Introduction: Linear Direct 3D Printer Extruder Drive. Fusion 360.

Designed to directly replace NEMA17 in any 3D printer
High Force to Weight Ratio

High continuous and peak forces

Weight 4.2oz (119 grams)

High Accelerations

Halbach Array magnetic system

Scotch Yoke mechanism

Step 1: Linear 3D Printer Extruder

An extruder with a stepper motor is used in my Pusa I3 and also in
almost all 3D printers. I decided to develop a lighter extruder drive based on a 3-phase linear actuator with a magnetic Halbach array. It is designed to directly replace NEMA17 in any 3D printer.

Normally a linear drive (like VCA or similar) requires metal parts, so I've focused on the applicability of 3D prints with a minimum of metal parts. Such a drive can be made practically at home.

Step 2: The Device

The device has a weight of 119 grams, that is, almost 2.5 times easier than the NEMA17 that is used in my Prusa I3.

This drive requires a 3-phase PWM driver and an ARM microcontroller. The drive seat coincides with the NEMA17 31 mm square and M3 screws. The assembly of the entire device is also M3 screws behind. In the same place is located the shaft of the optical encoder of the feedback.

Step 3: Principle of Operation

Like any electric machine, this drive operate on the principal of the Lorentz force equation. Simplified, this force is directly proportional to the current strength, the length of the conductor, the induction of the magnetic field, and the sine of the angle between the magnetic induction vector and the conductor.

To increase the magnetic flux, the Halbach array was applied.
This will increase the magnetic flux from the conductor and reduce it on the other non-working side. Halbach array does not require an back-iron behind the magnetic system, this is good for the weight parameters of the device.

Step 4: Modeling in Fusion 360

I made all the drawings and animation of the principle of operation in Fusion 360.
As can be seen from the video, the linear movement of the magnetic assembly is converted into a circular movement using the Scotch yoke. This is a very simple and reliable solution that is used in many devices.

Step 5: Components

The body of the drive, as well as the holder of the Halbach array and the details of the Scotch yoke mechanism are made using the 3D printing method. In addition, we use 3 bearings 625ZZ and pin 2mmX 10mm for the Scotch yoke.

Magnets are an important part of the design. We need 5 Neodymium Rare Earth magnets 5X5X25mm N42 magnets
You can use compound magnets, they will need 10 pieces.

Shaft D5 X30mm, copper wire AWG28

Step 6: The Coils

The coils are wound on a 3d printed base. The cross section of each coil is 4mm x 6mm. After winding, the coil is wrapped around the coil by PTFE tape.

Two coils are formed in the auxiliary part in such a way that the nested coils are produced.
After forming, all the coils are located in the grooves of the housing.

Step 7: Assembly

The assembly of the magnetic system is very simple, follow the direction of the polarity of the magnets. The direction "north-south" does not matter, the electronic driver will switch the original polarity of the coils in the desired direction.
The Scotch yoke also has no difficulties in assembling, but the shafts of the gear and the encoder must be firmly fixed. To do this, apply a good adhesive when assembling. Next, we collect the device completely. When turning the shaft with your fingers, there should be no mechanical resistance. Otherwise, it is necessary to grind the surface over which the magnetic system slides. As an electronic driver, any 3-phase driver with optical feedback can be used. It should be designed for a current of at least 15 amperes per phase and a voltage of 12 volts.

Step 8: Materials and Files

All the files are listed here.
Filament ABS. Do not use PLA and similar materials, since the drive coils heat up to about 60 degrees, which can lead to deformation of the entire structure and wedging of the extruder.


the_3d6 (author)2017-04-13

Thanks for awesome engine design! Have you measured any parameters, such as current required to produce given torque, max rpm, what power is required for stable extrusion, anything else?

TanyaAkinora (author)the_3d62017-04-14

Greetings to Kiev!
Tests are still in progress. The torque and other parameters are sufficient for use in the extruder drive. We will publish the results and details here as soon as we finish the test phase.

the_3d6 (author)TanyaAkinora2017-04-14

Thanks, waiting for the update! Meanwhile I've read a bit about Halbach arrays, it seems logical to use there cubic magnets - they will increase each other's field more, resulting in much higher torque (or much less coil current at the same torque) for just a few additional grams of weight. It seems that 5x5x5 mm cubes will fit there just fine. If magnets surfaces will touch each other, field amplification will be higher.

TanyaAkinora (author)the_3d62017-04-14

You are a little wrong. The longer the magnets the better for the parameters of the system as a whole.

the_3d6 (author)TanyaAkinora2017-04-14

That doesn't seem right. I didn't make calculations yet, but non-symmetrical structure with air gaps is extremely unlikely to be better than symmetrical one without gaps (because if non-symmetry and gaps increase field, then single magnet would be better than any array)

warhawk8080 (author)the_3d62017-08-01

There is a slight flaw with the scotch yolk's not a perfect sinusoidal operation...the rotation will have flat spots in both the rotation of the bar. Possibly resulting in a non-uniform filament extrude. Still quite an interesting little design for sure

TanyaAkinora (author)the_3d62017-04-14

I already made the calculations. Field graphics are made for the real case of magnets in the device.
Collect the Halbach assembly together without gaps and show us all. After we discuss.

warhawk8080 (author)2017-08-01

Interesting....any idea on the torque of the motor and how it counts steps to ensure accuracy of the extrusion of the filament

Very very interesting..especially if it can extrude without getting too hot..low profile direct drive...awesome!

id11vcna (author)2017-04-16

Wow! Amazing. Nice job.

TanyaAkinora (author)id11vcna2017-04-16

Thank you!

AcrimoniousMirth (author)2017-04-13

Any footage of it in operation?


The tests were done on a third-party PWM controller. Now we are doing an emulator for the printer's Addruino-board 3D printer (the converter of the NEMA17 commands in the drive commands). This will simplify the implementation of the drive in a 3D printer.
We will make a video as soon as it becomes possible.


I greatly look forward to it! An interesting design based on an interesting principle :)


Thank you!

itsonlym3 (author)2017-04-13

any chance we could actually see a vid of it in action?

thanks for the design!

TanyaAkinora (author)itsonlym32017-04-14

see above.
Thank you!

peabody1929 (author)2017-04-13

15 amps at 12 volts for each phase ----How much power does this filament driver use?

TanyaAkinora (author)peabody19292017-04-14

Estimated allowable current density in the coil is 7.3 A / mm2 (repeated short-time switch-on mode). For AWG28 the maximum current load of the copper core at 60 ° C is 1.4 A. But with PWM control, currents in the coil can reach much larger short-term values.
Therefore, transistors (not current consumption) should be designed for a current 15 Amps or higher.

About This Instructable




Bio: Hello! My name is Tetiana. One of my hobbies is 3D printing. My work covers various areas of engineering to the decorative. You can see ... More »
More by TanyaAkinora:Elveet. Kinetic Charger PowerbankGlueDart. Glue Pen With Electric Motor ExtruderLattice for Indoor Plant
Add instructable to: