AMS Purge Waste: What Multicolor 3D Prints Actually Cost

The first time you print a multicolor model on a Bambu Lab AMS, you’re probably going to laugh. The print looks gorgeous — colors crisp, lines clean, tolerances tight. Then you look at the build plate and there’s a small mountain of multicolored extruded filament spaghetti next to your part. And another small mountain of flushed plastic from the poop chute on the floor. Did you really need all that?

Yes. And the cost matters more than most multicolor sellers realize.

For Etsy sellers shipping AMS-made products, purge waste typically runs 30–50% of total filament used, and almost no pricing calculator counts it correctly. If you’re charging based on the slicer’s “part filament weight” you’re underselling by 30–50%. Across hundreds of prints, that’s the difference between a profitable shop and a hobby with extra steps.

This article walks through how AMS purge actually works, the real numbers from production prints, slicer settings to reduce it, and when you should skip multicolor entirely. The Pricing Calculator bakes in purge automatically — bookmark it for the math.

How AMS purge actually works

When the AMS swaps from one filament to another, the new filament has to push the old filament through the hotend until what comes out is clean. If you switch from black to white, the first 50–80mm of extruded white is actually a gradient — gray, then off-white, then finally clean white. That gradient is unusable for the part itself; it has to go somewhere.

Bambu Lab handles the gradient in two ways:

1. Wipe tower (also called “purge tower” or “prime tower”) — a small column off to the side of your part that catches the unusable gradient material as the printer warms up the new color. Looks like a small striped pillar next to your print. Solid plastic, recyclable but not directly reusable.

2. Flush volume / poop chute — extra filament purged out the front of the printer between objects or at the start of a swap. Lands as little curlicues on the build plate or the floor.

Both come from the same physics: when colors swap, the printer must extrude until the gradient is gone before it can lay down clean color again. The amount it extrudes is the flush volume, set in your slicer.

What the slicer actually does

In Bambu Studio (and OrcaSlicer), the flush volume defaults to 70mm³ per filament change, with a flush multiplier of 1.0. That means each color swap consumes 70mm³ × density of the filament it’s purging — roughly 0.087 grams for PLA.

For a print that uses 4 colors with 50 swaps per color (200 total swaps):

200 swaps × 70mm³ × 1.24 g/cm³ ÷ 1000 = 17.4 grams of purge

If the part itself weighs 60 grams, that’s a 77.4g total filament use — 17.4g of which is purge. 22% of your filament is going to the trash.

For more aggressive multicolor designs (16-color logos, intricate gradient work), 30–50% purge is realistic.

A real example — multicolor name plate

Real production print, four-color “name plate” art. Bambu Studio reports:

• Total filament: 142.6 g
• Filament tower: 38.9 g
• Part filament: 103.7 g
• Filament towers as % of total: 27%

So that “100-gram print” actually used 142 grams. At $25/kg PLA, the difference is $1.00 in material per print — not large per unit, but at 200 prints/month, that’s $200/month in unaccounted filament.

The slicer’s “Total filament” number already includes purge. If your pricing calculator asks for “part weight” only, you’re missing the purge. If it asks for “total filament from slicer,” you’re getting it right.

The Pricing Calculator uses the total filament number directly, so this is automatic. The “Multicolor breakdown” section is optional and informational only — it shows you what % of your total is going to waste.

Where purge waste comes from, in detail

Three sources, in order of contribution:

1. Color swaps (biggest contributor)

Every filament change pushes 50–80mm³ through the hotend. A 4-color print with frequent swaps can have 50–500 swaps depending on geometry. Math:

swaps × flush volume per swap × density / 1000 = purge weight (g)

A typical multicolor model that swaps 2–5 times per layer (modest) for 200 layers = 400–1000 swaps = 35–87g of purge. For an 80g part, that’s 30–50% waste.

2. Initial filament loading

Every time the printer loads a new filament from the AMS, it primes the nozzle to ensure clean extrusion. About 10–20mm³ extruded before the part starts.

For a 4-color print: 4 colors × 15mm³ = 60mm³ ≈ 0.075g. Negligible per print.

3. Tool change retraction / purge

Each color swap also includes a small “ramming” sequence to pull old filament out cleanly before loading new. Fully integrated into the flush volume above; not a separate cost.

How to reduce purge — slicer settings that actually matter

Five settings to tune:

1. Flush volume (per-color matrix)

In Bambu Studio: Filament settings → Flushing volumes table. Default is 70mm³ for every from→to color combination. Reality: some swaps need more, some less.

