Bob Puhakka on Bifilm theory

I think you are right; I think I am reading something into the story that is not really true.

I am really glad we ran across Bob.
As I have mentioned before, I learned more in a month about castings from Bob than I did in the previous 6 years.
Its like finding a big diamond in the middle of a farm field while you are plowing frozen ground.
All of a sudden, plowing does not seem so bad anymore.

.

 

It seems very plausible that hydrogen will escape the freeze front (lower solubility in the solid) so long as it can migrate away. An oxide layer, or the gap in a bifilm, could arrest that migration.

 

Did another see through flask today. Smaller sprue and runner (0.04 sq in) with a pouring basin and foam plug. I used some high temperature glass I ordered and it did better. I didn't clamp it as tightly either.

Thought you'd want to see me bust my glass.



I've started using a runner into a spin riser with a branch to the pattern but that was inconvenient with a vertically split flask.

 

Too bad your glass broke! But these visible-flow pours are very interesting and instructive.

You did get a nice smooth meniscus that rose smoothly up the mold. And you could see the plug actually held long enough for the metal to get a good head developed before it released.

With respect to seeing less porosity in your casting than the n the muffins: do you suppose that pre-existing bifilms might clump together if given a chance while they swim around during your muffin melt? They might then be skimmed. That might have a cleaning effect. As they form during a pour, they form and freeze singly or in clumps but don’t have an escape route.

Denis

 

I don't know. I assume cavities with oxide surrounding them will float in an undisturbed melt. The oxide will not melt and they are hollow inside. I also don't think the oxides are exactly the same density so they would tend to float or sink upon melting if you don't stir the pot.

For what I'm doing (no work for NASA) I'm not too concerned about broken oxides in my castings, kind of like aggregate. Bifilm, where an oxide layer has folded upon itself and creates a potential gap is also not a big issue as long as they are not connected. Since I do no degassing, and I'm seeing very few instances of cavities which may be due to hydrogen collecting in those gaps, I'm believing there is not a lot of places for it to collect. However I still believe the hydrogen in an ingot escapes as it's melting.

Yeah, I'm violating Rule #1 but getting good results at the level I'm inspecting, which is about 0.001". I say that because I can see my microscopic image move clearly when I move my mill table half a thousandth, but the image is not clear enough to discern cavities smaller than half a thousandth.

Obviously if you use good clean metal and keep it clean (which includes sources of contamination like oil on it) you will have a better melt. I think not stirring, pouring from under the crucible liquid level then using a pouring basin and plug goes a long way toward cleaning up a less than ideal melt. I can see the oxide layer surrounding the stream from the spout to the basin, and that protects the metal inside.

Breaking the glass was not unexpected this time and I took more precautions over last time. I was surprised it took two minutes and hope the air jet will eliminate it. I don't have any more glass pours planned, but we'll see.

The flow in the runner was faster than I expected, but I have really been enjoying using the foam plugs. Did you see the smoke shooting ahead of the metal? At first I thought that was metal in the sprue but the slow motion looks otherwise. With the gate coming off a runner that leads to a spin riser the flow should be slower into the pattern.

I have only cut up one muffin but have no reason to believe the others are much different. I think porosity will float if it is surrounded by an oxide which does not readily melt.

Regardless of reality, my castings are getting better assuming the bifilm theory is correct. When discussing theories I like to remember electricity. The mechanic thinks electricity comes out of the positive terminal of a battery, the engineers think its the negative terminal, the scientists tend to think it's the positive terminal. My car still starts.

 

Seems like the foam plug started leaking a thin stream, and then a big jet stream, which went into the mold way too fast.
Cool simulation.

.

 

Yeah, I'm stuck on rule #1. Campbell suggests using a dry hearth furnace. Does that remove all the oxide films?

 

The short answer seems to be you don't. They are still present somewhere. They can either float to the surface, or precipitate to the bottom of the crucible, or enter the casting system. Somewhere at the start of my experiments I charged with junk sprues and castings and ended up with much cleaner metal in my test without degassing or fluxing. Where did they go? Holding without agitation is one method mentioned to allow the folded bifilms to rise to the surface where they can be skimmed off.
Campbell mentions that the methods for refinement of aluminum is a closely guarded secret of the industry. The rotary degassing and flux additives have a tendency to add moisture (hydrogen) rather than reducing it , but raise the floating bifilms with the mechanical action of the inert gas bubbles.

 

Great test Andy. There's been a lot of commentary in this thread to digest over the past couple days. In regards to the speed I think adding the riser or a sacrificial area before the pattern cavity would have helped. As the cavity filled it appears that the meniscus held without overlapping any oxidized layer except at 2:36 where it flowed back over the initial entry of the cavity. The only other thing about the fill that was flawed in the cavity (from my perspective) was how the flow front launched at 2:33. Otherwise there wasn't really any jetting or splashing.
I'm very interested in the foam plug. It seems like when the metal met some resistance at the plug (cooling) it just looked for an easier way around. I'm thing more like a manhole cover than a plug. Thanks for taking the time to show the test. Very instructive.
I don't know if I heard Campbell or Bob P say they don't recommend degassing. Maybe they have, but I think they're critical of the way we think it works ie that the bubbles mechanically push the air-filled biofilms upward as opposed to dragging out Hydrogen. I may be off base there.

Pete

 

On the breaking glass: Why not pop it off 30 seconds after the pour? It took 2 minutes to heat up enough to break I think. So getting it off the heat source might prevent that?

Denis

 

This is going back, but don't get the sense that the two are related. Hydrogen is soluble and so is in the melt and can reach a saturation point, but still not be visible to the naked eye. The oxide is dry and has a higher melting temperature, if at all, than the molten charge and is visible. Especially when unfurled around an air bubble.
In the blower video are we led to believe the guy runs to the vacuum chamber with molten aluminum, vacuums it and when it solidifies, he cuts and inspects it and determines that it was a hydrogen bubble? More likely a bifolded oxide film that called for more of the degassing to mechanically float those films, while they thought they were removing hydogen. They used best practice to fill the ladles by immersing them in the melt, then use a waterfall sprue to dump the metal into the mold. Back pouring the riser has some validity to insure hot metal to feed that end of the casting based on conventional foundry practice. But from my video and Ironfarmer's first video we can visualize the splashing that took place at the bottom of the sprue and the effect of sessile height drop across the length of the casting before the mold filled vertically. In spite of this, they got a sound casting which leads me to believe that cleaning the melt was the most critical activity.
If Bob were to do this as a counter gravity pour the gates would be at the bottom of the mold and perhaps with blind risers in the cope. He would also get a sound casting based on his past results, but also start with a clean melt.
Rule#1 again.

 

Yes, that was a good one. I didn't see the smoke until it was mentioned, and had thought it was a spurt of aluminum which would have negated the reason for the plug. It does illustrate the need for a complete system of sprue, runners and gates so the filling of the mold proceeds evenly.

 

No, just a reduction, and then he mentions an instance of a dry hearth where the furnace tender uses the rake to push that mass of aluminum oxide husks and other trash into the melt rather than out into a receptacle to be reclaimed and in so doing voids any advantage the method provided.

This stuff is making my head hurt.

 

This is an important point to me. The little that I've read about hydrogen solubility in liquid aluminum indicates it behaves like lots of other liquid/gas solubility: solubility increases with temperature. One reference on measuring the solubility of H2 in AL noted that at low temperatures (still molten) the amount of hydrogen is very low. At higher temperatures the solubility increases.

Therefore, a charged crucible, upon first melting, will have melt fronts right at the melting temperature exposing hydrogen pockets. That hydrogen will likely escape. If the crucible is half full of hot melt and you plunge an ingot under the liquid level one would expect the surrounding aluminum to absorb most if not all of the hydrogen released. I try to not plunge material below the liquid surface but keep adding as the melt settles. Above the liquid level I believe substantial hydrogen will escape. I add material this way to keep the crucible as cool as possible, since it absorbs heat faster if it is cooler than the furnace temperature. But it also should aid in allowing hydrogen to escape and maybe even burn.

 

At first I thought the thin stream sneaking around was aluminum but have since been convinced it is a smoke stream. The color is different, and when the obvious aluminum stream comes through it is reflective. Looking carefully you can see the smoke ahead of the metal. Along about 2:31 the metal is just through the foam and smoke has filled the sprue. I was sure the foam would create a vapor which would push ahead of the metal but was surprised it was smoke. Duh!

I agree the metal went into the mold way too fast, but it was in fact right at 0.53 m/s which is right about the runner sizing guideline of 1/2 m/s (unless I'm not remembering that number). In any event, it has convinced me that I want no more than maybe 1 ft per second flow into a mold. Funny that I think we all have had successful castings with flow much higher than that and it does not seem to wash green sand away.