Casting a windmill gear in grey iron

Discussion in 'Sand Casting' started by Mister ED, Nov 17, 2018.

  1. Melterskelter

    Melterskelter Silver Banner Member

    Well, I decided to try burying SidiumSilicate bound sand the slightly shrunk (accidentally by me) half of the PLA pattern Ed sent me.

    Tomorrow I’ll start by heating it to 160 and testing the pattern’s strength and flexibility. My plan is to gradually increase the heat until I find a temp that softens the pattern enough that it can be worked out of the sand without tearing. Odds of success guessed to be 20%.

    Here is the pattern embedded in sand and coated prior to embedding with paste wax and graphite. CD0B089E-28D1-498C-BF64-3AD4C42A5642.jpeg

  2. Melterskelter

    Melterskelter Silver Banner Member

    Ok, I heated up the silicate bound mold pictured above to 160F. Nothing. Tight and not flexible. Next went to 190 for maybe 30to 40 mins. This was probably not enough time for the whole thing to heat through entirely. However I could see the pattern was bulging slightly due to expansion on heating. I started tugging on it with a hooked scribe and it started to move up a few mms. A few more circuits going round the pattern while tugging here and there and it popped loose! And the mold was completely intact. (I fumbled the pattern a bit when out and it fell onto the mold making the small breakaway shown on one tooth.) Actually extraction was fairly easy. I think some further testing of temperatures should be done to see what the ideal extraction tempt is. But the 190-200 range might be a good starting point. I suspect a higher temp like 220 might allow for a bit more flexibility while maintaining enough strength to pull out in one piece.

    I do think this is a viable method of working with PLA patterns. Given that the undrafted and fairly rough hub portion pulled cleanly, I think a little better print quality and some draft would be a pretty sure-fire setup. Wooohooo!

    Notes on the silicate bound mold:
    1) It was made up of 5% coal, and 5% silicate by weight of sand. The silicate was RU diluted with 15% water and weakened with 12% sugar by weight of silicate.
    2) I kicked off the silicate with CO2 with a technique as shown in the core strength testing video recently posted. It took only 5 holes in the sand and only 10 secs per hole to make the mold quite hard. The holes were placed one centrally and one in each quadrant midway between the corner and the edge of the pattern. Once gassed, the mold was immediately removed from its “snap flask.”

    A54BDACE-E23B-426C-A3C4-5498848F4AE9.jpeg 6F246A58-CC6F-4753-8BB1-AB8272BF09F4.jpeg

    So, if Ed wants to send me another set or two of patterns split as before and perhaps without draft, I would try casting. Then he can get both casting returned for his cost of only one USPS FRB.

    Last edited: Jun 30, 2019
    oldironfarmer likes this.
  3. Al2O3

    Al2O3 Administrator Staff Member Banner Member


    That could be a very useful technique for the right parts. Do the patterns shrink with heat or just soften to the point they deform and are compliant enough be removed. I asked because shirk could also shrink onto to some features as well, like feature that cores the shaft.

  4. Melterskelter

    Melterskelter Silver Banner Member

    Well, I wish I could say I had all that figured out. On cooling the pattern is slightly smaller than it was originally—-I am pretty sure about that. While it is hot it seemed to BULGE some from the mold. So, I assume it must have swelled some. But, because it was somewhat soft, I was able to cleanly extract it despite the surface not being slippery smooth.

    My previous heating of the part in sand had more or less partially embedded sand into its surface. That sand brushed off, but the paint I had applied was rough. I think that softness allowed the PLA to stretch thinning its cross section temporarily. And the softness allowed it to deform to accommodate surface roughness.

    So, I would not worry to much about shrinkage causing entrapment of mold features. The PLA, while hot, is too weak, I think, to grab on very hard.

    I really think the only value of the test I did is to serve as a vague starting point suggesting a lot more investigation by a number of folks to flesh out the possible advantages and limitations of the technique. Also to be determined is optimal temperature, wax/graphite/silicone/other separation aids, soak time at temperature, draft vs less vs no draft, and who knows what else?

    As I read it, the main advantage may be reduction in surface prep and very clean pattern extraction, but at the expense of destruction of the pattern. I was surprised it worked at all.

