I came across this video today, they used light weight foamed pla to make patterns for lost foam casting. They had pretty impressive results especially considering, this was their first attempt. This is the U of T paper he refers to near the beginning of the video. https://www.torontomu.ca/content/dam/frames/publications/Fedorov_et_al_2024.pdf
Remember this one over in the lost PLA sub forum? Lost PLA using light-weight PLA | The Home Foundry Thanks for the link Rocco. I didn't read the 27 pages of the paper (but will in time) but in a skim, they compared foamed PLA density(.044gm/cm3) to EPS (.021 gm/cm3). In my speak, that is 2.75lb/ft3 vs 1.31lb/ft3. Of course, EPS/XPS patterns are maybe just a form of high infill, but very low density, and printed patterns can achieve same or maybe even lower pattern mass depending upon skin thickness and infill settings at the cost of pattern strength. In general, I don't like the idea of hollow patterns for LF casting with low infills, but I do like the idea of being able to print LF patterns. To me it's just additive versus subtractive machining. It would be interesting to compare pattern machine times. I could cut their machine vice pattern in their video in 3-4 minutes which I bet is quite a bit faster than it can be printed. Only about 3-4 minutes of the video are devoted to the printed pattern. I suspect they had cold metal or may have had success both times. I left a comment and invited them to join the forum. Best, Kelly
It was interesting, but hard to watch being so long. The lack of safety gear was alarming. Seems like they have a nice shop with a mill, a band saw, ability to weld aluminum, etc. so in that regard I am a bit envious. They had fill issues at the top wrench end and the far tip of the handle. They missed a chance to give the handle a bit of flair where the end has a little larger diameter than the root. To make their part, it would be 7 minutes machining or 3 1/4 hours printing at 5% infill on my equipment. The part could have also been hot-wired manually and glued together like Kelly did on the rudder cheek demo. The print time to make that a permanent sand cast pattern here would have been a little over 4 hours for the part and maybe another hour for gating. It looked like they were on to making some other parts so good on them and it would be great to have them involved here.
I agree with both of you that this wasn't the most efficient way to make a part like this. But, I got the sense from this video that their primary purpose for making that part the way they did was not so much to have the finished part as it was to test out this casting method and in the process, to broaden their skill set.
Yes, the problem was there was precious little information about the details of the printed part. Who's LW PLA? What was the print temp, width, speed, infill, etc. They mentioned the foamed PLA was within 10-15% of the density of XPS/EPS. I think the odds that their print density was that low is quite slim and is more likely scalar multiples of XPS density. He tried to weigh a sample cube but didn't have a suitable scale. I still haven't read the paper front to back, but in my skim, it seemed to be tackling the typical problems encountered when trying to further reduce printed density with foaming agent.........nozzle pressure, controlling print width, sensitivity to print temp, etc. In the thread I linked, the OP of that thread listed what he called the "equivalent print density" to be 3.75/b/ft3. What that meant was divide the total polymer mass of the print by the volume of the print, not the density of printed foam PLA filament, which obviously is still quite dense. I really don't like hollow patterns or patterns with course infill for LF. Any time I have had voids in a pattern it meant casting defects. One of the reasons is, the molten metal will tumble chaotically into the voids, destroy the structural integrity of the pattern, before the metal pressure can support and hold the mold wall in place. It a solid XPS foam pattern, the metal front is separated by a liquid and vapor boundary layer which allows the metal front to advance in an orderly fashion. Still, it can, it can be at least partially addressed by bottom feeding and pattern position......but that limit your feed strategy. I would like to see the approach continue to progress and I think it will. I keep thinking in terms of my typical parts. It's not a linear world and as part size increases, the structural integrity of the pattern becomes more difficult, and with relatively thin walled very high surface area/volume part like one of my intake manifolds, low polymer mass in the pattern becomes very important. The intake manifolds I cast are cut from foam plank that measures 22" x 12" x 8"and that material costs me about $10. Yes, I cut them in pieces so the internal features can be cut on a 3-axis machine and because my designs are modular, and it allows portions to be re-sued on different intakes for the same engine platform. I typically have <4hrs machine time in a such pattern. You'd need a pretty big printer or have to print in pieces and assemble, have a (very) long print time, and I don't know the cost of the foamed PLA, but I bet many times more than the XPS foam. So for me, I'd have to buy a large printer, so I could make patterns that cost more and take longer to make, with thus far what would be higher pattern mass......... In the video, the wrench is a simple part and doesn't exploit the advantages of printing like the ornamental statue in the thread I linked. It's also the case that this was their first time casting, and they have a long way to go on the other process aspects that yield good castings such as melt practice, furnace tune, feed system design, pattern position, etc. It also made me smile a bit that despite the printed part, the runner was a big chunk of EPS, which probably was a major contributor to the degree of success they had. The wrench shouldn't be positioned vertically with the open hex end at the top of the mold. Still with all the above said, if pattern was tilted in the mold, and a proper metal temp, I bet both attempts would have been fully formed. My run time was 25 seconds. 180in/min on perimeter, and 90 on hex, both with .375 depth of cut on n .75" thick stock. I'd take a couple minutes and use my wife's nail file (as far as she knows she lost it in the couch) to sand the cutter radius out of the internal hex points. You need to step up your game Tops Maybe they/he will join the forum and we can hash out the details. I'd look forward to that. Best, Kelly
