Confessions of a Lost Foam Caster, 5 Years on.

Discussion in 'Lost foam casting' started by Al2O3, Apr 2, 2021.

  1. Al2O3

    Al2O3 Administrator Staff Member

    I was looking back on the thread below and the original version I posted at Alloy Avenue 5+ years ago. This part was actually what first drew my interest in lost foam casting.

    This is a lengthy post with lots of links to previous threads about process details. Hopefully you can use those to take in as much as you like and find it useful as a would-be lost foam caster.

    There’s been a lot of water under the bridge since that post. I still make versions of that water neck for folks that restore and campaign vintage race cars. There are more than ten variants now. Just thought I’d refer to this thread because looking back over my shoulder, in addition to where it all began for me, it really displays the versatility of lost foam casting. Since then I’ve made numerous process and material improvements summarized in this post.

    By way of example, when someone asks for a custom version of that part, it usually just involves a different location for the two snouts that interface with the intake. As long as they know dimensionally what they need, making the custom part is really not much more effort because it’s just cutting and piecing the tubular foam pieces together differently. That usually involves just an additional 30 minutes to make a crude wood assembly fixture.

    Here’s an example where the customer wanted a version without a thermostat, sent me a picture, and said, make it like this Cobra neck, only to fit “this” intake………The “Cobra” neck is a commercial part that was tooled and offered by an aftermarket company but the center lines of the snouts don’t line up with the desired intake manifold. I happened to have one on the shelf for many years.

    Like earlier in the thread, I make tubular elbows as the base building block.

    1 Foam Tubes.JPG

    Then cut, piece, fit, and assemble them as desired. -That’s it.

    2 Foam Piece Parts.JPG 2.1 MP C 2.JPG

    Gating: I almost exclusively top gate my lost foam castings.

    3 Top Gate.jpg

    I would never do that in conventional open cavity, gravity fed, sand casting because the molten metal would just hopelessly and turbulently tumble into the mold. But in lost foam casting, the evaporation rate of the foam controls the metal velocity and it is always way below the desired velocities for sand casting. Also, the gating rules of thumb for conventional sand casting really don’t apply for the same reasons and most of the time, the size of the sprue is more a matter of it being strong enough to support and process the pattern as opposed to merely feeding it.

    There are a couple other beneficial outcomes of top gating in LF. As the metal advances deeper into the mold, the hydrostatic sprue pressure increases helping the colder metal fill the mold extremities. With a bottom fed casting the coldest metal at the top of the mold would have the least head pressure. In a similar vein, it very nicely promotes directional solidification from bottom to top of the part. Despite significant differences in cross section, I almost never see shrink defects that surely would occur in same part if conventionally cast.

    For larger thin cross section parts extending the contact area of the gating the depth of the mold with a more generous runner cross section is helpful, like at the bottom of this post;

    ....and here:

    Compared to back then, these days, I enjoy essentially 100% yield and pressure tight castings. The contributors to the improved success rate are process and material improvements:

    Casting Stock: I cast with virgin ingot instead of scrap and many times remelted material and sprue stock. My sources of automotive castings started asking for $1/lb and that was the end of that. I still use scrap but just not on more critical parts. I won’t go into metal management here since it is the same for all processes, but since going to virgin aluminum ingot, I never flux or degass. I melt exclusively in a resistive electric furnace.

    Foam: I use extruded polystyrene (XPS or EXPS) as opposed to expanded poly styrene (EPS, the white beaded stuff), but I do like EPS for sprues and gates. The reason is XPS machines and finishes well. You can commonly buy XPS by the 4ft x 8ft sheet at the big box stores. The pink stuff is Owens Corning and the blue Dow. The pink OC product is most readily available for me so I use it. Their trade name is Foamular with the Pink Panther logo. There are a handful of densities but the two most common are Foamular 150 (1.3 lb/ft3) and 250 (1.55 lbs/ft3). I mostly use the 150 because it is lower density but will occasionally use the 250. It machines and finishes slightly better and is marginally stronger, at the expense of having more material to evaporate in the casting process. The EPS I use for sprue stock is anywhere from .5-1 lb/ft3. Don’t use Urethane foams! These are thermosetting foams like the expandable insulation in the aerosol cans and expandable packing foams. Polystyrenes are thermoplastic. The problem with thermosets, is they don’t melt like thermoplastics. They need to combine with oxygen, then burn, and produce residues and toxic (more toxic than polystyrene) gases. The two-phase transition (solid to liquid, liquid to gas) of Polystyrenes is important to the lost foam process because the liquid and then gaseous boundary insulates the foam pattern from the metal front while the liquid evaporates more controllably into gas at the metal front and is expelled through the coating into the mold media.

