More Fun With Aluminum

Date 04-09-2009 18:13

Hello Curt, here is one stab at your question. There is a slim possibility that the parting process for making the round pieces might be causing some surface smearing of the material. This surface might then exhibit the "surface cracking" that your photos are showing due to the smearing acting as sort of a "lamination" type condition. That's my stab at it. Do these cracks have much depth to them, or are they pretty much just superficial? Best regards, Allan

By OBEWAN

Date 04-09-2009 18:25

We saw a similar condition in Inconel next to a braze joint and it was indeed a surface condition caused by machining. They were like chip break indications. We sectioned through several and never found a crack with the microscope.

By Kix

Date 04-09-2009 20:18 Edited 04-09-2009 20:21

Man, what kind of camera are you using to take those pictures? That thing takes one hell of a good close up!! Are you welding on a anodized part or does the anodizing come later? If the anodizing comes later, are those cracks there before it goes to anodize?

By bozaktwo1 (***)

Date 04-10-2009 11:57

Canon Powershot SX10IS, set to manual in super-macro mode. Edit: Oh, and a Fenix LED flashlight for the contrast. :)

The part is anodized following weld.

I looked into Allan's idea of smearing, but ruled it out due to 1. inconsistencies in the occurrence among my sample group of 16 parts; 2. I personally witnessed the cleaning process prior to welding. The chemist at the plating vendor thought it might be a dilution issue between filler and base metal. I think that's half the problem. Lawrence, that paper brings up an interesting point; however the cracking is not occurring in the weld. Rather, it originates in the base metal and propagates into the weld toes (and only from the thicker piece). Not to discount the idea completely of the PMZ cooling too rapidly; I suspect that is happening anyway.

My corrective action for this is going to begin with spinning new parts, cleaning and welding, and then sending half of them to the same vendor and the other half to a different plater. I am going to 10x the entire lot beforehand; of course if what I suspect is true I won't see indications until after the chromic anodize. We have made these parts for decades without this problem, which is why it's so disturbing. The experiment will be the use of two different vendors for the same process; then we can evaluate the results and see where we stand.

Thanks, everyone for your replies.

By Lawrence (*****)

Date 04-10-2009 00:10 Edited 04-10-2009 00:28

I'm going to go with Liquation cracking in the partially melted zone (PMZ)

The Eutectic near the weld toe (PMZ) gets mushy and when it quickly cools bad things can happen.

Here is support. (Abstract found by our mettalurgy instructor)
http://files.aws.org/conferences/abstracts/2003/04b.pdf

This does not supply a solution.... other than a switch to 4043

Don't know if it helps.. .but it's interesting

Edit

Here is a real good powerpoint that has an image that looks alot like what we see above... Look at Page 6
http://www.sut.ac.th/Engineering/Metal/pdf/MetJoint/04_Weld%20microstructure02.pdf

And a bit more from TWI
http://www.twi.co.uk/content/jk21.html

Liquation cracking

Liquation cracking occurs in the HAZ, when low melting point films are formed at the grain boundaries. These cannot withstand the contraction stresses generated when the weld metal solidifies and cools. Heat treatable alloys, particularly 6xxx and 7xxx series alloys, are more susceptible to this type of cracking.

The risk can be reduced by using a filler metal with a lower melting temperature than the parent metal, for example the 6xxx series alloys are welded with a 4xxx filler metal. However, 4xxx filler metal should not be used to weld high magnesium alloys (such as 5083) as excessive magnesium-silicide may form at the fusion boundary decreasing

By 803056

Date 04-11-2009 01:58

I have seen it called "arc erosion" by one welding standard used by the Air Force.

As Lawrence mentioned, it is a shallow melting of the surface of the aluminum alloy by the welding arc. It is more prominent using AC. I don't know how much help will be derived from switching to a different filler metal since it appears beside the weld bead.

A vigorous brushing with a stiff bristled stainless steel brush is usually sufficient to remove the very shallow surface cracks. Removing the heavy oxide from the weld and immediate area just before welding will help minimize some of the "cracks" and switching to DC and helium can also reduce the tendency to develop "arc erosion".

I would imagine the use of a new inverter power supply with variable frequency that permits the welder to "focus" the arc may reduce the area beside the weld subject to the "arc erosion".

