Bifilm Pouring Basin

Discussion in 'Foundry tools and flasks' started by oldironfarmer, Mar 5, 2019.

  1. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    I hadn't mentioned it and not that it has anything to do with the casting defect, but I didn't understand all of the features on your basin design compared to the examples bifilm thread. I don't think the step should fall right into the sprue. As I understand it, there should be a more generous secondary well with a more radiused transition into the sprue. The step basin simulation showed in a fairly compelling way turbulence in the main pouring well (which seems undesirable) but not the portion on the other side of the step above the sprue. Without a radius, it will most certainly guarantee separation/turbulence and aspiration. The large square pouring well, yes, and the correct taper to the sprue....is your sprue tapered? Square versus round sprue....ummm maybe. The abrupt transition from sprue to large square gate....not so much. Did the part turn out fine?

    That first video was pretty interesting, especially that delay to the point where the well flashed with flame. The well looked rather massive compared to the volume of the part. Do you think the delayed flame was just a burp when the metal front reached the more massive section of foam? I would have like to have seen the resulting casting with feed system in tact. Thanks for the video. Try the phone on it's side instead of vertical next time.....bigger/better picture.

    Best,
    Kelly
     
  2. I was being a little facetious on not seeing that coming. Yes, a nice blind riser to feed metal back is a must. At some point several years from now maybe I'll see some of the obvious.

    My excuse here is I was in a hurry to pour and really wanted to prove the "small" sprue. 7/16" square seems fine and I really like dumping metal in a large basin and keeping the sprue flooded from first contact.

    In making the pattern for the basin and offset for the sprue I wanted to start a little large because I didn't think the small sprue would work. Seeing it does I'm ready to shave the transition off.

    At the bottom I just had nothing on my tool shelf to make a gate so I use what I had. I see wide shallow gates (not this wide obviously) but I'm thinking the narrower and deeper I make a gate the closer it gets to the bottom of the pattern, in this case anyway.

    I have no hope of reclaiming a use pouring basin at this point. With my success at green sand that seems to be the quickest and easiest method to get a pouring basin ready for each pour.

    Edit: Forgot to address temperature. I'm not good at measuring temperature. I usually get at least a hot spot measurement of over 1,700F and probably 1,650F average in the crucible before I cut the fire and proceed to pour. I can see that there is lots of metal movement in the hot crucible with temperatures varying 100F. If I just hold the thermocouple in the middle of the crucible I get a false security of accuracy. I think I lose at least 100F from furnace to tipping the crucible, and more in the liquid stream in open air.

    I saw the basin design in Marquardt's paper and really like the concept. From my vantage point once the level gets established in the well there is no aspiration. I think the small well and high surface tension play well together to keep any bifilm on the top surface of the pour (not in it). I may have misinterpreted Marquardt's paper. PatJ posted a link on page 4 of the epic bifilm thread.

    https://www.sfsa.org/doc/2018-4.11 Highland - Marquardt.pdf

    He credits Bob Puhakka and all the other prime suspects in the bifilm theory case so I think he is well tuned in. His basin is the green image on page 20.

    My sprue is not tapered. I thought it interesting the steel casting guys went with square sprues because they didn't have access to large lathes. Regardless, the square design in small sprues should limit falling velocities. I haven't tried a tapered sprue because I only had the straight one made for the lost foam setup. I was able to pull it out of the green sand with a bit of wiggling so I don't need a physical taper and it seems what they say it probably is of no consequence in sprues under 20 inches in length, and I'm at 2".

    I think I'll try a 3/8" round sprue next. I was shooting for 1 ft per second (2/3 of 0.5 m/s) to avoid turbulence. But I really need to make an interrupted pour to see my velocity by gauging how fast the pouring basin empties.

    The picture of the basin laying on top of the bucket shows it, it is a very small foam pattern. I'm guessing maybe the gas could not escape into the sand fast enough. It sure was late in the game. Initially it took some metal, then slowed, then started feeding like normal. Then the flame really was at the end of the pour, I wasn't pouring, just holding a bifilm stream over the basin.

