Weld impact toughness.

Date 08-23-2007 18:48

I'm thinking the lab tech didn't hold his tongue right when he let the hammer go. And I really only say this in half humor. Erratics, as anybody that has done enough toughness testing knows, are not that unusual.
A couple problems here that I see though. You say you obtained the welded plates. Does this mean you don't know how they were welded. The position of the Charpy notch in relation to unnaffected weld metal, weld metal high temp HAZ and weld metal Low temp HAZ can be critical.
It could be that even though your notch was relatively at the same depth, or even exactly, you still had more of one type of microsturcture in the notch in one specimen as opposed to the other.
If you're going to all the trouble of doing charpy curves doing a macroetch on adjacent material wouldn't hurt. Just to see what you got.
Also, Impact strength emphasizes just that,as opposed to lateral expansion that emphasizes ducitlity. and keep in mind that ductility may or maynot vary at the same rate as strength with a change in temp.
In the end, Charpies are sledgehammer. If you want real material properties you're better of doing CTOD's. But who really wants to do that. Its expensive and time consuming.

By VeryMetal Date 08-23-2007 18:58

Erratics are normal, yes, but a variation of more than 200 ft-lbs? I was thinking that there may have been a tempering effect in some of the weld material, given that it was a 1" thick multi-pass weld. I know the microstructural variation would change the impact toughness, but I've never heard of low-alloy bainitic/martensitic material reaching 245 ft-lbs at -30 F. This is what I'm especially confused about.

By HgTX

Date 08-23-2007 19:09

Are those your average values, or did you just do one test at each location?

Hg

By js55

Date 08-23-2007 19:20

At this point I have to say I do not remember what the chemistry of T-1 is and therefore, your answer may be something diluting from the BM.
A353-A553 9% Ni steel is martensitic/bainitic. The Ni is of course the primary element that gives it its toughness but its martensitic and bainitic nonetheless, in either the quenched or normalized condition.
What is the chemistry of T-1?

By VeryMetal Date 08-23-2007 20:52

Just one test at each location. Unfortunately material was limited.

I took the chemistries of two of the very different specimens, and they were fairly similar. There was probably a little bit of dilution, but nothing that would explain the huge mechanical differences.

By CWI555

Date 08-23-2007 22:02 Edited 08-23-2007 22:05

If the notch is not cut right, it can throw off the test, a small area of varying microstucture, a variation in the chill, the hammer not delivering it's energy (catching on a small notch) Even the best lab techs are going to run across the one off situation. I've performed a considerable amount of CVN's and so yes, those variations are not only possible, but are likely to be seen if you perform enough of them. 1:1000 odds sooner or later you catch the 1.

By VeryMetal Date 08-23-2007 22:33

Our lab has a special tool to make sure the center of the notch and the hammer are lined up. I examined the specimens after they were broken, and the specimens looked good as far as where the hammer struck, the notch, dimensions, etc. There was some microstructural gradient on the samples, and the direction of the gradient was parallel to the notch. Here are my test results:
Temp ft-lbs
+5 131
0 252
-5 267
-10   141
-15   86
-20   17
-25   14
-30   245
-35   94
-40   107
-45   149

By js55

Date 08-24-2007 15:08

This is unusual to be sure. But I'm thinkin your still sittin on the upper shelf. 14/17 though low is still not what I would call brittle. And the Charpy test is, using ft/lbs, a test that combines strength and ductility, emphasizing strength. Lateral expansion tests would emphasize ductility.
Toughness is a term we toss about but I'm not sure anybody really knows what it means. There is a functiuonal definition to be sure, but functional definitions don't mean we really know what is happening.

Perhaps, and this is just brainstorming here, while maintaining some ductility with decreased temp you get increased strength, and is therefore reflected in the increase of impacts at lower temps. But IMO this would only work if all is still on the upper shelf. Once brittleness steps in an increase of strength at decreasing temps would not be reflected in the impact results. IMO.

Have you done an FATT analysis?
Does it appear you have dropped below your DBTT?
I would be surprised if 14 demonstrated a majority volume fraction of brittle appearance.