Best regards - Al

By bozaktwo1 (***)

Date 01-29-2010 14:30

Sorry it took so long to update on this.

The problem was corrected by switching from a sulfuric acid wash to a solvent wash in the plating process. All is well!

Thanks for all the input to this topic.

By js55

Date 01-29-2010 14:51

I think Lawrence had it dead on. It looked like typical HAZ intergranular liquation to me. The sulphur in the acid wash seems is the culprit, creating the low melting eutectic.
Looking close you can see the phenomena is almost fractal, and fades the farther from the weld it gets.

By ssbn727 (*****)

Date 01-29-2010 17:29

Just out of curiosity Curt... Why was a sulfuric acid wash used in the first place? And what was the dilution ratio of the mix? Could it be that they didn't focus on what the material was when they cleaned it prior to plating and nobody wants to admit this???

Respectfully,
Henry

By bozaktwo1 (***)

Date 01-29-2010 20:54

I don't know the mix ratio of the solution used, but I suspect that the cleaning solution tanks were not properly prepared prior to the process. Additionally, the sulfuric acid wash was far too harsh for the 6061. This had the added inconvenience of making the mating surfaces and the threads in the center bore too rough for proper fitting. The wash was used because that's what the supplier most commonly uses to wash parts prior to anodize. I was also not too happy with time in tank, as it wasn't recorded accurately each time the process was performed. I don't have the time or the resources to chase this to an absolute end, and this supplier does a lot of good work for us. I suspect they just got moving a little too quickly on these parts. So I will keep an eye on these for a while. The last batch came back looking a lot better.

By 803056

Date 01-29-2010 21:22 Edited 01-29-2010 21:40

We used to use a buffered solution of sodium hydroxide to remov surface oxide followed by a nitric acid to remove "smut", followed by a clear water rinse as the chemical cleaning regime.

The "craze cracking" was associated with aluminum that was idle too long before welding and it developed some oxide on the surfaces. However, even the freshly cleaned aluminum could experience some of the problems depicted in the initial photographs. As I mentioned, the cracks were very superficial and could be easily removed by hand wire brushing. We didn't experience the problem to the same extent using DC as we did when using AC.

Are you seeing this problem right after welding or after anodizing? If after anodizing, are you checking them before anodizing?

Since the "cracks" are in the area adjacent to the toe of the weld I suspect it was similar to the problem I experienced rather than attack by the sulfuric acid. If the sulfuric acid was the culprit I would expect the attack to be more general and not confined to the area along the toe of the weld. If it was liquation cracking the cracks would be confined to the partially melded zone at the very edge of the toe and the penetrant test would provide positive results before wire brushing as well as after wire brushing. Hand brushing the weld and adjacent areas would not smear the metal to mask the cracks as would be the case with power brushing.

If you were to weld 6061 without filler metal, the weld would would look like a spider web of numerous cracks. When welding with AC the arc tends to wander somewhat, and in doing so there is a very shallow surface melt that can occur (arc errosion). In essence I suppose one could say that you have a partially melted zone in that case and liquation cracking, be it very limited in depth, could become a factor. However, I differentiate between liquation cracking that occurs in the partially melted zone of the weld bead (extending under the entire weld bead) and a surface condition that can be removed by wire brushing. It would be interesting to see how deep the "cracks" extend.

Best regards - Al

By bozaktwo1 (***)

Date 01-29-2010 21:31

I personally examined the test coupons with a 10x prior to anodize. We did one set with the sulfuric, one with the solvent. Definitive results, although there was some minor discoloration in a few spots on the solvent-washed pieces. I attribute this to the cleaning process, or some slight variation in gas flow...maybe the welder breathed on it. :) Now, it still kills me that the problem only occurred on the thicker section. owever, I have spent enough $$$ trying to solve this mystery. I now have a process that is under sufficient control for repeatability with acceptable results. That's what makes product move out the gates!

By ssbn727 (*****)

Date 02-01-2010 08:20

Hi Curt!

You may find this interesting:

http://www.eaa.net/eaa/education/TALAT/index.htm

Here's an entire library from across the pond for you:

http://www.eaa.net/eaa/education/TALAT/index.htm

Respectfully,
Henry

By Tommyjoking (****)

Date 02-01-2010 23:58

I have followed this one with significant interest!!!

Glad you worked it out Curt! Nice commentary by all.