Bob Puhakka on Bifilm theory

PatJ said”
An alternative to the pour basin and plug arrangement is what John Campbell mentions in his book, which is a quiescent pour, where the sprue is eliminated, and the lip of the crucible is put in contact with the opening in the cope, and there is no drop down a sprue.
This assumes you can fill a mold without using the height of a sprue, but I think ironsides uses this method with great success”

I sure wish I could do that. For plaques etc would be great. Unfortunately a majority of my pour goes into the cope.

Denis

 

Pat: I think Al was saying people who ignore all these items might not have perfect castings, not Bob.

 

When I first found out about bifilms, I thought that the pour basin and sprue should be eliminated since it has the potential to cause the most trouble, and indeed in John Campbell's book there are several details of how to do just that.
But for day-in day-out backyard castings, and many industrial situations too, we are stuck with the pour basin and sprue arrangement.
But with a stopper, I think the problems can be pretty much eliminated, and even without a stopper, I think with careful pouring we can get good results.
I think a consistent pour that keeps the basin full and the sprue also full is critical to eliminating air aspiration into the melt.

 

Could be I misinterpreted what he said, but I think it is still interesting to me to see just how good castings can be when using the 10 rules.

I mean why not use the 10 rules in a hobby setting?
Many of the rules just involve simple placement/arrangement of things like risers, pour speed, and such, and could easily be incorporated into what we do on a hobby basis.

.

 

[QUOTE="PatJ, post: 13314, member: 17]"<Snip>
I think a consistent pour that keeps the basin full and the sprue also full is critical to eliminating air aspiration into the melt.[/QUOTE]

I agree and I think it is very difficult to cleanly and fully fill the sprue and keep it full right from the getgo. Most of the time imperfect technique in that regard does not result in a fatal casting flaw. But way too often it does and much time and energy is down the tube so to speak.

That is why I am keen to make the plug work. BTW, I would use a hunk of graphite for the plug if I had one handy as Pat does. But, I think sand and SodSilicate will work fine and is very quick and easy to modify as needed and to fabricate. I think I have a very very simple plan for triggering it and lifting it. I’m anxious to get out of airports and into my shop to put ideas into tangible form. Should be in the shop tomorrow!

Denis

 

Made mold for an offset pouring basin just to try it. I'll probably make one with sodium silicate for the first test. This is based on Campbells conventional rectangular. Tobo's looks like the slimmed shape on the same page.



The test with molding sand. I just barely pulled with the sharp corners so I installed fillets and shellaced them.



The ten rules are not necessarily new. Choking the sprue, using a good quality melt, avoiding core blow and avoiding shrink damage are all lessons I learned from reading and hard knocks. The new rules have to do mostly with entrainment and bubble damage. I'll change the way I sprue and runner my molds just to eliminate some of those defects and see if there is any noticeable improvement.

 

Have you considered trying to let the plug float when it gets enough liquid around it?

 

You bet, but I am concerned about 3 probable problems:
1) there is a good chance that while pouring my flow rate could slow down enough at some point for the plug to get sucked down against the sprue opening obstructing flow temporarily until it floated off again. Thus the very problem I was trying to avoid would be exacerbated.
2) the idea of the plug swirling bumping around carried by currents in the basin would make me worry about dislodging green sand chunks.
3) missing the moving plug with the metal stream would be tough as it moved around (there are strong currents in the basin during a pour) and hitting it would cause splashing.

If my planned teeter totter method works it should avoid those problems. You will see that the heat-triggered release is quite simple and will/should prevent the problems listed above.

Denis

 

I just skimmed through some of the more critical aspects of John Campbell's casting book, and the one thing that is certain is that until about 20 years ago (+ -), the methods used for casting metal were about as accurate as assuming that the earth was flat, or perhaps assuming that the universe revolved around the earth.
The last 20 years of research has pretty much proven that many/most methods developed over the last 2,000 years are bad methods that result in often defective castings.
Its a rather shocking revelation, but it appears to have been verified by both radiographic methods, computer programs, and actual test castings.

The reason that the bifilms have not been noticed before is that they are very thin, perhaps 1,000 times thinner than a filter that they may be precipitated on, and so even under a scanning electron microscope they are difficult to spot since they wrap themselves around the filter structures.

But bifilms are very real, and cause very big problems in castings.

Flow rate and velocities have to be closely controlled, as does air aspiration and turbulence.

I can safely say that everything I have ever read about sprues/runners/gates/risers before this time is wrong from the standpoint of what has been proven as modern casting techinques.