    Video technique noted.
     
    Last edited: Mar 7, 2019
  3. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    It depends on the expert/guru we're talking about but Marquardt admits he has inventor's bias on the rapid prime basin. It is not the style Puhakka appears to prefer or uses in his examples or actual pours in his videos. Apparently Marquardt doesn't subscribe to the need to start with zero velocity enabled by a plug as Puhakka does. Perhaps this is the point of the rapid prime basin.

    Puhakka didn't say you don't need a tapered sprue below 20". He said a straight taper (as opposed to a hyperbolic curve) was a sufficient approximation below 20". We are way below 20" but it doesn't mean that a non-tapered sprue won't induce turbulence and/or aspirate and the tapered sprue and it's interface to the well are part of the system that prevents aspiration. The well layout, shape and geometry is supposed to buoyantly separate biofilms.

    It's more difficult to get laminar flow in a square pipe than a round one. The areas in the vicinity of the sharp corners are the culprits. For our velocities some rounding of the sharp edges on a square sprue probably takes care of that......and they are easy to make....even tapered. A couple of seconds on a belt sander will do. And BTW, I like my reason a lot better than Marquardt's, his being I don't have a lathe and it wastes material. Really? This whole thing is about Bifilm theory and you'd compromise a design over something like that?

    I don't think the shape of the cross section of the sprue limits falling velocities. Cross sectional area, poetntially yes. But the whole point of the Puhakka discussion was to achieve a sprue that would approximate the shape of molten metal falling unconstrained, sort of like the equation to a streamline and Bernoulli discussion. A falling liquid stream doesn't form a square sprue it forms a round tapered sprue. The acceleration of gravity will dictate the velocity at any given depth. Conservation of energy, the height of the well and sprue, then dictate size the area of the sprue openings at each end (neglecting friction losses which are small for us).

    Here is what I have done to estimate velocities on past pours: Video tape your pour. You can easily capture the duration of the pour in seconds by doing so. Weigh the resulting casting complete with feeder system. Divide the weight by the duration and that's your average mass flow rate. If you first divide the weight by the density of aluminum (about .1lb/in3), that will give you the volume of aluminum poured. If you divide that by the duration of the pour you now have volumetric flow rate (say in in3/sec). If you divide that by the cross section at any point in your feed system, that will be velocity in in/sec.

    I like application of theory and reducing it to practice. It's fundamental to engineering and a fun intellectual exercise. We need to keep in mind we're hobbyists pouring metal at home in our garages. From my view, if the intent is to apply theory to gain some improvements, then apply it all not just some......but that's just me.

    Best,
    Kelly
     
  4. Great points.

    I'm certainly not interested in a plug. I have never poured without sucking air in with the metal until I tried this pouring basin. All the rest were too shallow to maintain a level and prevent a vortex. The little offset basin does prime rapidly, then the next challenge is to establish enough head to allow some bubbles and mixed bifilm to float. I suspect the two inch basin has greatly different fluid dynamics than a ten inch one (I have no idea how big they are using). The aluminum seems quite viscous in the small one. Good for avoiding entrainment, bad for floating out contaminants.
    I confess I have not had the time to listen to all of Bob's videos. I did misinterpret the tapered/hyperbolic difference.
    I thought the lathe thing was a hoot. However the sprues and runners they all show seem to be square cornered. not just Wade's. My presumption is the shape is not a big deal. While it is more difficult to get laminar flow in a square pipe, it's not much more. If you are well under the turbulent range (which I think I am) the larger wetted perimeter of square helps with reducing flow rate and avoiding turbulence.
    The shape of a falling stream is also not round due to surface tension and turbulence. A small diameter stream of water will be round until the increasing velocity and surface tension breaks it into droplets. I do believe that we have a high friction loss due to combination of viscosity and surface tension. When you get at or under the seissle (sp) dimension I expect it to be dramatic, like water in a hypodermic needle. So my concept for small sprues is that it will not accelerate and actually be held back by the surface tension and viscosity, making a system that will not pull into like water droplets and injest air in the empty sprue below.
    Your method will approximate the overall mass flow rate and hence average velocity. I'd like to estimate the maximum flow velocity with a mostly empty mold. We know the available head will fall dramatically as the mold cavity fills.
    I have always been quite amazed that theory works in application and find it to be very exciting, especially when it is not intuitive. Even if I can't apply all the bifilm guidelines easily, I want to apply the ones I can. The pouring basin which prevent air entrainment from pouring is such a big improvement for me I will use it even if I'm not avoiding other bifilm inclusions in my castings. I had a successful engineering career following the premise "do something, even if it's wrong". Of course that does not mean do the wrong things, it means don't be paralyzed by analysis or scared to act because you can't tell which ok method is the best. When talking to competent engineers, I often had to say "do something, even if it's wrong" as there were usually no really wrong alternatives, just less desirable ones. And progress is better than perfection in my world. Else my long reach clamp would not be holding veneer at this moment.

