§9 — Output

NeoTower

The stock wipe tower decides its geometry up front, before the slice, from an estimate of how many toolchanges there will be. That estimate is fine for ordinary multi-material printing and completely wrong for a Sandwich, which inserts toolchanges inside a layer that nobody counted.

auto-promoted for any Sandwich
print-verified
variable layer height is experimental

Where Quality›Prime tower›Tower type

You usually do not have to choose

Any scene that uses a Sandwich or a multi-pass stack auto-promotes to NeoTower, whatever the setting says. NeoTower does need a prime tower to be active in the first place.

Planning before you know

A wipe tower has one job: give the nozzle somewhere to purge the old colour before it prints the new one. To do that it needs a footprint big enough for the purge volume at every layer, and a shape that stands up on its own.

The stock planner works that out from the model and the settings before slicing. Then a Sandwich pass stack adds three toolchanges inside layer 40, four inside layer 41, and one inside layer 42, at microscopic sub-layer heights nobody planned for. The tower has to absorb all of it with a footprint that was decided before any of it existed.

Stock planner

Geometry decided up front, from an estimate. Sub-layer primes arrive later and have to fit into whatever was already built.

NeoTower

Runs after slicing, with the complete toolchange list in hand, including every sub-layer prime. Fixed footprint, delta-Z aware, in sync with the real G-code because it was built from it.

Watch it plan

Add Sandwich passes and see what the tower has to swallow. The blue bars are ordinary per-layer toolchanges; the amber ones are the sub-layer primes a pass stack inserts inside a single layer, which is the part the stock planner never sees coming.

diagram The toolchange list, and the tower it needs add passes and watch the footprint

A diagram of the relationship, not the planner. The purge arithmetic is illustrative; the structure of the problem, and what each control does to it, is real.

Tower type

Setting	Behaviour
Classic (default)	The standard WipeTower2 planner.
NeoTower	The post-slice planner: fixed footprint, delta-Z aware, with the options below.

Zigurat taper

Default: on. It limits how fast the footprint may shrink between consecutive real layers, to one perimeter width per side, so every wall ring rests on the ring below it. Wall on wall. Disable it to save material and time, at the cost of rings that are only partially supported by the sparse interior grid.

Sandwich purge compaction

Sandwich and multi-pass purges happen at microscopic heights, and left alone they would inflate the tower enormously: a purge that needs 10 mm³ at a sub-layer height of 0.03 mm wants a lot of footprint. Purge compaction is a flow-boost cap that compacts those thin purges into a narrower band, with the surplus hanging into the hollow interior.

Value	Effect
`1.0`	Off.
`1.7`	The default.
`up to 5.0`	More aggressive, smaller footprint, more surplus hanging inside. Confirm with a test print.

Sandwich wipe reserve

The volume purged before each Sandwich sub-layer toolchange: Solid, ColorStitch or PathBlend, Top and Penultimate. The config key is multipass_prime_volume and the label is Sandwich wipe reserve.

Default 10 mm³.
Lower gives a thinner and shorter tower; higher gives a better purge.
0 disables it. Needs a wipe tower to be active.

If your tower is using more purge than you expected, this and the compaction value are the two numbers to look at first.

Adaptive layers × multi-tool × Sandwich

Libre Mode + Tower type = NeoTower
experimental

Because NeoTower plans from the real post-slice toolchange list and is delta-Z aware, it is the mechanism that lets adaptive layer height, multiple tools and a Sandwich coexist on one coherent tower. Stock Orca refuses to slice such scenes at all. NeoTower can purge each toolchange at the real per-layer height.

The option Variable layer height (Experimental) sits under Tower type and is exposed only with Tower type = NeoTower and Libre Mode active. It is visible but greyed out otherwise, and it is off by default. When on, the slicer stops blocking:

scenes that mix objects with different layer heights, and
adaptive layer height combined with more than one filament.

The wipe-tower issue that used to leave empty or short tower layers, including the "empty first layer" abort, was fixed in 2.3.1: the tower stays coherent under variable layer height now. The capability was proven on the older 2.2 line and is consistent on the current base too. It stays flagged experimental, so review the G-code before long multi-tool runs.

Since 2.4.8 the same setting carries one more box, Stacked objects: real layers: a part resting on another one starts its layers at its own base and the part under it ends at its real height, so the two meet with no gap. It has its own section and demo on the Placing page.

This is also the answer to the most-asked question about the fork: yes, you can use adaptive layer height with multiple tools and a Sandwich. Enable Libre Mode, set Tower type = NeoTower, and turn this on.

The catch

It runs after slicing. That is the whole idea, and it does mean the tower cannot influence how the objects themselves were sliced.
Compaction is a flow trick. Above the default it puts more surplus into the hollow interior. It works, and it is worth a test print before you rely on it for a long job.
Variable layer height is experimental. Proven, consistent, and still worth a G-code read before a long multi-tool run.
A Sandwich is expensive at the tower. Nothing here changes that. What it changes is whether the tower is the right size and shape for what actually got sliced.
Wipe tower work is the most delicate part of the project. If something looks wrong there, it is worth reporting rather than working around.