John points out in his book that the requirements for good castings has increased, and people now xray casting and use other methods to test them, and thus what was previously accepted as good castings are now revealed to be rather weak and defective under rigorous testing, and much/most of the problem is due to bifilms in the metal, which are back-to-back layers of oxides that do not bind to each other, and thus create weak spots that form cracks.
The more bifilm you churn into the metal, the weaker the casting.

Top fill systems are very bad.
Horn gates systems are also very bad.
Basically every system you ever heard of is bad.

Can you use the old systems to make great hobby-grade castings?
Yes you can, and you can be happy about that the rest of your life.

Do you want to make 21st Century castings that are generally free of defects and high in strength?
Follow the 10 rules, and arrange your pour basin/sprue/runners/gates/risers as John recommends.
No guarantees, but at least if you are lucky you can eliminate the known causes of casting failures/weaknesses, and you should be much closer to success if you are having problems.

I am aware of some folks that do a lot of iron work and are having serious problems, so this is beyond theoretical chat just for fun.
For me, time is money, so if I can make a solid defect-free casting first time, every time, then I have saved a lot of time/money, not to mention frustration.
Making a mold for the second, third, forth, or however many times can be very discouraging and frustrating.

Getting it right the first time, and having consistent repeat success is very exciting, and really makes it all worthwhile.

For the pour basin, it seems like the very latest design has the area around the sprue opening necked down in width to not much bigger than the sprue opening, and with no sharp lips at the junction of the basin and the sprue opening.

.

 

Nailed it.

Came across another article of interest. It was clearly still a work in progress at the time of writing, but the fellow (Joe Plunger) also clearly feels he's heading in the right direction. There are some examples of successes on the last few pages attributed to the changes in gating design.

https://www.sfsa.org/doc/2017-3.4 MMP - Plunger.pdf

Al

 

That is an interesting white paper, and very relevant.
One of his comments in the paper is that the client was happy because he did not have to repair the castings after machining them, as he had to do with his typical source of castings.
That sort of reinforces the fact that existing foundries produce a lot of bad castings, and people in the past have just accepted that as standard practice, and assumed that poor quality castings were the best that could be made.

So I guess everybody is asking the obvious question "What can I immediately adopt that will potentially improve my castings?".

Like ESC above, I am going to adopt the pour basin, but make the width at the sprue opening narrow as shown in the above white paper.

For the sprue, a hyperbolic shape, probably square in section, not sure what the exact taper will be, and not sure the top and bottom dimensions.

The bottom of the sprue will taper on both sides to transition into I guess a flat runner.

I will probably use the swirl expansion chamber(s) with vents out the top of the cope.

I am not sure about using a vortex slot gate since it may require a filter to be effective, and I am not sure if the round ceramic sponge filters I have will work.

Much of the discussion seems to center around pouring metal down a tall sprue, which speeds it up, and then using all sorts of methods to slow it down again and prevent it from moving so fast that it causes turbulence.
My immediate thought was to eliminate the tall sprue (as ironsides often does; maybe he is ahead of the curve), and pour at about the top of the cope level, thus limiting metal velocity from the start.

I think I will try the pour basin first without a stopper, and if that produces good results then no sense using a stopper.
I don't want to introduce methods that will not be effective for a typical small-scale mold setup.

As far as the sprue/runner size, Bob mentions you should pour as fast as possible.
Translated, what that basically means is that you should keep the pour basin full during the pour.
The pour rate is not determined by the person pouring, but rather by the sprue, runner and gate dimensions, so the statement "pour as fast as possible" can possibly be misunderstood, since it would seem to indicate you have some control over the metal velocity in the sprue/runner/gates due to how fast you pour metal out of the crucible.

Obviously for parts with risers, you need the pour basin and sprue to extend a certain distance above the top of the risers, else they will not fill.

If the sprue/runner/gating is too small, then the mold will fill too slowly, and you may get cold joints where metal fronts meet, and also may get some temperature differentials in the mold and delivery system that are not desirable.

John Campbell recommends tapered runners so that metal is delivered evenly to the mold cavity, but the white paper above seems to use a flat non-tapered runner, perhaps because one runner feeds only one gate?

The other thing that seems obvious is that any arrangement using John Cambell's and Bob Puhakka's methods may and almost certainly would have to be "tweeked" (adjusted), and so another question arises, does every part need a specifically designed sprue/runner/gate system dedicated to that part, or can more generic and modular designs work over a range of casting sizes, such as 10-20 pounds in iron?

One thing that sticks out in John's book is the advice he repeats over and over, which is "don't oversize the sprue/runner/gate system", because that system is used to limit metal velocity, and excessive metal velocity causes turbulence and entrains both air and bifilms into the melt.

