DIY 3D Printer | Concept

A DIY tool changing 3D printer

Time for a new project

With the ScaraClock finally finished, it is time for a new project!

I have always found 3D printers extremely fascinating. I have used them extensively during my student years and now at work for both prototyping and my own various projects. However, for some reason I never had a proper 3d printer myself. I had(have) a ooooold prusa i3, one of the very first ones. But it simply required way to much fiddling to print anything on it. It was slow and quality suffered as Z height increased. So now it’s in a box in the basement..

I have acces to printers at work, but I miss having my own machine at home. So for a while I have been looking at different machines out there and considered to build a VORON, annex engineering or a Vzbot. But the E3D tool changer (and all the others) are also just so cool! Using a Haas CNC at work from time to time, I know that these tool changing type of machines are the future of FDM 3d printing.

Now, I am not the type for buying finished products.. I always go for at least kits that I have to build myself. But none of the machines mentioned above had everything I wanted from a printer..

I see only one solution to this problem.. I have to design my own!

Draft Concept

This is what my current draft design looks like.

The asthetics are generally inspired by the very beautiful Voron 2.4. I really love the look of this machine. It is elegant and clean. Thats why my design generally looks like The Voron 2.4. When we dive deeper there are still resemblances to the Voron, but also very big differences.

Crossed gantry XY Motion sytem

The XY motion system is a crossed gantry cartesian style mechanism.

The entire XY mechanism is moving in Z instead of the bed. There are two main reasons to do this:

Most of my prints are not very tall. This means that most of the time the XY motion system will be close to the bed (bottom of the machine). With all the moving mass placed low, the machine will be more stable while printing. This will help reduce ringing.

It looks badass!

In addition, the XY motion system:

Have a 350 x 350 mm working area -> Printable area.

Can do 200 mm/s @ 1g acc. quality print speed – Hopefully.
I am aiming for the same qulity and speed as mr. Simon Vez is doing with his VzBot and the Annex team is doing with their printers.

Has a maximum travel moves speed of >1000 mm/s – Hopefully.
This will greatly help with issues caused by oozing.

Will use no 3D printed parts in structural elements. Only metal or carbon fiber.
This machine must be capable of working with a 80C chamber. While there are materials that can handle this (I.e PC) I simply do not want plastics in my motion system! They are not great for stifness and stability. Especially not at elevated temperatures.

Will be easily removable from the machine for maintenance and upgrades.

Will support both NEMA17 and NEMA23 size motors.

Will use normal toothed pulleys on a shaft with two large bearings.

Uses 9-10 mm belts.

Read On

Pick and Drop Tool Changer

The machine use a automatic pick and drop style tool changing printhead because, it is the only real good method of printing in multiple materials in one print. I.e soulible supports or flexible sctions in otherwhise stiff parts.

In addition, the tool chganger will:

Hold at least 4 tools.

For now be heavily inspired by the E3D tool changer.

Have tools parked in a rack that move along in Z with the XY motion System.
The tool rack must be able to move to even do tool changing with a CroXY motion system. In addition if the rack trails the motion system then it will keep the umbilicals as short as possible as they do not need to reach the full Z, but only the XY plane.

Belted Z Axis

The XY mechanism will be lifted by 4 motors. This is generally overconstrained as only 3 points are required to define a plane. But Eddie the Engineer explains it well.

The Z-axis will not use lead screws or ball screws but belts. This is getting more and more common I.e. Voron 2.4.
Having 4 motors lifting the Z-axis provides much more torque and power than a single motor. This combined with a belted Z drive provides much faster Z moves.

The only downside to belted Z is that it will drop down when the machine is turned off or the steppers are dissabled if the weight is to high. Therefore, something must be done to prevent this. I am still not sure how to prevent Z-drop yet. But my options are:

Worm gears
Worm gears are non backdrivable and will prevent Z-drop. They are rather slow though, as the typical minium reduction is 1:20. Quality worm gears are expensive. If I choose this method then I will have to make a DIY solution like Annex does (prooven tech).

Strain wave gear box
This is what is shown in the design. It is another none backdrivable style gear box that can be made to a 1:10 reduction but are very expensive. Using these would have to be a DIY and the question is if that will be precise enough?

Belt redcution with brake
Voron has a simple 1:4 belt reduction and with this the rail drag and detent torque of the motors are just enough that their gantry does not drop. However, my gantry will be much heavier. So I suspect a electromechanical brake to at least 2 of the 4 Z-axis will be needed.

Belt reduction with counterweight
Another way to prevent dropping is to add a countermass. This will also result in the fastest possible Z moves of the four options as the motors only need to move the inertia and not also lift the motion system.