• Light → dark (white → black): 30–40mm³ is enough. White can’t show through black.
• Dark → light (black → white): 80–120mm³ needed. The black contaminates the white badly.
• Same family (red → orange): 50–60mm³.

Tune the matrix once per filament set and save it as a profile. Can cut total purge by 40%.

2. Reduce color changes via design

This is the highest-leverage move. If your model has 200 layer-changes worth of color swaps, see if you can:

• Group layers by color — paint all the red parts in one block of layers, then all the blue, etc. Reduces inter-color swaps dramatically.
• Use color overlays for thin features — a logo etched into the surface only needs 3–5 layers of accent color, not the whole part.
• Print smaller logo elements as separate parts — embed cleanly with friction fit.

A model that takes 50 swaps instead of 500 is 90% less purge. Same product, ten times less waste.

3. Wipe tower size

Bambu Studio sets wipe tower size based on flush volume. Smaller tower = less purge, but only if the math works out (the tower needs to absorb the full flush volume per swap, otherwise some leaks into the print itself).

Default tower size is conservative. You can shrink it 20–30% if your flush volumes are well-tuned.

4. Print on textured plate to skip skirt

The default skirt and brim on multicolor prints uses one color. Skipping the skirt (printing direct on textured PEI) saves ~5g per print on small models.

5. Use single-extrusion mode for the first layer

The first layer of a multicolor print is often a single color even in the slicer’s default. If yours isn’t, set it manually — saves the first-layer color swap purge.

When to skip multicolor entirely

For sellers, the question is always: is the multicolor premium worth the cost increase?

A four-color name plate that costs $1.00 more in filament (purge included) vs a single-color version. If the multicolor version sells for $5 more, that’s a great trade. If it sells for $0.50 more (because the listing is in a saturated category), you’re losing money on multicolor and should drop it.

Three categories where multicolor is almost always worth it:

• Custom name plates / signs — buyer paid for the customization, multicolor is the customization. Premium typically $8–25 over single-color base.
• Functional + branded products — where the brand is part of the product (logo enclosures, branded jigs). Multicolor adds perceived value.
• Detailed art prints — intricate work that singles can’t replicate.

Three categories where multicolor is usually not worth it:

• Articulated dragons / generic models — already a saturated market, customers won’t pay much premium.
• Engineering parts where color is decorative — function dictates use, color is a side note. Use single-color and apply paint/stickers if you need branding.
• Quick-turnaround production runs — multicolor adds slicer time, swap time, and failure rate. Single-color is 30–40% faster end-to-end.

Failure rate goes up with multicolor

One more cost: multicolor prints fail more often than single-color. Reasons:

• Wipe tower can fall over and crash into the printhead
• Filament jams on swap (especially with flexible or carbon-fiber filaments)
• AMS humidity issues compound across more spools

Realistic failure rate for tuned printer:

• Single-color PLA: 3–5%
• 2-color PLA: 5–7%
• 4-color PLA: 7–10%
• 16-color PLA: 12–18%

The Pricing Calculator defaults to 5%; bump it to 8–10% for multicolor prints to be honest. The failure buffer protects you — every failed multicolor print wastes more material (purge + failure together compound badly).

The honest math summary

For sellers running a Bambu Lab P1S or X1C with AMS, here’s what you should be doing differently after reading this:

1. Always use the slicer’s “Total filament” number — never just the “part filament” number. The first includes purge; the second doesn’t.
2. Tune your flush volume matrix — 30 minutes of work, saves you money on every print thereafter. A balanced matrix can cut total purge by 40%.
3. Reduce swap count via design — group colors by layer when possible. A model with 50 swaps prints 5× faster and uses 5× less purge than one with 500.
4. Bump failure rate to 8–10% for multicolor pricing.
5. Skip multicolor in price-saturated categories — single-color articulated dragons are already a money loser; multicolor versions lose more.

Run your numbers in the Pricing Calculator. Plug in the total filament from your slicer (purge included), pick your printer (Bambu defaults are loaded), and see if your current Etsy price covers true cost. For most multicolor sellers we know, the answer is “barely” — and that’s why the math matters.

For a fuller picture of all the costs in 3D printing for sellers, see The Real Cost of Resin 3D Printing for the resin equivalent, or How to Price Your 3D Prints in 2026 for the full pricing strategy.