    Great interaction on this forum. I appreciate your input.
     
  5. Tobho Mott

    Tobho Mott Gold Banner Member

    FWIW on round vs rectangular sprues: In the SWdweeb video about sprue design, Puhakka mentioned in a reply to one commenter that "square/ rectangular is the best", then went on to say he'd explain why in the following gating video. He did not, but this is convenient for me since I also have no lathe. :D

    Jeff
     
  6. ESC

    ESC Silver Banner Member

    I caught the comment about not being able to cut the round tapered sprue because of a lack of a lathe/equipment. As with the skillet mold that Jammer posted, the rectangular sprue is easier to make and can also be molded in one side of a vertical split flask using a match plate .
    Cambell's preferred design has the rectangular basin with the radius of the dam equal to the radius of the sprue and the front wall aligned with the sprue. He shows a Puhakka preferred design which is called a reverse delta that uses a rectangular sprue, longer dam, and smaller basin.
    Another note from Campbell, he mentions that the lost foam casting method is accused of being launched onto an unsuspecting foundry industry before it had been developed into a reliable process. With that in mind, the offset basin and small sprues and runners were not advocated for lost foam, but rather conventional lost air.
    I have another cast iron lost foam in the works, and he does mention problems with the additional heat of the iron pours. I'll need to spend more time on that to see if I can cut down on the folds I have gotten in my samples.
     
  7. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    I don't think round versus square is as big of a deal as getting the area right. They just didn't fit as nicely into the theoretical aspects in the example. Round the corners a bit and use them if they are easily adapted. -I will.

    Can you post or send the excerpt to me ESC?

    So is that to say he doesn't think they are applicable to lost foam at all or just not tested and optimized for such? I would think once the foam is out of the feed systems, most principles apply, except velocity may not be controlled by the feed system geometry but potentially by pattern evaporation rate. In any case, that would result in lower velocities than that designed for an open cavity system......but you may not be able to feed/fill a lost foam casting of the same mass.

    I took great interest in your last one and the use of mud as a coating. My understanding is LF iron needs high refractory and permeability coating because of the higher heat and feed pressure from the denser metals. Without, you have a lot of gas with nowhere to go. Vacuum might help that cause. I don't seem to need it with aluminum, I presume because it just doesn't vaporize foam at the rate hotter denser metals do.

    Best,
    Kelly
     
  8. JCSalomon

    JCSalomon Copper

    He kinda did—you know that thing where your explanation makes sense, but only to people who already understand? He fell into that trap.

    Consider any time you need to split the flow. With a rectangular cross-section, just split the runner in two, keeping the same height and dividing the widths according to which fraction of the flow each division needs. Minimal turbulence created. With a round cross-section, you first need to carefully calculate the two cross-sections, and then design a careful split which… never mind: no matter how careful you are, you’re going to add a lot of turbulence.

    Now rewatch the gating video from about 14:30, and realize he made all those points with his traffic analogy.
     
