Chaucer – A Vertical 3D Printer, Lego Compatible Parts

One of the print quality goals of the Chaucer Vertical 3D Printer is to be able to print snap-in LEGO® compatible plastic building blocks that are equal to or better in quality than LEGO parts. To be clear, the goal is NOT to reproduce LEGO parts, but to develop a 3d printer with a print pipeline good enough to high quality print parts that share similar interface characteristics with LEGO, but not copies of LEGO.

Today I was able to print compatible test parts that do indeed snap-in to Lego parts with an audible snap, which I might add, was music to my ears. The printed parts are also compatible with themselves. The quality is not there yet, but it will be when the print pipeline is fully developed and tuned. Nevertheless, I am pleased with the results thus far, and pleasantly surprised that Chaucer was able to print this level of quality, knowing the current print pipeline is far from being tuned.

Here is a photo of two of my test parts. The red part is the official LEGO® part, the black, LEGO compatible printed parts. The test parts were printed with 1.75 mm PLA filament.

Chaucer – A Vertical 3D Printer, Layer Registration, Flare Defects

For my first print using Chaucer, a vertical 3d printer, I chose the Nickel Calibration Test found on Thingiverse. The test print is a 3d model STL file of a rounded rectangular medallion with a Nickel-sized hole. A US Nickle should fit perfectly in the printed part, matching both diameter and thickness. My printed part had two glaring defects, a layer registration difference between printed layers, and flare at the base of the part. These are both classic print defects.

Here is a photo of the printed part.

If you look carefully on the left edge of the Nickel hole, you can see three or more stepped ridges. These are different edges of layered print, caused by driving the x and y-axis at too high a slew rate, causing the steppers to skip steps when rapidly accelerating and decelerating around corners and stops, causing the printer to lose it’s location, and subsequently print layers to print off grid. If you watch and listen to the video of the first print, you can hear ratcheting sounds as the stepper tires squeal when taking corners too fast. A Nickel would not fit into the hole with these registrations errors.

I temporarily fixed the registration error by slowing down all the default print speeds to 30 mm/s in the Slic3r configuration. Slic3r slices the STL file 3d model into horizontal slices, not unlike a CAT scan. Slic3r is where you set hundreds of critical parameters to characterize the print sample against the printer. Blindly setting all speeds to 30 mm/s was somewhat arbitrary, and far from optimization. At some later time, I will take a more engineering approach, and decide how I can print at the maximum speeds with the desired print quality.

You can see flare on all the edges of the medallion, a ramp of sorts, that spans both the outside perimeter of the medallion, and the inside perimeter of the Nickel hole. The edges should be square, but they are not. The flare is also enough to prevent the Nickel from seating flush the base.

Technically, the flare could be considered a design feature, and not a defect, as flare is actually printed on purpose, to get more surface area against the print bed, and subsequently better adhesion. The first layer, when the print head is closest to the print bed, is critical. If the first layer does not stick, the part will warp and move due to changing temperature gradients, and the print will fail, sometimes bad enough for the corners of the part to lift high enough to cause the print head to crash into the part. Not a pretty site.

To make the part stick on the first layer, we bring the print head close enough to literally squeeze the extruded hot plastic against the print bed, essentially flattening it. In this sense, 3d printing more resembles wire welding than classic printing.

I reduced some of the flare, by simply adjusting the z-axis end stop to raise the print head slightly higher than “normal”, which applies less pressure to the plastic against the print bead, resulting in less flare, and with it, less adhesion. It sounds simple, but it is not. If we raise the distance too high, we risk our first layer adhesion. Subtle changes produce radical results.

My print head normally starts about 0.10 mm above the print bed. I tried two additional heights, but first adding 0.1 mm, then adding a second 0.1 mm.

Here is a photo showing four print samples.

Starting from left to right:

  • Default high print speed, 0.10 mm height – layer registration error, large flare
  • Slow 30 mm/s print speed, 0.10 mm height – no layer registration error, large flare
  • Slow 30 mm/s print speed, 0.20 mm height – no layer registration error, medium flare
  • Slow 30 mm/s print speed, 0.20 mm height – no layer registration error, tiny flare

Please note, my nomenclature on the blue tape indicates the amount I raised the height from the default, not the absolute height.

The print on the right produced the best print quality, but I was not happy with the adhesion, so I went back to the default height of 0.10 mm, and decided the flare is acceptable.

Here is a closeup of the most acceptable print quality, going back to the default height, and accepting the flare as an acceptable defect.

Chaucer – A Vertical 3D Printer, First Print

Chaucer is a 3d printer modeled after a vertical milling machine, derived from Dave Kennell’s Pocket Printer. As in a vertical mill, the cutting tool, or in this case the adding tool, a plastic extruding print head, moves in the vertical z-axis. The z-axis is driven by a stepper motor and lead screw drive train, and can therefore support a heavy bulky load, making this printer suitable for experimenting with new multiple print heads.

Also like a vertical mill, the print bed moves horizontally on both the x and y axis directions, each driven by a stepper motor and timing belt drive train, making the table fast and agile, suitable for experimenting with extreme print speed and acceleration.

I completed my first successful print today. It’s not perfect, but easily fixed. Check out the video here.

There Be Tadpoles!

This evening, I attempted to print for the first time on Chaucer, my Vertical 3D Printer. Some things worked. Instead of printing a precision test cube however, it printed a few hundred tiny black whisker-sized Tadpoles. I’m actually pleased with the results. I have to fix the firmware, and continue calibration, but the fine detail and directionality of my accidental Tadpoles tells me this printer is capable of finer detail than I had initially hoped for.

If you look carefully at the reflection of the sky on the mirrored print bed, you should be able to see a school of tiny errant Tadpoles.