Heating PLA Patterns to Facilitate Removal from Sodium Silicate Molds

Discussion in 'Sand Casting' started by Melterskelter, Jul 16, 2019.

  1. Melterskelter

    Melterskelter Gold Banner Member

    Preamble:

    I am looking at the potential benefit and practicality of heating (and consequently destroying) a PLA pattern to make removal of the pattern from a sodium silicate mold easier. In this thread I will simply be posting my observations and progress in exploring this question. I don't anticipate having some sort of cookbook approach to the question developed by the end of the thread. (For instance, different PLA formulations have different melting points and almost certainly will vary in response to heating.) But, I do hope to provide some basic information that will at least help folks get started along this line should they so choose. I would welcome input and questions from anyone interested in the question. I would hope that some effort will be made to more or less stay on topic meaning the thread may be more useful if side talk about the weather, politics, aviation, socioeconomic concerns etc. are avoided.

    [It is true that the patterns I am using were supplied to me related to the Casting a Windmill Gear in Grey Iron thread and I will provide a post in that thread linking to this one. But this thread will have general applicability to PLA patterns and heating them and will only be tangentially related to that thread. Burying this information in that 380-post thread would make subsequent discovery of this subject nearly impossible.]

    The rationale for heating a PLA pattern that is embedded in a sodium silicate mold is removal of the pattern, especially complex patterns with lots of nooks and crannies, can be challenging whether they are in green sand, sodium-silicate-bound sand, linseed bound sand, or other binders like epoxy. I found that to be true with pattern I will detail in this thread and accidentally got started down this path when I heated my silicate mold to about 220 deg F not to ease removal but intending to harden the silicate mold. It was then that I noticed shrinkage of the pattern and got to thinking about "partially lost PLA" methods of mold making. So that's how we got here.

    Let's see if this exploration is a dead end or if it yields some useful results. At the very least I/we will learn what doesn't work and even that can be useful as it might point to other avenues to investigate and also suggest not wasting time trying this particular method.

    Denis
     
    Last edited: Jul 16, 2019
  2. Melterskelter

    Melterskelter Gold Banner Member

    Here are a couple of vids and a photo.

    The first video shows probing of a pattern that has been heated to 340 degF and the second to 380 degF.

    Enjoy seeing me flick the pattern off the sheet right off the bat!





    8EC272DC-32F8-4F46-ADCA-12A9651CB865.jpeg

    Denis
     
    Mister ED likes this.
  3. Melterskelter

    Melterskelter Gold Banner Member

    OK, I packed up a sodium silicate mold today and then heated the mold with patterns in place. The results were not spectacular, but I think it may be possible to do better. One problem inherent in the gear pattern I am using is that the teeth recurve in such a way that they key into the sand. In other words the space betaween teeth at the pitch circle is less than at the root circle. This causes the pattern to resist pulling out when I grasp portions of the softened pattern and tug toward the center. And that caused some fracturing of the sand. But not all the sand tooth profies broke---about half did not. I interpret that to mean I need to heat the PLA hotter yet to make it even softer---almost liquid and then attempt to lift it out. Were it not for this keying phenomenon on this particular pattern I think I could have extracted the pattern cleanly this time. Since it is planned to clean up the teeth by machining, it would be better, I think, to simplify the tooth profile so that there is no keying. For now I will forge ahead with the other pattern set supplied. PLA Melt (2).JPG PLA Melt (4).JPG PLA Melt (5).JPG PLA Melt (6).JPG PLA Melt (8).JPG PLA Melt (15).JPG PLA Melt (16).JPG PLA Melt (10).JPG PLA Melt (11).JPG PLA Melt (12).JPG PLA Melt (14).JPG PLA Melt (17).JPG

    Well, there you have it.
    I have 3 more pieces to try.
    More heat...
    Any other ideas?

    Denis
     
  4. DavidF

    DavidF Administrator Staff Member Banner Member

    Set the mold up on some bricks and heat it up until the 3d printed pattern falls out on own.
     
  5. Melterskelter

    Melterskelter Gold Banner Member

    image.jpg Ummmm, I like it!!!

    I was just restarting the oven when I read your post. Terrific idea.

    Added 20 mins later once the part was safely in the oven and propped up on bricks as DavidF suggested:
    Before packing the mold with sand I had sprayed it with aerosol graphite which wet the surface a bit with whatever volatile they use as a vehicle and then shook the pattern in a bag of graphite hoping to coat it as best I could. Now anxiously waiting for the "baby" to "crown."

    Denis
     
    Last edited: Jul 18, 2019
  6. Jason

    Jason Gold

    Boron nitride maybe???
     
  7. Melterskelter

    Melterskelter Gold Banner Member

    I do not understand this post. ??

    Denis
     
  8. Jason

    Jason Gold

    If it doesn't release cleanly with the graphite, you can try boron nitride. It tolerates TONS of heat and even sticky molten glass slides like owl poop on a door knob. Only downside is it's 50bucks a can. This is steel sprayed with the stuff. If you do multiple coats, it will build like primer.

    20190624_004856.jpg
     
  9. Melterskelter

    Melterskelter Gold Banner Member

    I just returned home after leaving the mold inverted and blocked up as suggested by DavidF. I had left it on timed bake at 460F. It was good that I had to be away as I could not give in to the temptation to “check on” the progress of the melt out. I just left it undisturbed for for 3 hours with the oven at 460 for one hour followed by natural cooling.