    Last edited: Mar 7, 2019
    Tobho Mott likes this.
  9. ESC

    ESC Silver Banner Member

    I take it that they were just not tested. The book dates from 2011, so almost 10 years old. As seen on utube the engine castings use a large conical pouring basin and the ladle pours a stream that looks to be a 3 or 4 inch oval stream of aluminum very quickly. In addition those sprues look to be straight and are strong enough for the robotics to handle them, so they don't appear to be hollow. From that I think the filling velocity is controlled as you mention by the rate of foam evaportion and the head pressure from the height of the pouring basin.

    This is the latest foam. I gave it one coat of mud and pierced the coating with a pin. I had poured the ball turner base horizontally with no coating and ended up with a good lower surface and folds on top. This is my preferred orientation for aluminum, but for cast iron we shall see. I'm with Oldironfarmer in that regard, just do something.

    IMG_20190305_162833.jpg
     
    Last edited: Mar 7, 2019
  10. One more step. I decided to try to make the curved transition between sprue and runner/gate. I guess its a gate as it goes into the mold but it's a runner because it doesn't choke the flow. Wound up with this.

    IMG_5186.JPG

    A three piece

    IMG_5188.JPG

    Upper two in the cope and lower is the runner. Or gate. Instead of trying to dowel them I decided to hold in place and pack sand with the other hand.

    IMG_5190.JPG

    It helps when the pattern maker is the molder. No complaints.

    To avoid the shrinkage problem I decided to add an open riser on top of the boss.



    OK, that was a mistake. And I'm pouring too hot. Way too hot. But look at that good surface finish.

    IMG_5199.JPG

    Still had shrinkage in the boss from pouring too hot, I believe. This is such a simple pattern I'm going to pour it tomorrow at about 1,300F without a riser.

    So fun facts, from the video you can't see whether the sprue was ever full. From watching it I think it was but I never poured fast enough to establish a level. I think cooler metal and a closed system will help that, but the numbers don't look too bad.

    Final casting as pictured weighs 11.8 oz. From the time metal entered the sprue until it exited the riser was about 3.3 seconds, from 51.7 to 55.0. Average pouring rate was 3.58 oz per second, or 0.22#/sec, which equates to 2.6 cubic inches per second at 0.086#/cubic inch. My sprue/runner/gate is actually 0.17 sq in so that means the average fluid velocity was 1.3 ft per second, or 0.39 m/s. That should be under the 0.5 m/s limit described in the literature.

    The curved transition worked OK, but the flashing due to lost sand surely created some turbulence at that interface. I want to be able to establish some level in the basin to be comfortable with the velocity calculation.

    The gate showed a little porosity, and the riser as well. One would expect some in the riser, I suppose. Top of picture is up, tit on left is flashing at parting line.

    IMG_5200A.jpg


    IMG_5201A.jpg

    I guess I'd better get my good camera out if I'm going to try to take macro pictures.
     
    Tobho Mott likes this.
  11. Tobho Mott

    Tobho Mott Gold Banner Member

    I don't know if being easier to split into two runners that add up to the same cross section is THE reason why square sprues are better than round that he promised to explain, but I can definitely agree that it is A reason... For me not having a lathe is another pretty good reason. :D

    I like the 3 piece sprue-to-runner transition setup idea, nice work. With the sprue plug mechanism also in progress, seems like you're just about all set up to go full Puhakka!

    Jeff
     
  12. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    I'm taking notes Andy. A couple random thoughts on this part.
    • Lower temp may/will help. Any idea on the alloy? Was it recycled casting or wrought? When it comes to pour temp and shrink, that can have a significant impact too.
    • Don't think your riser is big enough. There are rules of thumb but the riser needs to be more massive than the part it's feeding and in this case more massive than the rest of the valve body. A lot of times such features are positioned on the parting line so the riser can be parked adjacent to it. In current orientation you could dowel a larger (open) riser to the boss and pull it from the top of the mold.
    • Or, go the other direction. Is the boss big enough you could machine a conical depression into the pattern to reduce some mass?
    • Or, since some shrink and some don't, just take your yield rate.....how many burners are you going to make?
    For my selfish interests, I'd like to see more video'd lost foam offset basin Bifilm pours. I might be coming around on the rapid prime configurations. Sort of interested to see if the initial puke from foam vapor can be eliminated with a foam sprue or even with a hollow sprue. If not, there's not much sense in worrying about a plug.

    Best,
    Kelly
     
  13. With slow flow rates I'm not sure there is really much difference between square and round. I really need to calculate the Reynold's number of the runner flow. Square (or trapezoidal, I had no draft on this first go around) certainly is much easier to deal with.

    I realized last night I could have just cut the parting line down, duh! I'll do that to make life easier and for this part go all the way to make it closer to be bottom feeding.

    I'm going to do a test pour today realizing my average flow rate may have been far from the maximum flow rate as the pattern filled and reduced available head. My plan is to just run it out into an ingot mold to get a more accurate assessment of velocity as the metal front is approaching the pattern. That's where it can be turbulent and entraining air, making oxides. I'm grasping better why I need a tapered sprue as well, the sprue needs metal to feed the increased velocity as it's falling down the sprue initially, to limit mixing there.

    That wasn't me working on a plug, but you've challenged me. I can see lifting a plug to help drop the maximum fluid front initially. It may be more challenging with a green sand mold, and at this point I'm thinking a square tapered ceramic plug which will just compress the green sand a bit. With lots of parting compound. Now the trick is the lifting mechanism. I'm not convinced a foot pedal is out of line but I've also thought about a counterbalance assisted by bouyancy. Once it tries to float the counterbalance takes over and pulls it away. A wooden plug may work as well. Wood floats, doesn't it? Like small rocks...

    Lower temperature and no riser is going to be tested. Of course I'm melting unknown extruded crap. Scrap with no S is crap, no? I keep my cast stuff separate. My goal is to learn on a standard transmission then go to an automatic. I mean, learn on extruded then cast should be an improvement.

    Thanks for the advice on risers. I had not thought of putting a big riser on top of the pattern while filling.

    I had thought about putting in a core but because it's a burner manifold don't want a piece of stray sand plugging the nozzle. Probably an unnecessary worry.

    Well, for what I''m doing now, the burner manifold pattern is just convenient. Even the shrunken parts will make good manifolds. I think I can sell several manifolds.

    I will do some more lost foam. Interestingly the lost foam provides a plug. I've also thought about using a short piece of foam as a fusible plug instead of lifting a plug. I don't know what the foam would do in the way of making oxides we're trying to avoid. It may actually be of advantage to fill the sprue with products of combustion instead of the dreaded air.

    For that matter, an aluminum plug may work very well, there's plenty of mass in the basin to heat up a cold plug and away we go. 1/4 second delay is plenty to let you flood the basin.

    I guess I've got an opportunity to try other stuff...
     
  14. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    Buoyancy sure would be a nice mechanism. Thing is, buoyancy may provide a slow/gradual lift. A small seat with an extended bobber may do the trick. If the well fills quickly it wont matter, but better keep it full because if it re-seats or partially obstructs.......interrupted pour. I think I have the hot material for aluminum duty with that moldable ceramic fiber.....feather light, completely non-wetting, and suitable for metal contact.....guaranteed to float! I bet wood would work just fine too....once!

    Yah, I just don't know if you'll ever get away from that initial burp with foam in the sprue, and maybe not even without a completely hollow sprue and feed system.......sort that out for me would ya?

    Best,
    Kelly
     
  15. ESC

    ESC Silver Banner Member

    The volume of the riser at the top of the cope needs to be greater than the volume of the offset basin above the dam so you can snap back after you first see metal in the bottom of the riser. Somewhere Bob mentions that there needs to be a place for the reservoir to drain to.
    As Kelly mentioned a larger riser, but since that metal is post casting and bottom fill there would be nothing wrong with just scooping out a cavity on the surface of the cope. It could even be offset to clear the basin.
    I like the sprue /runner and might try something like that myself.
     