    Vacuum Assist: My molding rig(s) has provisions for vacuum assist. I can pull ½ atm of vacuum.

    4 Pouring Cup Vacuum.jpg

    The short answer is I find it is rarely needed in LF aluminum casting except in extreme circumstances. Like here:–-the-lost-foam-edition.1020/

    It can also provide improved mold stability but can produce as many problems as it solves. It is more important in high melt point metals where the gas evolves and needs to be expelled more rapidly.

    Pattern Detailing: For fillets, appliques, and lettering, I use a specially formulated wax for lost foam casting. I use it very sparingly. I always had pattern maker’s fillet wax available. This was very good. The specially formulated wax is very similar with a slightly lower melting point, supposedly the same melting point as polystyrene. It’s easier to use and seems less likely to produce defects. I always prefer to machine a fillet from foam rather than lay wax fillet after the fact but sometimes it’s just not practical. I won’t typically use a wax fillet larger than .25” radius. The use of ball fillet tools and scrapers is essential to get a good result. Heat the ball and scraper tool with a hot air gun or alcohol lamp. I can’t make these waxes burn. If you put a torch to them, they just melt and evaporate.

    5 Ball and Scraper.jpg

    I have tried all kinds of paraffins and waxes. Many have used wax seals for toilets. The latter is very low melt point stuff. For a while, I used warmed wax in a plastic syringe. If you buy one of those inductive coffee cup heaters, fill a cup with water, and just return the syringe to the water bath, you can dial a temperature that easily allows the wax to be extruded out the tip, applied, and then smoothed/formed with a ball fillet tool. This will work for most any wax.

    Glues: I use low (the lowest I can find) melting point hot melt. The higher melt point glues tend to melt the polystyrene and create thick, dense, blotchy, joints.

    5.1 Low Melting Point Hot Melt.jpg

    I also use PVA (Polyvinyl Acetate) which is white (Elmers for example) glue. I use them as sparingly as possible. I coat and smear the joint to promote a completely covered and thin as possible joint. Hot melt is good for fast work. White glue when you need more working time. I’ve tried many different PVA/white glues but my favorite is Arlene’s Fast Grab Tacky Glue.

    6 Arlenes Tacky Glue.JPG

    As the name suggests it has high tac and in most cases will self-fixture, much better than ordinary white glue. In cases where I want compression of the joint, I use strips of packing tape which is removed after the joint cures. For laminating sheets, I use Polyvinyl Alcohol. Because the carrier is alcohol, it will eventually dry better than water-based materials, and it can be applied very thin. I briefly used shellac for same, which worked, but does not burn out as cleanly. The carrier in most sprayable contact cements will attack the PS foam but can be used if applied very lightly. There are commercial spray glues specially formulated for foam. Don’t waste your time and money. They are all crap for lost foam casting IMO. I still keep my eye out for better glues. The one thing about them all is they are harder than foam and if you sand a joint, the glue won’t erode as fast as the foam. If you try to machine through a hot melt joint with a router bit or belt sander, it will instantly reactivate the glue and load the belt/bit producing a pour result in the aftermath.

    Cutting & Machining: For resizing sheet and block, you can’t beat a hot wire.

    You can also pattern cut profiles with them. They are fast, easy to build, and create little mess. For machining woodworking bits at high surface speed work well. Woodworking bits have generous gullets and clear foam chips well. Climb cutting will often produce better finishes. I like these tenon cutters for bosses and sprues.

    6.1 Plug Cutter.JPG

    This is probably one of my more complicated foam machining projects.

    Sanding & Finishing: I most often use 220 grit, but finishes on Foamular 150 will improve all the way to 320 grit. Use light pressure or the foam will tear locally. If you have a lot of rough shaping I’ve found the wire screen type abrasive sheet used for plaster/drywall finishing works well because it clears the fines well. An evaporative dip coat for surface finish would be great. I have not found a good one. The obvious one is a wax. It will inevitably be too thick and produce undesirable results. I have had limited success with rubbing and polishing successive layers of Carnuba/Johnson’s paste wax. Test a sample because sometimes the carrier will attack the foam. My parts are machine parts and already have better finishes than most sand casting processes. Frankly if I wanted that high of a quality of finish I’d use a different process.

    Molding: One of the beauties of LF casting is the use of unbound sand. The only requirement is for the sand to be dry. Otherwise it requires no special attention or care. Vibration is key to good mold packing. I made a molding rig with better vibration capability. This produces higher density more stable molds that support difficult features……less “leakers”.

    I just use Quikrete fine sand sold in the construction materials section at the big box stores because it's convenient.

    Quikrete Fine.jpg

    It's about 50 mesh which would be very course for conventional sand casting but with pattern coating in lost foam it makes no difference. I've added some here and there for more capacity but have been using the same sand for 5 years and treat it with disregard......except, I try to keep it vey dry, and it passes through a coarse sifter on it's way back onto the sealed barrels so I don't have to smell it during storage.

    The Quikrete stuff is nearly white and turns tan/brown with use. Sometimes there is some residual polystyrene/sand clumps and just I just discard those. If you raise the temp of the sand to about 1800F it returns to looking like it did when it came out of the bag, but the energy to do that is more expensive than just replacing it, even at retail store prices of 6c/lb. Fresh sand does seem to flow a little better under vibration.

    Just about anything will work for LF as long it is dry and of sufficient refractory. If you are using beach sand, you just need to test for organics. Sometimes beach sand will contain a lot of fine shell and the Ca will react with iron and high melt temps. Other natural sands can contain organics. The shape can matter too with regard to how easily and well it packs but I have no problem with the granular silica shape.

    Commercial foundries will often used rounded particle shape in Mullite or Zircon.

    Pouring Basin: I started using “Kush Cups” and then added expendable foil sprues which was an improvement but the conical feed system was still prone to aspirating air.

    I now use reusable offset pouring cup with a Wier.

    This was a surprisingly significant improvement. I have three of them two have 50-100 pours each on them. The pours became more tranquil with little (and often none) flash or smoke. The sprues are round but that’s just a matter of convenience because I have a set of round tenon cutters but they could be (and arguably better) square.

    Pattern Coating: Instead of brushing drywall joint compound as a coating, I transitioned to dip coating with a drywall mud slurry.

    ……but now use a commercial refractory coating which described in the dip coating rig thread below.

    The reason for dip coating was twofold. Compared to brushing it’s very fast and can coat interior or delicate features with ease. It also produces a very repeatable and uniform coating thickness. This produces a more uniform coating permeability which further aids in high yield and quality castings. Sometimes instead of dipping I use a ladle to pour the slurry over the part because the patterns are often so buoyant, they can stress and can break due to the buoyant forces. I made a dip coating rig that is essentially just a vat of slurry with hanging arms for positioning and the initial drip drying.–-the-big-dipper.776/

    People ask me if the commercial coating is really worth it. I think it depends how much casting you do. I think thinned drywall mud can produce near equal result, but the commercial stuff is better when it comes to wetting the pattern, self-leveling, staying in suspension between use, and drying faster. It may also be more permeable but this is a strong function of coating thickness, which I recommend to be no thicker than that require to cover the color of the foam.

    I have and continue to cast many parts with lost foam. To date, they have all been aluminum and 20 lbs or less because I also have a motorsport hobby and most of my castings are automotive parts. I usually want to make one or maybe a handful of parts. Lost foam is perfect for that. If you look at the tubular part at the beginning of this thread, imagine the effort to core, tool, and part the mold. Those issues along with draft/backdraft aren’t even a passing thought, and a complete non-issue in lost foam. It provides wonderful design freedom. I wouldn’t want to make hundreds of parts with my pattern making methods, but I wouldn’t want to make hundreds of parts with any method.

