Spindle pattern making

Discussion in 'Sand Casting' started by Mark's castings, Apr 16, 2020.

  1. So I've decided to bite the bullet and make a single use pattern for the sand mixer project, these are the parts I attempted in lost foam earlier. This time round I'm making it from materials lying around like some scrap aluminium tube from gas cylinders and MDF sheet. MDF is pretty easy to work with, it soaks up epoxy resin like a sponge and machines quit well on the lathe afterwards: a bit like plastic does. This was supposed to be a quick and dirty pattern for resin bound sand but has gotten a bit out of control. Right now it's all had a coat of epoxy instead of spray bog filler as that can take days to dry enough to be sandable whereas the epoxy is rock hard 24 hours later and has a gloss finish. The fillets are polyester car bog filler and I use some contoured cheap hacksaw blades as putty knives as well as some scrapers made from hacksaw blades to shave off the excess filler before it fully cures and goes hard. A bit of a sand and you have some nice hard durable fillets. The central core mould is made from split PVC pipe with extra to take it from 150mm/6" to 175mm/7" diameter.

    Starting ingredients: Aluminium pipe and MDF sheet
    spindle pattern 0.JPG


    Top part of the pattern showing the flange and the locator for the core.
    spindle pattern 2.JPG

    Everything is machined and the epoxy is cured, needs a sanding to get rid of the furry texture on the end grain.
    spindle pattern 1.JPG

    PVC core mould.
    spindle pattern core 1.JPG
     
    Last edited: Apr 16, 2020
    Tobho Mott likes this.
  2. Melterskelter

    Melterskelter Gold Banner Member

    Looking good, Mark. Looks like you are doing a first-class job.

    Forgive me for rambling about a few things that are probably very familiar to you, but might be useful to some reading this thread.

    Being a rather impatient pattern maker, I often substitute polyester resin as a surface treatment rather than using epoxy. I like to bump up the kicker ratio in the poly to accelerate the cure rate as it will cure hard in a few hours rather than a day. With epoxy I am stuck with standard mix rates and cure times. Also the poly is a bit cheaper. 6-minute epoxy often mixed with fine wood flour (dust) as a thickener is nice for small gluing/filling but quite expensive for larger jobs.

    To fill those niggling little pinholes that seems to show up as I am applying white lacquer finish, I use a blob of the fine dust that accumulates at my disc sander to the hole and then drop a drop or two of thin cyano glue followed by accelerator to get an instant hard fill that I can sand off immediately.

    By the way, white lacquer (Rustoleum makes a good spray can) really helps me see surface defects much more easily than clear finish and dries in under an hour, especially in the sunshine or an intentionally warm shop which speeds up everything. If you have used silicone release agent on the pattern during fabrication, pinholes will bedevil you. (Johnson’s floor wax works well and does not cause pinholing). To get the silicone to stop telegraphing through as pinholes, a coat of spray-on shellac (Zinsser brand) seems to cover and seal the silicone and is very fast drying. After a coat or two of the shellac, the silicone will stop pinholing your lacquer.

    Looks like you will be ready to cast this very soon.

    Denis
     
  3. That's a good idea using the faster polyester resin, it's good stuff although I'd like a longer shelf life than I get in the tropics: I've had a tin or two go solid after 6 months. The epoxy I used was soft enough after 12 hours to take a fingernail impression but was hard after 18 hours, I suppose I could put it in an oven to speed cure. The epoxy didn't wet the aluminium until it had partially cured: plenty of areas it retreated from so possibly some surface contamination although I did wipe it down with toulene to clean it before painting on the resin. I'll block sand the epoxy to flatten and smooth the ripples and fill any void with nitrocellulose blade putty that dries quickly and then prime it. I actually have a few cans of Rustoleum auto acrylic lacquer in white, black and satin black so I can give it a try too, it's the fastest drying spray paint and cheap too: $10 a tin on sale. I do have some thick superglue and accelerator, so it's an option for pinholes too. I'm trying to get this ready by the weekend if possible so I can make the sand mould.
     
  4. Melterskelter

    Melterskelter Gold Banner Member

    My wife has endured the perfume of curing epoxy and poly in her oven on many many occasions.

    Any petroleum-based solvent does not leave a clean surface good for coating with any liquid as trace amounts of low-volatility ingredients are always present. Those low volatility (read that “grease”) ingredients remain after the main ingredients evaporate and trick you into thinking the surface is clean. On the other hand, isopropyl alcohol is free of low-volatility ingredients as are other alcohols. That’s why manufacturers suggest isopropyl as a pre-application prep.