  16. The key to buoyancy is a counterweight which is lifting more than the plug with the plug gently jammed into the opening. Once it moves it will then lift and stay up. I think you're stuff is just the icing on the cake. Primarily by being non-wettable.
    Burp may have been from the tiny sprue and not enough area to absorb gas. I'm thinking I'll try a big sprue.

    I like having some place for the stuff to go. Pretty basic. If you saw the video it was too fast for me to react, and I thought I was watching. But I was trying to establish a level. So a big basin on the surface would be great to add to reaction time. My next pour for this pattern is going to be closed riser and cooler. I think the internal air pressure will help slow the flow.

    I think I improved my sprue runner joint this morning. That design worked well but the tail of sand on the cope is very fragile and I had trouble controlling the breakage at the edge of the sand so the metal was seeing a gap. Adjusting the parting line will help that a lot.
     
  17. Tobho Mott

    Tobho Mott Gold Banner Member

    Oops! :oops:
    Case of mistaken identity. Apologies to melterskelter, the actual guy who's working on this forum's first family friendly plug.

    Jeff
     
  18. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    One thing I noticed in that video is the basin doesn't appear to fill very deep. Do you think it still acts like a Bifilm separator at that depth? A plug would obvious fix that but not sure it acts like a Bifilm design when there is no flow either. 4 seconds is a short fill time. Since the velocities are already low, you could probably further reduce the sprue cross section. It's a pretty small part. What's the volume of the well compared to the volume of the part? When parts get this small, there are some practical limitations because the well size still needs to be a sufficient size target.

    I'm thinking I'll make my first well about the same volume of a 2 1/2" diameter x 4" tall kush cup because that has proven about right as far being able to respond to changes in fill rate. Experience says I'll need a bigger one around 8 lbs.

    Best,
    Kelly
     
  19. Certainly no harm done, and it has made me think about going that way.

    Of course it is of no value for separating bifilm until a decent level is established. However I chose what I thought would be a three second volume per Wade while Bob recommended a 1 second volume. For small sizes I can't see being too big as a major issue. The height of the weir is what would primarily assist in separating floating bifilm so long as you have sufficient level over the weir. The well is about twice this part volume.

    Today I made a tapered sprue and new runner with draft. That required a modification of the basin to a rectangular opening.

    IMG_5206.JPG

    I buried the runner in the drag.

    IMG_5207.JPG

    Hmmm, the pattern looks a little close to the flask? This can't come out good.

    I cut the parting line down to fit.

    IMG_5209.JPG

    "She said 'I've never really done this kind of thing before, have you?' and I said, 'Yes I have, but only a time or two' ". I need more practice cutting parting lines.

    Purpose of this mold was to calculate the maximum velocity obtained with no backpressure in the mold, and no metal head in the mold.

    I was concerned about how fast the hot metal might exit the mold with the snap flask removed, so I put up a firebrick wall and stood on firebrick in case the metal ran toward me. Much ado about nothing.



    3 ft per second, 0.9 m/s, pretty fast. But the goal is less than 0.5 m/s going into an empty mold. I guess I'll do another test like this with a smaller runner, maybe 1/2 this one. It seemed very tiny at first, but it is looking bigger now.
     
  20. Like I said I would, I poured another burner manifold with the small runner system. I can make good manifolds with a large air-aspirating system. I would like to be able to try keeping a runner system full instead of just splashing it down a giant sprue.

    Since I poured the last manifold too hot, with an open riser without a basin, I wanted to see what a closed mold with a small runner would look like, and try to assess runner velocity. I did a little better at cutting the parting line down, but didn't quite get a clean square clean corner. I'm learning but slowly.

    IMG_5210.JPG

    Video mania. At least I was able to establish some liquid level.



    Average velocity was 1.41 ft per second, .43 m/s, so the maximum velocity was likely over 0.5 m/s.

    Sprue and runner are a bit ugly but we're getting there.

    IMG_5213.JPG

    But there's a spectacular shrink hole. This will be the last one without a riser.

    IMG_5214.JPG

    That hole is about an inch deep.:eek: But look at that nice surface finish.:rolleyes:
     
    Last edited: Mar 8, 2019

Share This Page