    I think Ed may send me another pattern or two to try. I hope he has a lot of PLA....

  5. Mister ED

    Mister ED Silver Banner Member

    I just ordered two more rolls, LMBO.
  6. DavidF

    DavidF Administrator Staff Member Banner Member

  7. Melterskelter

    Melterskelter Silver Banner Member


    How does one get a shiny smooth surface?

  8. DavidF

    DavidF Administrator Staff Member Banner Member

    Not sure I understand what you are asking...
  9. Melterskelter

    Melterskelter Silver Banner Member

    Why use the Amazon listed PLA? What makes it special?

    How can one improve surface finish on a PLA pattern since final finishing by hand or with sanders etc is so time consuming.

  10. DavidF

    DavidF Administrator Staff Member Banner Member

    It's a "silk" filament. Hard to really explain but let's just say it prints great and gives you a nice surface finish. It also sands well with some 220 or 320 grit. Down side is that it has less bond strength between layers especially if you print too cold.
    On the sanding part my answer is a catch 22.
    Most pla filaments can be buffed smooth and shiny with a rag and styrene. But it does not work well on silk filaments and it works better on some filaments than others. And yes I mean glass smooth shiny... But you need a good print for starters, not some 300 micron Lincoln log looking crap...
  11. Mister ED

    Mister ED Silver Banner Member

    That stuff is kinda weird, LOL. Before I turned the temp all the way up, it was like it was swelling coming out of the nozzle. Ran a temp tower and now have another gear half printing. Early on, but it looks like it is printing pretty decent ... we shall see in the morning.
  12. Melterskelter

    Melterskelter Silver Banner Member

    Is the PLA you are using “silk?”

  13. Mister ED

    Mister ED Silver Banner Member

    The gear printing now is of the same material that David referenced above. I ordered a roll of it after I ordered a couple of other rolls.
  14. DavidF

    DavidF Administrator Staff Member Banner Member

    Solutech calls it "ultra" pla. But in the 3d printing world it is a silk filament.

    LOL yep thats what it does. Seems to fix alot of the lost pla issues that I experienced with other filaments, and its super easy to print with...
  15. DavidF

    DavidF Administrator Staff Member Banner Member

    Man 19 page long thread.
    Might be pouring this weekend...
    dtsh, Tobho Mott, PatJ and 1 other person like this.
  16. Mister ED

    Mister ED Silver Banner Member

    I showed this to my wife without telling her what it was. She thought it was some kind of fancy toasted marshmallow, LMBO.
    Tobho Mott likes this.
  17. PatJ

    PatJ Silver Banner Member

    It does look rather marshmallowish now that you mention it.

  18. DavidF

    DavidF Administrator Staff Member Banner Member

    Soon to be one crispy marshmallow :p
    Mister ED likes this.
  19. ESC

    ESC Silver Banner Member

    I used a taper reamer to add a little draft to the internal bore, and created some draft inside the hub with modeling clay since that surface will be machined.
    I rolled some aluminum rings to act as forms for the molds and made three. There was slight distortion in the molds, so it took a little hand work to get the pattern to go back in to full depth so the locating dowels could be used to locate the other half. They also came out good, but no photos.
    This is the material for the baked molds. A liter of sand with some flour and 25 ml of boiled linseed oil.


    My standard core plates, mold rings and the sanding strips that I used to clean up the teeth. They were cut at 14 1/2 degree to match the pitch and sandpaper was glued on one side. The other side acts as a safe surface and when the depth is correct for the dedendum the shape of the tooth is not altered.


    Filling and striking off, then rolled over on another core plate and baked at 350* for 90 min.




    I also thought I would try a horn gate, so I made up three different sizes and a square tapered sprue. These will allow me to feed the mold from the bottom without a cheek. They extracted in a test, but I am coating with shellac and sanding between coats.

  20. ESC

    ESC Silver Banner Member

    Here are the other parts for the mold. On the right is a two piece mold, and on the left, 2 parts of a three piece mold. The mold in the back is a spare.

    Tobho Mott and Mister ED like this.

Share This Page