Bob Puhakka on Bifilm theory

Thanks, Pete. I did learn a lot from this pour. A lot of information here for sure. And it is difficult to sort through it.

The first section of the pattern was indeed intended to be a blind riser. On my first glass pour I had the runner entering below the riser and exiting below with a dam in between. That indeed slowed the flow. My normal way of gating now is to gate perpendicularly off the runner like this.





The runner directly to the open riser will have to slow the flow through the gate into the blind riser and pattern.

I find a small plug works very well. It can't float, and you can make it short to burn through quicker, but I like it about 3/8" long to give me more delay. I also like the smoke flushing oxygen out of the sprue and runner ahead of the metal front.

I'm not set up for degassing and am delaying it now until I figure out whether it will benefit me. If I see no or very few defects after melting dirty ingots then the value seems diminished. I'll be interested in seeing whether I can meet Kelly's challenge of a leak free tube with dirty ingots as my melt source.

 

I had thought about trying to remove it, but I'd like the metal to be solid first. The first time the glass broke most of the metal ran out. Trying to remove it will likely break it anyway, being unstable with high temperature gradients, but I wouldn't have to spend twenty minutes sweeping up glass if I just broke it in half. I just don't want to splash molten metal as well. It may be an air jet to keep it cool then early removal to get it away from the heat. Interesting.

 

Trying to cast a small pipe (1-1/2"x8" 1/4" wt) I tried to incorporate what I thought I should do to minimize bifilm production. It is a bit of a challenge to implement what you have read about.

First is flow rate. I did several pour videos into ingot molds, then weighed each ingot and measured pouring time on the videos. Out of 28 ingots poured I only had four which were good smooth pours. They varied from 3.2 to 3.9 cubic inches per second.

Then runner flow rate. Maximum rate is dependent on depth of metal. I'm going to use a 3" cope so velocity for a 3" free fall is 48 inches per second (square root of 772 times depth). I have observed this in the glass flask efforts.

Now runner size. 0.2" runner at 48 inches per second flows 1.92 cu in per second. 1/4" runner is 3.01 ips. 7mm is 0.276", flows 3.67 ips. I chose 1/4" to be sure I can keep the system full on filling with my crucible and how I pour. Pouring tests are critical because everyone is different. It doesn't matter how well the system is designed if you can't keep it full. 6mm (.236") would have worked for me too. I was concerned that 8" long with 1/4" wall might freeze before filling if filled slowly like a o.2" runner would do.

Sprue design. The experts say that for a short sprue a straight taper is fine. After noticing my inability to keep a sprue full just below the top it seems it needs to be flared more. So it seemed reasonable to actually calculate the ideal curve. Calculate the velocity for each depth down the sprue and the calculate the area to give 3.01 ips at that velocity. I have found it difficult to cut a tapered sprue accurately (should 3D print it) and even more so for small sizes so I decided to make the sprue 1/4" thick with curved sides. Calculation results:



Laid out on paper



And stuck to a 1/4" board



This sprue was placed on a 1/4" runner feeding a trap riser with gate going to a feed riser in contact with the mold.



I forgot to radius the inside corner at the sprue and runner.

Sprue, gate , and feed riser:



Runner at trap riser



The bad: I had baked a sodium silicate core but broke it trying to cut off the end of the copper tube I used as a reinforcement (left it too long) so made another in a hurry and forgot about baking it and did not vent it. 1" by 9" is a long core for me. The core made vapor and created a bubble inside the casting ID, barely visible in the last picture.

The good: The casting is bad anyway so it's easy to cut it up for inspection. The steam from the core bubbled back out the riser so there will be some suspect areas there.

Try again...

Video if you like.

 

I'll try to get a better picture.

How's this?



There's a little thin area on one side.



It's a little worse than it looks.



Aww, you made me cut up my masterpiece.

 

I use a PVC pipe to make cores, and I slit it down one side.
Once the core has set, I use a screwdriver to pry open the pipe at the center (via a small relief slot) and the core drops right out.
Makes for a very straight and accurate core, and eliminates the two piece core box.

Sometimes I drill the vent hole through the core after the core has been made.

 

Thanks, good idea. I have cut both sides and taped them back together but not tried a slit. I really wanted 1" not 1" pipe ID (1.049") so I made the core box.

Probably splitting hairs.

 

Make the slit bigger so when the hose clamps are tightened you get your desired OD. Done this way they are self releasing.
Core wash helps a little, but direct venting is critical.
What results are you getting as far as a clean casting without unfurled bifilms?

 

That is the drawback of using split PVC as a core box; ie: the diameter of the core is dictated by the internal diameter of the PVC sizes.
For me, I generally want some machining allowance inside the bore, and so it works out pretty well.
For an exact core box size, the traditional corebox method may be the only way.

Making the slit a bit larger can change the core diameter slightly, but also gives more core size distortion which is proportional to the increase in slit size.

.

 

If you need to be that accurate, cut the slit, clamp and then bore to the diameter desired.

 

Good idea, I'll try that.

Core wash helps with core blows? I don't have any core wash, or don't think I do.

These 8"x1" cores keep breaking on me if I don't out reinforcement in them. I'm not good with sodium silicate.

The end of this casting looks pretty good to me with a 5,000 grit sanding. I'm seeing six specks. Biggest one about 6 o'clock, and another at 5. Ring is 1/4" thick. Defects are less than 0.002".



Near the core blow it's filled with little stuff. Smudge at 5:30 is not a defect, I don't think.

 

I just feel that sealing the exterior of the core and adequate venting helps direct the gases out through the center. Relieve the upper half of the core slightly on the end where it rests in the coreprint, and vent vertically through the cope. I use a molasses and graphite mix and bake. That is the only way I have been able to get durable NaSi cores. It also improves the surface finish although your finish looks very nice.

 

So how do the sanded sections look for dirty ingots of extruded material?

 

Dang, they look pretty good from here. You've done some real good work here.

Can you elaborate a bit more? Do smear the mix onto hardened cores, then bake?

 

I make molasses water in a 10 to 1 ratio then I add fine graphite to paint on, or flour to make a paste depending on what is needed. It needs to be baked, but since most of my cores are Lindseed oil bonded the parts just go back into the toaster oven to set and dry.

 

I forgot to add, and remelted material with no degassing...

I'm gaining confidence that the way to get rid of bifilm is to let it float out upon melting if it has any bit of unfurl it has buoyancy even if it were heavier than aluminum, I've heard even steel ships can float.

 

I see you got off dead center.

Are those bob-tailed tubes?

Made some plugs from new castings and bored the pipe ends for the o-rings.



Short one fitting and I was out of time to make a new one. The plated fitting was old and cross threaded, I tried salvaging it but wasn't good enough. In any event I got it up to 30 psi with the fitting leaking and soaped it down. So far so good.



I plan to put water/glycol mixture in it an run it up to 150 psi and see whether it leaks (after it's proven at the specified pressure which I have to ask for again) but it's going in the press for high pressure.