Wednesday, June 15, 2016

"IvorBot", the early days...

I found some short videos of early days testing of my core-xy movement (stacked, fishing-line, rectangular format). Worth sharing I thought, as even a short video shows so much more. In the first clip you see it plot, probably doing about 40mm/sec. It was my first experience of a core-xy movement and was pleasantly surprised at how precisely it plotted, and how accurate the repetition of the same drawing was. Fitting a pen holder to an x-carriage is always a practical way to get an initial feel for movement and dimensional accuracy, in x/y directions at least. See video below.
(In the audio you can hear the poor quality groove bearings beginning to grind, which was it's ultimate downfall. The stepped nature of the circle is just the way it was drawn.)

Driving it at higher speeds didn't help the failing bearings, but you could really open the throttle on this chassis. Having no weight at all on the x-carriage helped. In this other short video clip (below) you can see the movement running at faster speeds. I used an old Gen6 board to drive the experiment, which it did without issue once I updated Marlin firmware for the core-xy support. The tests were controlled from Repetier Host. If you don't have an extruder fitted and want to trick the printer into working you need to set a dummy thermister "998" in Marlin Configuration.h, and un-comment #define DUMMY_THERMISTOR_998_VALUE 185". You can then slice something and send the gcode file, which it will pretend to print, or plot if you have a pen fitted, and no Z mechanism.

I learned a lot from this 'prototype'. I was able to draw on my existing body of knowledge and understanding of 3d printing, and build on that to construct this experimental chassis. The plans were drawn up in Sketcup primarily, which I use quite a lot for design visualisation. With some plug-ins I can export for printing (STL), and export for laser cutting (SVG).

I've also found Autodesk 123D very good also, and move to it for more complex components.

By building the physical working model I was able to get a true sense of scale, construction challenges, and mechanical performance. I was then able to modify design and rapidly improve through iteration, swapping out both printed parts and laser cut framework. I'm constantly keeping an eye to my primary design goals of compact size for given print volume (200x300x250 approx), along with dual extrusion capability with room for experimentation.

More detail of the journey to follow.
Tks for viewing.

Sunday, June 12, 2016

Code named "IvorBot"! (Part 1)

I've been lax in blogging of late. Sometimes it seems there's always "just one more thing" to do or make or print before a new blog post is worth writing. Then the weeks and months pass and the motivation to post historical work just isn't the same after a time. Lets see if I can do a bit better on the blog front going forward!

I'm always keeping an eye on 3d printer developments, innovations and direction, as things continue to evolve at some pace. Digesting these influences, I've regularly found myself subjectively noting the 'desirable' and 'best' features of the changing printer vista, with one eye, as always, on building my next printer.

A design is always given focus with some constraints, and for my "next printer" I imposed some goals that formed the general basis of the prototyping and experimentation that followed, for some time now. Key enhancement objectives included such goals as faster print speeds without loss of print quality, dual extrusion, moderately larger print volume, but minimum increase in outer printer dimensions. Secondary goals might be a flexible carriage system to allow for easy print head swap-outs for greater flexibility.

The choice of movement system and filament feed mechanism was left open in the early research. Also, I wanted to explore and experience bowden vs direct drive, 3mm vs 1.75 and the pros and cons of these now common machine choice variables. My new printer would give greater opportunity for testing and learning, reconfiguration for experimentation.

The attractions of CoreXY has stood out for some years, and has enjoyed much adoption in both experimental machines and commercial printers of late (FABtotum, Airwol3D Axiom, SmartrapCore). I quickly settled on this as key design feature and pushed on with some basic plans.

Armed with Sketchup, my trusty Mendel90 and access to a local Fab Lab laser cutter I was quickly able to stand-up a basic core-xy chassis. For a fast and relatively accurate movement system I also took inspiration for RichRap's use of Spectra line as in his Sli3RD Printer, and even Nicholas Seward's CoreXZ printer. The following fishing line CoreXY "plotter" was put together (Photo below.).

Using fishing-line allowed compact and flexible "belt" paths. By routing the line over small groove pulleys it was possible to position the motors below the x/y plane resulting in space saving and more compact external dimensions. (Illustrated in photos below)

Access to a laser cutter was transformative in terms of flexibility of chassis design, manufacture and build speed. It allowed me to iterate quickly, and a very low cost. I worked in Shetchup, exporting to SVG via a plugin for laser cutting comparability.

Above is one of my printer frame work revisions being cut from cheap plywood on the Trotec laser cutter at Fab Lab Limerick. Back at home (below) the next revision was quickly constructed and motion testing resumed.

The prototype printer, pictured below (actually just a plotter at this stage), made many trips to the Fab Lab - Thursday evening sessions over the Winter, for peer review and valued feedback. It was during one of these visits that it earned it nick name, the "IvorBot", which has stuck!

Testing with the fishing-line core-xy chassis ran it's course. It was functional and with the addition of an extruder might even have done some printing, but it didn't have the robustness I was striving for. Also, my choice of cheap grove bearings was a mistake and they were grinning and failing as I'd push the speed and tension on the fishing-line. 

The next major revision of the design was a move to belt drive, and brass pulleys. I designed and printed a whole new set of motor mounts, corner brackets and x-carriage ends. I chose 9mm GT2 belt, matching brass pulleys with double inserted bearings, LMS8UU linear bearings. I ran a "stacked" cord-xy belt path so there was no cross-over in the design. I felt the wider belt and pulleys wouldn't lend itself to the cross-over belt path that typified some core-xy layouts.

Another frame was cut, this time with more z-height, and when E3D started selling the 200x300 heated bed from their BigBox printer separately, I jumped at the opportunity and matched printer frame size to accommodate this larger print bed.

Above, "IvorBot" current revision. It prints reasonably, but there were many changes to get it to this stage as you might appreciate. I recently got it tuned up for dual colour printing, which was a new personal mile-stone, and gave me the incentive to resume some bloggin. There's lots more detail to share, and much planning going into changing and refining my design, but for now I'll sign-off and get this much published.

As always, comments and questions welcome!