Your prime tower fell over four hours in. Now what?

The model survived, the tower didn't, and every color change after that smeared. Two fixes for this are confirmed by the people who had the problem. One variant is still unsolved, and this post says which is which.

Striped prime tower printed next to a small multicolor figurine on a build-plate-style surface

TL;DR: If your prime tower keeps popping off the build plate mid multicolor print, the confirmed fix is boring: your plate has a soap film on it. Re-wash it with a residue-free dish soap and rinse thoroughly. If colors also bleed after swaps, raise the flushing volume for the offending pair. The H2D variant, where the print head physically knocks the tower down, has no confirmed fix yet. Below: how to tell which failure you have, and what actually worked.

Why does the prime tower fall over?

A prime tower falls over for one of two reasons. Either it loses grip on the build plate, usually because an invisible film of soap residue or grease is under it, or the print head physically knocks it loose, which mostly gets reported on dual-nozzle machines like the Bambu H2D. The adhesion case has a confirmed fix: wash the plate with a residue-free soap and rinse well. The collision case is still an open problem with workarounds but no confirmed cure.

The tower is the canary of a multicolor print. It's tall and thin on a small footprint, and the nozzle comes back to it at every filament change. Any weakness in plate adhesion shows up there first, long before your actual model lets go.

First, work out which failure you have

Look at the base of the fallen tower. If it released cleanly, first layer intact, the whole thing tipped like a book falling off a shelf, that's an adhesion failure. If the base is still stuck to the plate and the tower sheared or crumbled partway up, the head hit it. That's the collision failure, and it's a different problem with a different (and less satisfying) answer.

The confirmed fix: your plate has a film on it

In a Bambu Lab forum thread on prime tower and bleeding problems, user Schumacher had the classic version: tower popping off the plate mid print, general adhesion getting worse. The culprit turned out to be the soap. A spray cleaner had left a film on the plate that a normal-looking wash didn't remove.

The fix was a re-wash with a dish soap that leaves no residue, followed by a thorough rinse. Schumacher came back and confirmed it: "Prime Tower is no longer popping off the build plate and adhesion overall is back to normal." The fix was never in the slicer. It was a clean plate.

The tell that you're in this bucket: everything on the plate sticks slightly worse than it used to, and the prime tower is just the first thing to actually fail. If your first layers looked great two months ago and mediocre now, wash the plate properly before touching a single setting.

Tip

Rinse matters as much as soap choice. Use hot water and keep rinsing until the plate squeaks, then let it air dry. If you wipe with isopropyl alcohol between prints, know that IPA spreads an existing soap film around rather than removing it, so IPA "cleaning" can hide this problem for weeks.

Toppled prime tower lying on its side with curls of purged filament next to it
An adhesion failure releases the tower base cleanly. If the base stayed glued and the tower sheared, the head hit it.

The companion fix: bleed at color changes

The same thread had a second symptom: green bleeding where it met black. Forum user Bullocks pointed at the flushing table, raising the purge volumes for that specific pair. The OP's reply: "That did the trick, thank you for that." If your tower failure comes with muddy color changes, treat them as two problems and fix both. We've covered calibrating flushing volumes for a painted 3MF in detail, including which pairs deserve higher values.

The H2D knock-over variant is still unsolved

There's a three-page thread titled "prime tower knocked down, this is NOT an adhesion problem", and the title is accurate. On dual-nozzle machines the head can clip the tower during travel, and no post in that thread, or the two sibling threads, ends with a confirmed cure.

What people try: a bigger brim on the tower, moving the tower to a corner away from travel paths, and switching the model's infill pattern so the head lifts differently. All of these are guesses that "seem to help a bit" in the reporters' own words. If you're in this bucket, try them in that order, but know that nobody has posted a verified fix yet. We'll update this post when one appears.

Shrink the prime tower so less can go wrong

The prime tower's size tracks your swap count. Fewer filament changes per layer means a smaller tower and less standing plastic to fail late in a print. When you paint a model in Layerpaint, turn on the "Show swap bar" toggle: it shows a live per-layer tally of color changes while you paint, so you can see a swap-heavy design before you slice it. Palette order matters too, because reordering your palette changes which transitions happen and can cut purge volume by a noticeable margin on the same paint job.

Common questions

Does a bigger brim stop the prime tower falling over?

For adhesion failures, no. A brim on a filmy plate just delays the failure by a few hours. Clean the plate first. For the H2D collision failure, a bigger brim is one of the unconfirmed things people try, and it's cheap to test.

Can I print multicolor without a prime tower?

Partly. Bambu Studio and OrcaSlicer can flush waste material into the model's infill or into other objects instead of a tower, which shrinks the tower or removes most of its work. There are trade-offs around color contamination inside the model. See flushing into infill with a painted 3MF before switching it on.

Does moving the prime tower on the plate help?

It doesn't fix adhesion, but for collision cases moving the tower away from the model and the head's travel paths is a sensible unconfirmed tweak. It costs nothing to try and several reporters in the H2D thread do it as standard.

How do I know soap residue was the cause?

Wash the plate with residue-free soap, rinse well, and reprint the same file. Schumacher's case went from tower failure to normal adhesion in one wash. If a clean plate doesn't change anything and the tower base was still stuck when it fell, you're in the collision bucket instead.

Try it now

Layerpaint paints per-triangle colors on an STL, OBJ, or 3MF in your browser and exports a Standard 3MF your slicer opens natively. The swap bar shows you the cost of your color choices before the prime tower has an opinion about them. First export is free, then $2.97 for your next 3 models or $39.97 unlimited. No subscription. Open the painter and see what your model needs.