Jacob over on HMEM posted this in the casting forum. https://www.homemodelenginemachinist.com/threads/evaporative-pla-casting.35760/ Looks kind of interesting. Don
Don, I'm not a member of that or the RC plane forums but did read through that thread. Thanks for posting it. I have been and remain a bit of a skeptic when it comes to printed lost foam patterns. I have had several people tell me that pla with no infill works for lost foam process (no seperate burn out, and the metal front evaporates the pattern during pour). I always say, ok, have you ever actually done it? Please show me your sample and results, and they usually instantly become ghosts. In a couple instances they show very poor quality castings riddled with flaws and defects. I must say, the discussion at your link at least address the primary issues, that being dimensional accuracy of the print and (activated) density of the foamed filament. I went to the filament links and read the tech data and use instructions. The process they describe to achieve dimensional accuracy seems a bit nebulous and difficult to control. The affect this has on printed line width is less concerning than the affect it has on printed line height. There wasn't any information from the fellow at the HMEM post on the nozzel size he used but I suspect that the print lines may be quite pronounced. Putting that aside..... The activated density of the two LW-PLA filament products linked indicated "as low as" .4-.5 grams/cm3. That's 24-31lb/ft3 density. By comparison, the maximum foam density I will consider for patterns is 1.5lb/ft3 and the lunatic fringe is 2lb/ft3. If you were able to have enough foaming agent in the filament to achieve XPS densities, the filment expansion would be 10x the amount and problem it already is in this filament.......thus my skepticism. At $40-$75 per kg, it seems pretty pricey but with just outer walls printed and no infill, you won't use much material so it may make patterns at a reasonable cost. It would have the added benefit of printing much faster. Without this, at the pattern sizes I use, it would take days or a week to print. To have a hope or prayer of it being viable for lost foam process, it better have darned little material at those printed densities. But for discussion's sake, let's just say, with only outer walls and no infill, you could achieve a similar total polymer weight as XPS foam for given pattern........ -I still don't like the idea of hollow patterns for lost foam method. Any time I have seen hollows with XPS patterns it spells disaster in the form of defects and localized mold collapse. The patterns being solid keeps the molten metal organized and the gases escape as they are created at the boundary between the molten metal and the pattern. The head pressure of the molten metal keeps the loose sand mold in place and stable and the polystyrne pattern does that for the rest of the mold until it is consumed. I'd love to see viable LF printed pattern media, but until I see someone actually casting quality parts of any appreciable size, I'll be happy to watch. Best, Kelly
So I went and took a look at what he had done. Basically he printed a hollow mold and coated the outside with drywall mud and poured in a fashion similar to lost foam. It looks like it worked well for his purpose.. That's all I have to say....
I was trying to decide if he actually did that casting or just posted the pdf, cuz then he asks if anyone is using the process. It's a model engine forum and it didnt look like a model engine part. Best, Kelly
Kelly: I totally agree with you, for 1:1 scale sized parts, this is not really a viable option. For casting model parts it looks like it might just work. I know he said he used a single perimeter wall, but I'm wondering how far down those holes in the part went. The picture of the partial print looks like it has a solid bottom so they don't go all the way through. With no infill, how did he manage to print the bottom of the holes? There are a few people on the model engine forums that are using Lost PLA to cast some of their model parts, Foketry, Myfordboy, and GreenTwin all come to mind, but they are using a more conventional Lost PLA where the pattern is burned out of the mold, and all three use different methods to make their molds. This was the first time I had seen a 3D printed pattern as if it were Lost Foam. I just thought that it looked interesting and might be worth doing a little experimenting. I'm still getting things set up so I'm not ready to jump down that rabbit hole yet. A while back I bought a dead jewelry furnace and got it working, I recently bought another dead jewelry furnace with the intention of turning it into a burnout oven. It's still well and truly dead, but parts of it will be cannibalized to actually build a burnout oven. Don
Yes, I noticed that too, and wasn't very sure what he did there, or what happens within the interior of conical section of the casting. He only has 3 posts and they are all within that thread. He posted links to LW PLA filament but is yet to respond to any of the more substantive questions. The casting looks like it is about ~2" in diameter and 3-4" tall. My comments about use with the Lost Foam method not withstanding, I do like and think printed lost PLA is very useful in block investment and shell methods but it loses its shine for me as part size increases. For the size parts I do, LF is the hands down winner in economics, invested time. I get these comments on my YT channel all the time that say I should be using lost PLA. When I ask why? They usually say because you can just print the pattern. To which I say, have you ever actually done it to make a casting this size? Of course the answer is no because if they had, they'd know the following: For instance for one of my intake manifolds, the pattern is roughly 24"x 12"x 6". That needs a mold that is 16" in diameter and 40" tall and will weigh 500-600lbs. If you compare LF to block or shell for that casting: The cost of pattern and mold materials: In lost foam the mold media is common silica sand that can be reused many times. At $20/ton that's essentially $0. Even at 10 cents/lb retail store cost, still essentially zero with many reuses. My foam pattern costs range from a few cents to $10 for the largest patterns. For one of my intake manifold patterns with the filament discussed in this thread, with no infill, I'm guessing $50-$100. Now, 500-600lbs of real investment or shell media would be an astronomical cost. Well over $1000. Much less if you used plaster and sand for investment, but still $100-200 you still need to mix 500-600lbs of plaster. I just pour the sand in the flask while vibrating. Usually takes about 15 minutes. No need for pattern burn out with the LF method......and molds this size would take a very large high power kiln and over a week to burn out. Got one of those? The labor to demold is quite different. In LF, I just dump it on the ground, pick the casting out of the loose sand, and blow the coating off with compressed air. Takes about 5 minutes. Compared to chipping/hammering off 500-600lbs of investment or shell and washing or media blasting it out of the crevaces. Then you have all that mld media left over. For shell, it needs to be disposed of and for investment you can potentially reuse some of it but it needs to be pulverised for reuse. I just shovel my sand back into the flask and put a lid on it. As far as pattern construction, the labor to CAD model the part is the same for either processes. The effort to process the model fr printing or CAM program, similar. Most of my more complicated patterns would have 40min run times, usually per side. But like printing they run unattended. The big difference is I might have 4hrs of run time and the same printed pattern would be days approaching a week, if your printer was large enough. It's not labor, just calander time, but print failures are painful. The prints do have lines that can be smoothed, but I also sand my foam patterns......so some finishing is required with either process for the best finishes. So if LW PLA was to work effectively in these large LF castings, for all the same reasons, I'd say it would be pretty comparable in overall economics and labor with maybe +$50-$100 increase in pattern costs. OK, I've had my coffee now.... Best, Kelly
Ah, I see now. It was just so the print cross section all started from bed height. The print or the casting? Best, Kelly
Actually, I retract my previous statement after looking at the photos again.. My new theory.. Witchcraft!!
This must be a second print added to the first print.. you can kinda see it protrudes a bit from the edge where it is sitting on the first print.. So yes, cigar tubes printed vertically..
There is also this hole that they could be tied into. But you only see it in this photo, seems purposely hidden in all other pictures...