    FB50490E-51E2-406D-B655-4E3A410B9D11.jpeg

    4BFCBEA0-EF79-401B-83DF-842184CAACB3.jpeg

    The result was superb. The goo was in a neat little blob under the mold and the mold was CLEAN. Nice results and relatively painlessly achieved. How this would work if there was not a clear path for drainage is a good question. In other words, would it clear because the gooiness of the PLA would pull it off an undercut as the PLA sagged out of the mold.

    I can say that the pattern I used was packed as printed. No sanding, filling, or trimming. I would not have been able to extract it from green sand or silicate sand as such I am quite sure.

    So, I think this method is worth pursuing. Clearly, various PLA formulations will require temperature adjustments and maybe other slip agents (like boron nitride suggested by Jason) might work better.

    BTW, in these tests I used
    -Olivine 100 mesh sand
    -5% Sodium Silicate mix by weight of sand
    -Silicate mix was RU diluted by adding by weight 15% water and 7.5% table sugar
    The mold was gassed using compressed CO2 at 10PSI for 10secs each through four holes (one each quadrant) and 5 secs each the more central shallow holes.
     
    Last edited: Jul 18, 2019
    Jason likes this.
  10. Melterskelter

    Melterskelter Gold Banner Member

    Thanks, Jason. I will remember that option.
    Sound potentially useful.

    Denis
     
  11. Jason

    Jason Gold

    Looking good! POUR that sucker!
     
  12. Mister ED

    Mister ED Silver

    Wow, it sure is interesting to see how the two different PLA formulations reacted. The grey (and blue) PLAs are Hatchbox PLA. The green is Solutech Ultra (silk) PLA. It looks like I need more heat when printing the silk stuff. I had it in the middle of the recommended range, and printed a temp tower prior to printing the gear.

    I also wonder if varying the infill pattern would help with the collapsing and extraction?
     
  13. Melterskelter

    Melterskelter Gold Banner Member

    I just cooked out another Hatchbox PLA pattern mold provided by MrEd. I put it in the oven propped up as before on two bricks open side down as suggested above by DavidF. For an hour it cooked at 460F and that might have resulted in full melt out, but it looked like it was hanging up with a good portion of the PLA having dripped out, but there were still some sizable stalagtites of PLA. It is not possible to actually see inside the mold when it is in the oven and I am reluctant to handle it for fear of damaging it. So I upped the temp to 485F and waited 20 more minutes. By then all stalagtites were gone, so I pulled from the oven and inspected it. The resulting mold is very very clean and sharply detailed. Incidentally, there was no smoking and no objectionable odor—-just a hint of warm plastic.

    87A7D712-DE63-4F46-A5C5-5BD26245DEF6.jpeg 8BE4390C-6328-44A6-872E-88E9E2D65F87.jpeg

    Additional comments: I was able to get a very solid coating of graphite onto the pattern prior to molding by spraying the mold with aerosol graphite and while it was still wet from spraying, transferred it to a plastic bag with a quarter cup of graphite powder in it and then shook the pattern. No graphite was added post melt out. WSIWYG. The bright metal bit you see at the center is a pattern for gating metal into the mold. I’m thinking this will be a very useful PLA mold making strategy for use when the pattern does not lend itself well to cold pulling from green sand or sodium silicate bound sand.

    I will be baking the other mold half this evening. A Solutech filament mold (green) it is all packed and gassed, but we need the oven for supper... Priorities ya know.

    Denis
     
    joe yard and Jason like this.
  14. Mister ED

    Mister ED Silver

    That looks very nice Denis! PLA is nice in that the smell is minimal and not offensive at all. I understand that some of the other plastics need higher heat and some ventilation when printing.
    Oh my, in the kitchen oven? Your better half is more understandable than mine!!
     
  15. Melterskelter

    Melterskelter Gold Banner Member

    Yup, kitchen oven. She’s put up with a lot over the years!

    Mold number two came off without a hitch. I went straight to 485F and it was all clear 50 minutes later. I feel like I can start to trust this method.

    62AEBDBD-3266-4A19-8A54-2F2B7A693CE8.jpeg 3CB2B9D1-EE6C-4F05-ABFC-07CC45B6580F.jpeg

    You can see two specks of PLA on the corners of the pattern at the root circle of the gear. Once I am preparing to close the mold I’ll carefully pick or scrape those off though they would likely do no harm if left.

    Next step is to pack up these silicate molds to incorporate them into green sand molds and cast a gear.

    Denis
     
    Mister ED likes this.
  16. Jason

    Jason Gold

    Wow! That is really clean! If I cooked stuff in my wifes junk samsung oven, she'd kill me. I put the old oven in the garage and use it for this kind of nasty stuff... Funny how the old clunker is dead on balls accurate and the new one is piece of shit.:rolleyes:
     
  17. Chazza

    Chazza Silver

    Well done Denis!

    Your perseverance and thinking is admirable with a an excellent result!

    Cheers Charlie

    PS What does "PLA" stand for?
     
  18. Melterskelter

    Melterskelter Gold Banner Member


    Thanks, Charlie! I’m pretty happy that it seems to be a workable solution. Ya never know what you’ll end up with when you start down one of these paths.

    Polylactic Acid, commonly known as PLA, is one of the most popular materials used in desktop 3D printing. It is the default filament of choice for most extrusion-based 3D printers because it can be printed at a low temperature and does not require a heated bed. ... ”

    I actually know nothing about the printing aspect at this time. But I can see a printer in my future someplace down the line.

    Denis
     
  19. Jason

    Jason Gold

    I wonder if ABS pooped out in those printers would also do this?
     
  20. Melterskelter

    Melterskelter Gold Banner Member

    Seems likely that it would behave similarly. Only one way to know for sure... Time for you to try it out. ;-)

    Denis
     

Share This Page