.

 

Interesting analysis.

I agree on your assessment to "pour as fast as possible" means simply keep the system full.

What this all means to me is I need to keep making sprues and runners smaller until I find they're too small.

So, blind risers at high spots?

 

Here’s my sketch of Puhakka’s “top gate”, in a form suitable for hobbyists.

The “surge cylinder” is a relatively simple two-part form—the “spinner disc” (the same height as the runner) in the drag along with the runner, and the “surge cylinder” proper in the cope.

Above the runner is the “top gate”. The void for the “bubble trap” and for where the filter goes can be a fairly simple form in the cope; the gate continues past the filter and enters the mold cavity from the side at the bottom. (The “top gate” proper and it’s more complicated sibling the “trident gate” fill the cavity from the bottom, but that requires the trident gate to be formed as a separate core and be placed in a cheek between cope and drag. Bob says this simplified version is plenty sufficient at the small scale.)

What I’m hoping Bob will explain in his swdweeb webinar is how to compute the size of spinner disc, surge cylinder, and bubble trap needed.

The runners and gates are rectangular in cross-section so splitting the flow can be done without introducing turbulence. They can probably best be created with forms in the drag and cope. Some things I’ve read suggest their height should not exceed the 0.4″ sessile drop height. Runner cross-section should be chosen to match the flow-rate at sprue-bottom velocity; gate cross-section should be much wider to slow the flow down to below the “critical velocity”.

 

Or calculate, perhaps erring on the small side but not by too much.

Vents, rather than risers, is my reading—or vented risers.

 

X2

That was my take on things from the Sprue design video. Pretty much says that flat out near the beginning and again after ~13:47 mark....or maybe just how I keep "hearing" it....

@JCSalomon From my post on pg 6, he never mentioned a bubble trap though.

6) From some comments Bob made re step gates (video since taken down, but mentioned/pictured in Wades article)...cross sectional area of the step is equal to the runner cross sectional area. Surge chamber is 1Q with equal diameter and height. Bob said the slot gate was sized to the casting, but I did not understand and did not follow up. Mea Culpa. It makes sense to me that once the surge chamber is full, the metal velocity into the casting is going to be the same as the runner velocity (ie too high) unless we size the slot gate with an eye to reducing that velocity to sessile drop height velocity.

Al

 

Exactly: if the runner velocity is (to use his example) 96 in/s and we want to reduce this to the critical velocity of 0.5 m/s = 20 in/s, the gate should have cross-section area just under 5× that of the runner. I tried to show that in my sketch, with the gate much wider than the runner.

 

Here is a paper by Wade Marquardt, and he apparently worked at one time with Bob Puhakka.
https://www.sfsa.org/doc/2018-4.11 Highland - Marquardt.pdf

The pour basin that I would propose to use is like his "rapid prime offset basin" shown on page 20.

I am thinking the impeller casting arrangement shown on page 29 is probably all that I will use on my parts.
I just don't see using a vortex step ingate on a 10-20 lb casting.
A vortex step ingate may be good for a 1,000 - 10,000 lb casting, but I think it may be overkill for a 10-20 lb casting.

I would use risers on most parts, and before I was planning on using blind risers, but after reading about air expansion and how that can push the metal back down the riser, I will use open-top risers.

I want to keep it as simple as possible, and I don't want to add features that I don't need.

So the plan is to try it like that shown on page 29 for the impeller casting, with no filters, and assuming any air and trash will go into the spin trap.
If that does not work well, then go from there, but I think it would work and is simple enough to implement quickly.

 

Is there enough information on the geometry? I’m not seeing it in that paper, but perhaps elsewhere?

The vortex ingate is deprecated—the folks involved (Campbell, Puhakka, Marquardt, Plunger) all say there are problems with it sufficient to avoid it entirely. The system they recommend for large castings is the “trident gate” and for smaller ones, the “top gate”.

Marquardt p. 29 shows a “top gate”, same as my illustration a few posts back. I think my sketch is a bit clearer than Marquardt’s diagram but to quote the paper, “I will admit that I have creator’s bias”.

No idea how it’ll work without a filter, and I’m curious to find out—please report back!

 

Marquardt's paper is very good. His conclusion is worth reading twice. The concept of 3 seconds of material in the basin, and the design of the basin to neck down into the area above the sprue are critical. His formulae are not well defined, I'm going to work on that. I think the information is there, just not clearly presented.

 

Could perhaps use the following to generate some approximate ratios as a starting point....

https://www.researchgate.net/figure...c-view-b-front-view-and-c-top_fig13_315896027

Al