    I use “real” lacquer as it dries harder and faster than “acrylic” lacquers. upload_2020-4-16_6-38-38.jpeg

    https://www.amazon.com/Rust-Oleum-1904830-Lacquer-Spray-11-Ounce/dp/B000PIIKDG

    Denis
     
  5. I should know better regarding solvent residue when cleaning, I do have some methanol I bought by mistake and need to use up (can be absorbed through the skin to poison you) that may do the trick. Methanol can be used for nitrocellulose thinner too. This acrylic lacquer is about the only commonly available stuff over here, it dries quite well so long as I can override my natural instinct to slather on a thick coat that forms a skin.

    https://www.autobarn.com.au/motospray-satin-black-400g-sb400
    https://www.autobarn.com.au/motospray-primer-putty-400g-oh400
     
  6. The actual hollow spindle pattern was completed today, no lathe work this time round, just scribe circles with the dividers and cut out four discs with the bandsaw and then round them to the scribed line with the disc sander and give them taper at the same time. The main body is two 90mm PVC drain pipes with one split put over the other, then packed with small PVC rings at the base and one ring about 40% of the way up to give taper. The core mould is 65mm PVC pipe split and held together with cable ties. Once it has the release agent painted on, both pattenrs will be ready to use.

    spindle pattern 3.JPG
     
  7. I made some more progress today with the spindle moulds, these are using CO2 hardened sodium silicate bound sand with some sugar added to aid the shakeout after casting. The two core moulds and the small inner spindle tube mould all worked flawlessly and the pattern pulled out easily. I'll have to get some more sodium silicate to complete the rest of the moulds. Having silicate sand allows me to combine multiple batches of sand from the mixer as I made up to 13 kilos of sand for one mould. The second brown piece at upper left is the hollow core for the outer spindle tube, it has a 4" cardboard tube inside sealed at both ends.

    spindle mould1.jpg
     
    Melterskelter and Al2O3 like this.
  8. Melterskelter

    Melterskelter Gold Banner Member

    Those molds look crisp and clean. Nice work. Any observations you’ve made about things like release agents, mixing ratios, type of silicate, CO2 technique?

    Denis
     
  9. In case anyone is wondering, the white silica sand mould isn't one of mine, it's only the dirty brown sand moulds.
     
  10. I have been on a steep learning curve, I'm using homemade sodium silicate made from silica gel kitty litter and sodium hydroxide to bind the sand. The wet mix seems to dry a little while sitting before use and the narrow core piece was made with added water instead of neat sodium silicate which seems to improve the sand coating: it hardened fine with CO2 gas but I did put it in the sun for a few hours just in case. So it seems a little added water may aid the 5% by weight sodium silicate to coat the very fine sand. The sand had to be processed to remove the super fine grains, so it would tend to need more sodium silicate to work compared to coarser sand. I accidentally made the big core with 1% sodium silicates yesterday and it actually went solid but was super crumbly, 5% has enough strength to be handled but is similar to PUNB resin in that it can be shaped with metal tools by scraping/filing. The CO2 technique needs some work: I was able to harden the narrow core with gas from one end and the same for the big hollow core, while the mould at the middle of the photo was gassed from the top extensively and was still soft at the bottom so I had to use a stiff wire with a flattened end to drill 8 holes in the sand to allow gas to penetrate from the bottom. The pattern was temporarily screwed to a board along with the PVC tube around it so it was easy to flip upside down and remove the board to expose the sand for gassing. Gas needs something like a plywood lid with a rubber sheet on one side to seal well enough to ensure a good dose so I'd fabricate some in different sizes if I keep using sodium silicate.

    Gassing could be a bit hit and miss in regular use, but I understand propylene carbonate can be used instead to get a 20 minute hardening time, so may be more practical than gas.
     
    Last edited: Apr 27, 2020
  11. Melterskelter

    Melterskelter Gold Banner Member

    I also have found holes in deep molds are needed to allow gas penetration. I just used 1/8” rod pressed into the moist sand to make the hole. Once the holes were made, placing the nozzle of the low-pressure CO2 tube in contact with the hole opening and giving a seven to ten second flow of CO2 did the trick.

    If I recall correctly, propylene carbonate took considerably longer than 20 minutes (quite a few hours ) for good hardening. If you try it, please post your results. If there is a more effective way to use it, I'd really like to know.

    Here is thread where I posted much of my experience making silicate-bound molds.
    http://forums.thehomefoundry.org/in...-ru-for-core-making-in-cast-iron-molding.702/

    Denis

    PS: I was never able to find the 5 to 10 minute catalyst PatJ talked about in a couple posts in that thread.
     
