difference in the mechanical properties of the filler metal and base metal. The 309 (especially in the root where it gets lots of dilution) will be much stronger than the A-36 material. I would guess up to 90ksi or higher UTS. If you compare that to the 66 ksi of the A-36, the difference in strength places a lot more strain on the base metal adjacent to the weld during the bend. The Canadian metal was 83ksi UTS, so it was a lot closer to the weld metal. To confirm if this is the case, try a longitudinal root bend an a weld with the A-36. If it doesn't crack, you will have discovered your problem.
As far as your production welds, if a mismatch is specified with regards to tensile strength, the disigner should take this into consideration. There are situations where it can be detremental (like fatigue), but normally it will not cause problems.
The longitudinal bends are a good point. I'll pass that on to try.
The only problem I see with the UTS mismatch idea is that our #3 material has an 88K UTS and that also gave problems while the #2 material didn't. I'm not saying you're not right just that this is puzzling.
The only possibility I have been able to come up with in regards to Mn content is that when welding CS with austenitic SS filler metal, due to segregation while solidifying, there is always a very thin layer of martensite next to the base metal. It is usually thin enough that it doesn't cause any problems. If you were having a problem with it though, the higher dilution at the root would tend to be the place it might happen. However, Mn is an austenite promoter, so the more Mn present, the thinner the martensite layer might be. This is just a guess. (there was an AWS article by Kotecki regarding a martensite boundary on the WRC FN diagram with this info a while back). If you wanted to investigate this potential problem, take one of the bend tests that would fail if bent and stress relieve it at about 1150. That would temper the martensite and increase its ductility. If it bends after this, it could indicate that this is the problem, but wouldn't rule out many other problems. Another way to check would be to weld up a test with nickel based filler metal like ENiCrFe-2. The alloy content is high enough that no martensite layer is formed. If it bends ok, that martensite layer could be a culprit.
Thanks for your response on this. The plates cracking with low(er) manganese content but not with the higher content would tend to back
up what you're saying about martensite/austenite.
We haven't had a chance to try the longitudinal bends yet but we will. I don't know if we'll be able to stress relieve a coupon or not, but it would be good to try. But I think I will try the ENiCrFe to see how that works out. Maybe it will prove to be a good indicator on future heats of material.
It's just puzzling because we've used the same method of qualifying welders with stainless for more than 10 years and haven't had a problem with it. But now, with this batch of material every plate cracks out.
Again thanks for your thoughts and I will post a follow up if we find out anything useful.
Chet Guilford
Hi Chet
I believe that what you are experiencing is "interface cracking". This is a hydrogen cracking phenomenon that occurs in dissimilar metal welds in the "diffusion zone". The diffusion zone is a very thin zone where the base metal did not actually melt, but due to the very high temperatures, some of the alloying elements from the weld pool has diffused into the base metal. This zone can then form martensite upon cooling. Generally this is not a problem, because it is such a thin zone. However, if you have too much hydrogen around, then you can get cracking in this zone.
Interestingly, the higher the alloy content of the weld pool, the less this is a problem. Also, as with hydrogen cracking in general, the base metal composition (even from micro-alloying additions) is rather critical.
You may find that your sudden problems may have something to do with a moisture contaminated batch of filler. The variations depending on the different batches of plate will probably have something to do with the micro-alloying additions between the different base metals. (Including the Mn contents.)
Try using a different batch of filler that is definatelly "dry", and see if your problem goes away.
Regards
Niekie Jooste
Fabristruct Solutions
Forum