In addition, the Z-Axis will:

Have a 450-550mm working range.
I am using 600 mm rails as I already had 3 of these. The final working range depends on the tool changing mechanism.

It will be capable of >100mm/s travel moves.
This help keep tool change times low. It helps create cleaner multi material prints where Z-hops are important. And it homes faster… I hate slow homing!

It will allow for two independent carriages to move on the same rail
This is required to support a independent tool rack and possible nozzle cleaning.

Magnetic Heatbed

The heatbed will be a Mandala Rosework style magnetic bed. It will be machined from a cast aluminum plate. These plates are very temperature stable and are very flat.

In addition, the bed will:

Be AC mains powered with a keenovo silicone heater.
An AC powered bed takes a huge amount of load from the 24V power supply and allows for much more heating power -> reduced heating times.

Be used to heat the enclosure.
Fans will be used to pull the hot air from the top of the enclosure down and blow it under the bed. This circulation will heat the chamber and create a more even chamber temperature.

Be capable of reaching at least 150C.
These temps are required to print materials like PC and will also result in higher chamber temperatures.

Have high temperature neodym magnets screwed in place.
This is the Mandala Rosework style bed. I really like this approach as there is no risk of adheasives falling off or loss of magnetism at high temperatures.

Accept standard PEI coated spring steel sheets.

High Temperature Hotend

I plan to use either the Mosquito magnum+ or the Phaetus Dragon in their water cooled version.

The requirements for the hotends are:

Be capable of reaching at least 400C.
This is required to print in more technical materials like PC with some heater headroom.

Be water cooled.
The heated enclosure could result in jamming issues if not properly cooled. Besides, the hotend fans are usually the loudest part on a 3D printer.

Be capable of high flow.

Frame and Build Approach

The frame is built with 30×30 and 30×60 extrusions to increase stifness.

In addition, the frame will:

Use no 3D printed structural parts.
All structural elements of this machine will be made of metal. This is done to increase stifness, to improve fire safety, and to avoid any long term sagging/deformation.

Be designed for easy CNC machining.
To avoid 3D prints I may have to machine my own parts at work or on a future CNC machine. Therfore, all parts are designed for as easy and simple manufacturing as possible. To help keep cost down I also try to re-use as many parts as possible.

Be as fire proof as possible without hurting form and function.
I want to be able to print when I sleep or when I am not at home. But I do not want to burn down the house. So all materials like pannels and structure will be fire proof.

Enclosed and insualted
I want to be able to reach a chamber temp of 80C. Why 80C? Because almost everything has a maximum working temperature of 80C. I.e. Rails, Fans, Electronics in general like to stay cool.

Cool and Clean looks
I want the machine to look good. I am no Voron designer, but I love the looks of these machines so I may borrow a lot from them. Including the glass front and sides. Mine will not be acrylic though (yikes!) I will use a dual layer (for insulation) of Polycarbonate (Fire retardent).

Klipper Firmware

I am done with marlin. I simply hate having to re-compile for even the smallest firmware changes. Klipper and reprapfirmware are so geniously made, that you can simply edit the firmware through the built in web interface – Using one of these two is a no brainer.

I have chosen to go with Klipper, as I really like the features, and webinterface they provide. It is also a fast moving firmware that is rigorously updated by the developers!

Control Electronics

The control electronics will most likely be two Fysetc Spider or BigTreeTech Octopus Pro boards. Currently I am leaning twoards the latter as they allow you to use any voltage for any driver.

The reason I need two of theses 8 driver boards are that this concept uses redicoulesly many steppers:

4 XY motors
4 Z motors
2 Tool rack Z motors
4 Extruders
1 tool holder

Thats at least 15 steppers. In addition to the steppers I also need 5 heaters and thermistors, and 11 end stops.

I want to use boards where I can replace drivers without replacing the entire board. Both in case of failure but also for experimenting.

I will do anything to keep the noise of the machine to a minimum when printing and it MUST be SILENT, when idle. Therfore, I have chosen to use 4 120 mm PC fans to cool everything (including the watercoled hotends).

Display

I was first not sure I even wanted a Display.. After all Klipper has a awesome webinterface that works on the phone. So Why?

But I had to remove the display from the printer at work (running klipper) and I now know.. I WANT a display!

Still not sure what I will use in the end but right now I am thinking of either a 10″ tablet simply running the Web interface or Klipperscreen running on a 10″ touch screen. I am leaning towards the latter as it does not require internet connection. Only downside is that it does not allow you to view your webcam (I want a nozzle cam).