Typhoon: A Polar Dual-Extruder 3D Printer
Typhoon: A Polar Dual-Extruder 3D Printer
Typhoon is a 3D printer built around a spinning bed and two independent nozzles. I designed and built it with Alexander Dornback as a graduate mechanical engineering project at Columbia. The goal was to make multi-material printing faster, cleaner, and far less wasteful.
The problem: printing with more than one material
Anyone who has printed in more than one material knows the aftermath: piles of tangled, multicolored purge strands and towers of scrap built just to flush the nozzle between colors. On a print with frequent material changes, the waste can easily outweigh the part itself.
This isn't the fault of one brand. It's baked into nearly every commercially available FDM printer, because they all feed every filament through a single nozzle. Each switch means pushing out the old material before the new one can flow cleanly.
That waste creates a second, quieter problem: time. Every purge, wipe, and tower layer adds minutes, and across a full print those minutes pile up into hours.
Our solution: polar dual extrusion
Typhoon attacks both problems at once. With two extruders, both materials stay loaded the whole time, so there's almost no need for waste extrusion or purge blocks.
Cutting out the purging alone shortens print time dramatically. On top of that, the polar IDEX (Independant Dual Extrusion - Two nozzles capable of moving independantly) layout opens up new printing modes where both nozzles work at once, pushing speed even further.
How it works
The design comes down to three ideas working together:
The bed rotates. The print bed spins around a central axis, which is the θ (theta) axis.
The nozzles slide independently. Each extruder rides its own carriage along a single rail across the bed, the r (radial) axis, and moves independently of the other.
Together, they cover everything. Combining rotation with radial travel means each nozzle can reach the entire bed while only traveling half the rail.
Less waste, by design
Traditional multi-material printers purge old filament before every material switch. Typhoon dedicates one nozzle to each material, so there's no cross-contamination to clean up: no purge tower and no waste extrusions. A material change is simply a handoff to the other nozzle. The result is less wasted filament, a lower cost per print, and a faster print overall.
Faster through symmetry
With two nozzles able to print at the same time, Typhoon gains two speed-focused modes:
Duplicate mode: both nozzles print the same part simultaneously, doubling throughput.
Mirror mode: both nozzles print symmetric halves of a single part at once, roughly halving print time.
That makes Typhoon well suited to rapid prototyping, whether you need a symmetric part or a matched pair of identical parts for a subassembly.
Why dual extrusion?
Independent dual extrusion hit the sweet spot for this project:
Less wasted filament: no purge towers or scrap between material switches.
Faster prints: both nozzles can print at once in mirror or duplicate mode.
Simpler than the alternatives: no tool-changer mechanics or complex multi-material feeders.
Feasible in one semester: few moving parts kept the scope manageable.
Why polar?
Putting that dual-extrusion system on a rotating bed brought its own advantages:
Low mechanical complexity: fewer moving parts than a full Cartesian IDEX machine.
Shorter toolpaths: each nozzle only covers half the bed radius, which also makes collisions far less likely.
New printing modes: the natural symmetry of a round bed makes simultaneous symmetric printing possible.
User experience
A printer is only as pleasant as it is to use. Alex set up a touchscreen interface running KlipperScreen, which gives live temperature readouts for both extruders and the bed. It also puts movement, extrusion, and print controls a tap away.
Software and hardware stack
Typhoon is built on a Creality Ender 3 frame. A BigTreeTech controller board runs Klipper firmware, with Mainsail for web control and KlipperScreen on the touchscreen. SuperSlicer handles slicing, and Sunlu filament feeds both nozzles. Alex handled most of the Klipper configuration. I wrote the Python post-processor that translates standard slicer output into motion a spinning bed can follow. Below is our custom machined polar bed, with countersunk bolts, a ceramic resistive heater, and a "lazy Susan" style bearing.
Analysis: sizing the bed motor
The spinning bed is the heaviest moving part of the machine, so its motor needed careful attention. A NEMA 17 stepper typically operates at around 2–5 N·cm. Ours drives a 72-tooth gear on the central bearing from a 20-tooth pulley, a 3.6:1 reduction.
The analysis showed that if the belt is over-tensioned or the bearing friction runs high, the motor can be pushed close to 20 N·cm. That's roughly four times its typical operating range, and it shows up as heat. The fix is more headroom: a larger drive gear, a stronger stepper, or both.
Physical demonstration and prototype
After a semester of machining, wiring, configuring, and a healthy amount of debugging, Typhoon came to life. The heated bed, which Alex built a good share of, spins beneath both hotends. The whole machine homes cleanly and prints. The highlight: two filaments printed simultaneously on a polar bed, as shown below
Future work
Typhoon is a working prototype, and there's plenty of room to grow. Next up are hardware upgrades, including a stronger bed motor and camera-based print monitoring. We also want to add sub-symmetry printing: splitting an asymmetric part into its symmetric sections so both nozzles can share the work, with a single nozzle finishing the rest.
Credits
Built by Maysarah (Miso) Sukkar and Alexander Dornback. I led the polar dual-extrusion mechanical design, motor analysis, slicing pipeline, fabrication, and assembly. Alex set up the touchscreen, did most of the printer configuration, built a good share of the heated bed, and helped with assembly.
Comments