If that (the cigar tubes) is a second print that has been added, not sure why it was necessary given what the finished casting looks like. It's also a little odd that there appears to be a four holes poked through the center of the tube array in the pattern. Why not just open unless the whole interior is solid. I noticed that hole too? There may have been more than one attempt to print and cast it..........we may never know.... Best, Kelly
Yah, looks like the entire top cylinder is just riser for the conical section. Can't be sure what goes on in the interior of the lower cone. I see Jacob, the OP for the original post over at HMEM joined the forum here today. Maybe he'll stop by and fill in the blanks for us. Best, Kelly
I've been puzzling over the pictures too. The picture that David has in Post #10 has circled the hole in the back of the part. I believe this is an "in progress" picture of the same pattern as shown in Post #9. The stringing between the riser and the actual part appears to be continuous. However, if we assume that what we are seeing in Post #9 IS two pieces, that kind of makes sense. The lower half is seen in progress in Post #10. The upper half could have been printed upside down. That way you could get by with possibly only a single layer forming what is the very smooth top, with nice round holes. There is no way you can get that smooth of a top surface in a single layer, it would look like corduroy at best. The openings would be, less than round. You can bridge in a straight line easy enough. A curve, not so much, you're trying to print a curve in mid-air. To get that smooth of a top layer over zero infill would probably require a minimum of three layers. But if it's 2 pieces, and you printed the top upside down, then you would get the nice smooth surface - and it could be done in a single layer. If you were sneaky enough, you could completely print all of the "cigar tubes" at the same time. If the tubes connect to that mysterious hole in the part like we suspect, you could even print the tubes and that transition at the same time. I can see such a transition in my head, and it would be printable. My biggest conundrum is with the machined part, why was so much of the part machined away? Seems like a waste of time and material. Don
Perhaps that lower cone is mostly solid and he was just being conservative with plenty of riser to insure against shrink. Keep in mind, it's a small (at least by my standards) part, and even though the riser is a high percentage of the part volume, it's probably only a 1-1.5"D x 2" high slug, so it doesn't require a lot of metal, and it enabled him to cap the riser cavity so the mold could easily be packed to sprue/pouring well height. Looks like he has a little spin trap in the feed system too. Then he just poked a couple vent holes in the print above the riser. Best, Kelly
I thought it was a spin trap too, but it looks like everything enters and exits the "trap" on centerline and at the same level. Won't any crap just be pushed straight through it?
Hey All! I'm Jacob, the OP from HMEM. I'll try to provide some more info and keep up with any questions from here on. First off, for some context...The casting I shared on HMEM is a valve head for a "dynajet" pulsejet engine. The cylinder on top of the conical casting is just a riser, the 10 air intake holes extend through the riser to ensure they can be filled with sand. I print a solid "roof" on the riser and sprue so sand cannot enter the casting cavity as it is buried in sand, then drill vent holes just before pouring. To really push the limits of this method, I attempted to cast this husky statue. I believe this would be a challenging geometry for any molding technique. The horizontal areas of this pattern would never print correctly if it were truly hollow, but my slicing software (Cura) has the ability to print infill under areas that exceed a specified angle. In this case I printed infill under sections that were less than 10 degrees from horizontal. The post you see under the husky's tail was added for venting. A hole was cut between the riser and casting to provide unrestricted venting. This is the complete gating setup. In this case, the "riser" is only there to provide as much venting as possible. The entire pattern with gating weighed in at 23 grams, with a volume of 484 cubic centimeters. That gives an equivalent density of 0.047 g/cm^3, or 2.93 lb/ft^3. Since the print is mostly hollow, this value would vary wildly for different sized patterns. I believe this method will work better for larger castings, since there would be less plastic to vaporize for a given volume of aluminum. The pattern is buried in loose sand, the tops of the riser and sprue are opened up. And here is the finished product. At this point the casting quality is really limited only by my print quality. The PLA seems to vaporize completely, and the aluminum captures detail down to the layer lines. My next steps are to dial in my print settings to minimize the amount of PLA used, and to try this method with larger castings.
Here are some other castings I produced with this method, they were poured through risers attached directly to their bases. Also, this is a link to the exact filament I've been using: link
Welcome Jacob. I might join HMEM just to see that engine in action. That's a very high polymer density. Since volume of a casting will go as a cubic function of the 3 linear dimensions, I could see how this might work for you in this respect as size increases as long as the pattern structural integrity comes along for the ride. Other things can get tougher with bigger castings. Presuming the wolf is solid, that will be difficult to scale without shrink problems........but that's more of a casting design isssue than casting process issue. IMO, that is a much more impressive casting than the valve head. I like it! Best, Kelly