By js55

Date 08-24-2007 15:21

A question:
You still haven't clarified what "recently obtained" means.
A Charpy curve will be of little help to you I would think if you don't have weld parameters to go along with it. You are after all testing the welds and not the BM.
And after doing a search on T-1 and realizing it contains Cr, Ni, Mo, V, and B, in reasonably large amounts, there's no telling what kind of soup you have in the weld metal, especially mixed with high Mn and Si from the S-6.
You have elements that respond well to Q and T. And that is exactly what takes place in a multipass weldment. So I suppose with all of this variety of elements, a multi pass application and no telling what mircrosturctures were emphasized in each Charpy bar, it could be argued that erratic results would be inevitable.

By VeryMetal Date 08-24-2007 21:26

Recently obtained means I had it welded up last week. The welder used the specified weld parameters for our company.

Again, the chemical analysis showed minimal dilution. The B, Cr, and V levels were no different between the Charpies. Carbon had a difference of about .01 wt-% between the 17 ft-lb and the 245 ft-lb sample, and Mn a difference of 0.03 wt-%.

The 17 ft-lb sample is the only one that's near the value for the AWS spec

By js55

Date 08-27-2007 17:55

In the end I would normally chalk this up to just one of those things, but what bothers me is not so much the vast difference in the high and low result but the tendency to drop to 14 and then to rise again. This indicates some regularity that argues against abberants.
It would be interesting to do microfracture analysis, and note any differences between those areas of high toughness and those areas of low toughness.
It still seems to me your on the upper shelf with some metallurgical regularity causing difficulty at certain temp ranges.
Note that the rise at the lower temps does not quite come back up to the level of toughness of the higher temps. Throw out your -15, -20, -25 and you have fairly regular and typical upper shelf behavior as it slopes off to the ductile to brittle transition.

By Sean

Date 08-27-2007 18:31

Personally, I would try doing some more metallography on the coupons.

Recently, we did some PQRs on some HSLA steel sections and we had some "interesting" results with CVNs too. The procedure tests kept failing due CVN values in the base metal. It was strange because the MTRs and subsequent base metal CVN testing done by both the manufacturer and supplier were great. The results were also consistent with each other. After the first two attempts at the procedure tests with CVNs failing in the base metal we took a look a closer look at the microstructure of the base metal. The microstructure through the flanges varied a little bit; however, we did find that the microstructure on a section of the flange, where we took the base metal CVNs was much different than the rest of the coupon... I think it was bainite (not sure though). At the end of the day we made sure our PQR coupons came from the more consistent microstructure and the results came out as expected, similar to the original CVN values obtained by the others.

Now the first thing we do if we get "interesting" CVNs, after we check the parameters is check out the microstructure.
Sean

By js55

Date 08-27-2007 18:56

Sean,
I thin thats right. If any more digging for answers is going to take place thats where the money needs to be spent.

By VeryMetal Date 09-04-2007 23:10

I had some metallography performed on the low and very high samples. The difference was from large columnar grains with ferrite fields (low impact values) compared to very fine recrystallized zones of martensite (high impact values). Thanks for all your suggestions.

By js55

Date 09-05-2007 10:51

Clearly the columnar grains are as welded strucutres, and the fine grain recrystallized zone is the low temp HAZ, and I am assuming an equiaxed grain structure as well.
But if I may, for curiosity sake, does the as welded columnar zone appear as a ductile or brittle fracture? As I said, I think th ewhole thing is still within predominantly the ductile zone. There isn't any other way, other than rather thin logical contortions, as far as I know, to explain the increase of toughness at continuing lower temps.
PS: D10 should be coming out with a document of toughness testing soon which IMO is a very good one. Jon might be able to help us on a time line here. I'm not sure where in the priocess the doc is right now.

By VeryMetal Date 09-05-2007 16:03

The large grain structure fractured in a brittle manner, and there were no real shear lips to speak of. The toughness increase at low temperatures was because we basically tested two different materials with different heat treatments and chemistries.

By js55

Date 09-05-2007 16:17

Ah. The fly just came buzzin outa the ointment.

By VeryMetal Date 09-10-2007 15:20

They weren't actually different materials, it was the same weld pad. Base metal dilution and (chemistry change) and recrystallized zones caused the material to behave differently in different places.

By Joseph P. Kane (****)

Date 09-11-2007 00:47

I have seen a curve like this once before. It had different temperatures, but the curve was similar. It was the Nil-Ductility diagram on a fuel tube from the Chernobyl Nuclear Reactor.