    Last edited: Apr 27, 2020
    Mark's castings likes this.
  12. Today I ran the furnace and successfully cast the two sand mixer components. The sodium silicate sand moulds did the job, the particular fine sand I used needed to be rammed solidly to get a fine finish and about 8% sodium silicate with 12% sugar (of the silicate weight). This sand needs nearly double the usual PUNB resin when compared to pure silica sand of the correct mesh size. There was a lot of variation in the strength of the cured sand so I need to develop some consistency in what I'm doing, some pieces were friable, some were tough to drill holes to carve the runners. The furnace performed fairly well: it used 10.8 litres (measured with a dipstick and calculator) of fuel to melt a brimful AT30 crucible of aluminium (12Kg) in 26 minutes from cold to poured in the moulds. That was stopping to thoroughly skim the melt and then heat for five minutes after, if anything it may have been too hot as I had a lot of sand in the metal surface. The castings have sand embedded in place and other areas are comparable to resin bound sand. The moulds charred black thanks to the sugar and the sand softened with water and disintegrated so it's easier than resin sand to remove once you're finished and the lack of smoke was a bonus too.

    Silicate Spindle casting 1 .JPG
     
    Last edited: Apr 30, 2020
  13. Melterskelter

    Melterskelter Gold Banner Member

    Nicely done, Mark.

    Do you think the unevenness of binding was due to not quite enough mixing of sand with binder? What mixing method do you use? What do you think is the problem with uneven bonding?

    Denis
     
  14. The very first mould had 1% sodium silicate as a traveling salesman was pestering me as I did the dosage calculations: it actually held together enough to pull from the pattern but was super crumbly, so I can't rule out further errors like that. The second two moulds with 5% silicate were definitely inadequately rammed out of worry the sand would not vent, and the sand grains were friable too. Fine silica sand and polyurethane no bake resin shrinks a bit when mixed and needs a few raps to pack tight enough to make a mould, whereas this combination mix expands significantly and needs ramming like green sand. For the final lid mould I used 8% silica sand, rammed tightly and it was a bit stronger than polyurethane bound sand and the sand particles held firm. The finish of the metal in contact with 8% sand was the best out of all the castings.

    I borrowed my friends sand mixer machine: it has to coat sand with resin in under a minute once the second component is added. These castings are for a duplicate machine, by the way.

    mixer.jpg
     
    Last edited: May 1, 2020
  15. Melterskelter

    Melterskelter Gold Banner Member

    I see. So, you didn’t have hardness variation within a given batch. You’d think your mixer would provide even results. In my case, I just use a common paint mixer in a half-inch drill motor to mix the sand and binder in a 5 gallon bucket. That seems to be adequate as well.

    upload_2020-5-1_5-55-37.jpeg

    FWIW, I do quite firmly tamp in the 4% weakened and diluted RU mixed with 100 mesh olivine . I use a robber mallet and light blows. The silicate-bound sand seems very porous and requires no venting other than the gassing holes I have to make anyway.

    It sounds like you are getting your particular mixing parameters sighted in as you try various options. I really like silicate bound sand for its detail rendition and strength.

    Denis
     
  16. I can see it's a process that will need a fair bit of experimentation to zero in on good results every time. If I get a recipe that works for my material, I can build dedicated gear to make it easier. I see in paragraph thirteen of that Tom Cobett article you linked to, that propylene carbonate gives a ten minute cure time for the silicate sand: I'd be inclined to opt for that over gas as I can mix it in easily with the sodium silicate (and I can use it to make K-bond sand).

    After machining casting has particles of sand 3mm or 1/8" below the surface so there was still loose sand in the mould even after blowing it out during assembly, or washing off the mould walls: it looks just like a casting made from recycled runners with sand still attached when they go into the melt.
     
    Last edited: May 2, 2020
  17. Rocketman

    Rocketman Silver Banner Member

    Are you mixing Olivine & silica sands?
     
  18. No it's local beach sand: mostly silica with some mica, garnet and mineral sands such as rutile and ilmenite. I wash it four times to remove salt and organic material, sieve it and this batch I used an industrial fan to blow the fines out of the sand as it normally takes double the polyurethane resin when using that stuff to get strength.
     
  19. Melterskelter

    Melterskelter Gold Banner Member

    I never was able to replicate short setting times of propylene carbonate that Cobett reported. I think it aided responsiveness to CO2 and did provide slow set times---many hours. But, however I used it (maybe incorrectly?) I could not get the hoped-for fast setup.

    Denis
     
  20. Good to know in advance, it's hard to get here without paying top dollar from a lab supply company. I think Ebay UK was my best option for the stuff.
     

Share This Page