A production manager once asked us to squeeze 48 pendants onto a single flask to save a casting cycle. We held the line at 32. The instinct was to fill the flask, but density has a ceiling: beyond it, metal cannot flow evenly into every mold and you trade a saved cycle for a 30% scrap rate. Tree sizing is a balancing act between throughput and defect rate.

What limits how many pieces a tree holds

Three physical constraints set the number. First, flask diameter: a 3 inch (76 mm) flask cannot hold a central sprue thicker than about 12 mm plus branches reaching 30 mm out. Second, metal feed distance: each sprue channel can only feed a mold so far before the metal cools and cold-shuts. Third, mass balance: a tree lopsided with heavy pieces on one side and light on the other spins unevenly under centrifugal casting and gives uneven fill. We covered the physics in vacuum vs centrifugal casting; tree density is where that physics hits the schedule.

Rough piece counts we plan to

These are starting points for a standard 3 inch flask with a 10-12 mm central sprue:

  • Plain rings (size 6-7): 24-36 rings per tree, arranged in tiers.
  • Small pendants / charms (2-4 g): 30-45 pieces, but only if sprue lengths stay under 35 mm.
  • Bracelet links: 40-60 small links on a well-branched tree.
  • Statement rings with large stones: 6-12 pieces max, because each stone pocket needs a generous sprue and vent.

When stones sit in the tree, the rules tighten further; see spruing around large stones in investment casting and our earlier note on tree piece count and layout for the geometry. The sprue itself is sized to the heaviest piece, covered in sprue design for jewelry casting trees.

The cost of over-packing

Push past the counts above and defects rise fast. Cold shuts appear on pieces farthest from the sprue. Porosity clusters where metal flow collides. Flash and investment bleed happen where branches crowd. We would rather run two flasks at 80% density than one at 110%, because the scrap and rework cost of the dense flask wipes out the cycle saving. Weight planning also matters: we track casting weight loss and recovery percentage so the tree mass feeds directly into the gold melt calculation.

Balance the tree, not just fill it

A tree that is heavy on one side spins off-center under centrifugal force, and the heavy side casts short while the light side runs fine. We plan the tree so the mass is balanced around the central sprue, alternating heavy and light pieces at each tier. We also keep the heaviest pieces closest to the sprue base, where the metal arrives first and hottest, and put the lightest pieces on the outer branches where flow is weakest. That rule is the same investment-casting logic described in the investment casting overview on Wikipedia. We count the tree weight before investing it and compare it against the melt we weigh out, so a 30-piece ring tree does not get a 50-piece melt of gold that we then have to recover and re-refine. After divesting, we count how many pieces survived, how many need repair, and how many were scrapped. That yield number feeds back into the next tree layout; a design that consistently loses four pieces a run gets a re-sprue rather than a re-cast. We also photograph the sprue tree before investing it, so if a piece comes out wrong we can trace whether the problem was sprue position, metal temperature or investment. Buyers on repeat orders get the same tree layout, so the yield stays predictable rather than wandering batch to batch. We also photograph the sprue tree before investing it, so if a piece comes out wrong we can trace whether the problem was sprue position, metal temperature or investment. Buyers on repeat orders get the same tree layout, so the yield stays predictable rather than wandering batch to batch. We also photograph the sprue tree before investing it, so if a piece comes out wrong we can trace whether the problem was sprue position, metal temperature or investment. Buyers on repeat orders get the same tree layout, so the yield stays predictable rather than wandering batch to batch. We also photograph the sprue tree before investing it, so if a piece comes out wrong we can trace whether the problem was sprue position, metal temperature or investment. Buyers on repeat orders get the same tree layout, so the yield stays predictable rather than wandering batch to batch.

FAQ

Does a bigger flask let me scale proportionally?

Partly. A 4 inch flask holds more, but the central sprue and feed distances grow too. We scale tiers, not just piece count, and re-test the first tree of any new size.

Why not pack pieces tightly to save investment?

Investment cost is small next to metal and labor. Crowded pieces weld together on the tree, which means hours of cutting and finishing, plus a higher breakage rate.

How does stone size change the count?

Each cast-in stone needs its own sprue and clearance. A tree of 10 mm gemstone rings might hold 10 pieces where a plain ring tree holds 30.

Do rings need to be spaced around the center sprue?

Yes. We stagger rings in tiers so they do not shadow one another from the metal flow direction. Directly stacked rings give the upper ring the metal first and starve the lower.

Need a foundry that sizes your tree for yield, not just fill level? Talk to our production team — we optimize tree density per order and share the layout with you on repeat runs.