    At this point, LF casting is all about the pattern for me…..the quality of the pattern dictates the quality of the casting. I try to make and detail my castings to look like tooled production parts.

    7 Part Assortment.JPG

    I have more recently purchased a cnc router and am learning to use it which should help speed and accuracy of pattern making. I also want to produce some iron parts for machine restoration.

    hatta, Bob10, MrCrankyface and 26 others like this.
  2. Melterskelter

    Melterskelter Gold Banner Member


    This is a really nice monograph on lost foam casting. It is a terrific resource for someone with no experience but starting out in LF casting (like me someday) or someone wishing to refine their process. How valuable is a summation of 5 years’ experience in LF by an astute observer and creative and capable problem solver? I’d say pretty darn valuable. You deserve a heartfelt thanks from anyone reading it.

    Thank you.

  3. Petee716

    Petee716 Gold Banner Member

    Couldn’t have said it better.
    Patrick-C and Michael like this.
  4. Al2O3

    Al2O3 Administrator Staff Member

    Thanks for the comments and likes fellas. Much appreciated. I woke feeling a little under the weather yesterday and though I wanted to work on a casting project, I couldn't seem to get out of my recliner all morning, so I decided to write about it. Feeling better now. Hopefully it will be helpful to some folks.

  5. BattyZ

    BattyZ Silver Banner Member

    I also appreciate the absolute plethora of LF info you have posted recently. I started hobby LF casting in 13-14, built a "light" CNC in 2017 for foam patterns. Swayed over to sodium silicate/ greensand in 18-20. Casted/machined a 'heavy' CNC router upgrade in '19. But as you've said, it is all about the pattern making. The information that you've laid out on here and on youtube has really put the fire on...looking to complete my 5-axis CNC Foam Router upgrade here in the coming weeks. Hopefully will make for some push-button pattern making(outside of the CAD/CAM). Looking forward to talking gating/riser tech with you and the rest of the home foundry team.
  6. Al2O3

    Al2O3 Administrator Staff Member

    Glad you found the vids and posts useful.

    I'll be looking forward to that. I bought a 4th axis along with mine just to experiment with. Haven't used it yet. It's small but a larger one would be an easy build. I figure all things in time.

    Certainly a different world in LF but as like all things, many ways will get you there.

  7. BattyZ

    BattyZ Silver Banner Member

    I had thought about just making a big 4 axis or really an indexer that could hold a thick 2-4 in foam piece and index at the needed angle. I ended up thinking any backlash by moving/rotating a big piece of foam would offset seams and whatnot. Hoping that a moving 5 axis head and static block to the machine will be better? We will find out.
    I noticed, lol, but as soon as I saw your reusable pouring basin, "This guy knows" I have since made some. Excited to try them. 3d printed the molds and did a smoky burnout.
  8. BattyZ

    BattyZ Silver Banner Member

    20210530_131430.jpg 20210530_134410.jpg Snapchat-498271187.jpg
    Here is an open-face-sodium-silicate-lost-foam-burnout-hybrid cast I did for the gantry of the machine upgrade. 40x8 inches. Worked well enough, came out straight and the top part where there was no sand contact will be the only place that gets machined.
  9. Al2O3

    Al2O3 Administrator Staff Member

    I suppose it could depend on stepper resolution and holding power but you're probably right about the 5th being superior. I saw people doing small sculpted busts which piqued my curiosity. My system uses GRBL for the post which can be a bit limiting.

    Sometimes they get sooted up which doesn't really matter unless you want to patch them. If you cook them at 1800F for a few minutes they will return to pristine white. You can break them in half and still patch them back together good as new. Biggest issue is for me is remembering to extract and dump them out after the pour.....LOL!

  10. BattyZ

    BattyZ Silver Banner Member

    Sounds like the voice of experience!

    Oh, I don't want to be right...I just would like it to work, consistently. I have run LinuxCNC from the beginning but heading to Dynamotion with a KFLOP board, so getting a post-processor to work will be a journey.
  11. theroundbug

    theroundbug Silver

    Great post full of information! I've avoided LF casting until now because I figured it was a pain to make the patterns and each one was consumable. This is making me re-think that opinion

Share This Page