Hardware primer hosted by Armanicreations and Vescarobotics was a great success

HARDWARE primter

On Feb 3, we helped the DC robotics group host its first ever robotics hardware workshop.  Thanks to PUNCHROCK in Adam’s Morgan for hosting us as their first technology outreach event.  PunchRock is a social entrepeneurship space.

The session consisted of 5 learning areas.  These included 3d printing, motor control, tapping, joining parts, and total hardware assembly.  For the next 1.5 hours, over 30 guests switched between the stations taught by Michael Armani and David Jones. At the end, a summary was given to the group explaining how to make a robotic prototype that matches the scale and stage of the endeaver.  As a result of these learnings, people got the best of our many years of experience and 10′s of thousands of dollars spent on trying different hardware strategies in our own product development.

The presentation is included here Punchrock presentation 2003 format.

Thanks to all that came!  Everyone was very smart, asked questions that were spot on, and learned some quick methods for robotic creations.  We are honored to share our knowledge with this group.

Printing with a 1 mm diameter nozzle.

2012-11-06 17.09.40-2-crop2

FYI – The above picture is a 1mm printout.  It weighs 35 ounces (a measure of it being quite solid) and took only 20 minutes to print!!  It is very hardy, and this is because the infill is more sturdy.

The idea of reprap is that you can print excellent resolution by relying on highly accurate steppers to form a perfect outline, and filling in the center.  This makes it possible to use a 0.5mm nozzle to create and object that is say, 2.3 mm wide, and to fill it perfectly, because the steppers have 10 micron resolution (assume no backlash).

So…why not 1mm?  Or more?  Why is the push towards ever smaller 0.25mm nozzles smaller nozzle sizes (see link to makergear site with nozzles)?  Well, 1 mm is a great solution that makes stronger, faster parts, with virtually no disadvantage!  And 3D printing is SLOW, but within the realm of being really useful.  So any speed improvements here are a big deal.

 

Procedure:

I succeeded at this mod – by drilling out the prusa makergearv2 nozzle (plastistruder) to be 1mm diameter (order exact drill bits online). Then smooth the nozzle by brushing it against a flat brass surface (acts as superfine sandpaper).  There is nothing shockingly new here, but the take-home-message is that it was totally unintuitive before doing the mod how much easier plastic would flow, and how much stronger and faster the parts print.  Always challenge the status quo, it doesn’t know why it exists.

Some assumptions/requirements

  • You will need to have optimized the pololu RAMP output for the plastistruder to maximum amperage for maximum force
  • You need to balance the force on the filament – too much and the bearing roller won’t turn. Too little, and the filament won’t move.
  • In my case, I drilled and tapped the plastistruder so that i could use screws with stronger springs and to be able to remove them more conveniently.
  • You need to understand that for PLA, the filament does not behave normally when it has been sitting hot in the hot end for more than 5 minutes. Even if it doesn’t flow out, it will get “sticky”. Once you push it through (about 50mm filament) you will not believe how effortlessless filament flows through.
  • You should know that for 1mm, even though 4 times the material can go through, there is exponentially less resistance to flow through the nozzle, so the plastistruder can keep up. Also, technically, the layers only print twice as fast because slicer prints based on diameter, not diameter^2. I’m not quite sure why that is.
  • You will print about 1.15 mm in reality.
  • For a strong part, you have to let slicer overlap
  • I’m also assuming you optimized firmware for max acceleration and 350mm/s between print moves (prevents ooze)
  • There are some obvious modifications for slic3r. I am printing around 55mm/s for the part linked here. I also print onto cardstock paper to help set the first layer.

Advantages of 1mm as compared to a 0.5 mm nozzle

  1. Because more comes out of the nozzle per second, your prints will be more consistent on the first layer, and stick well.
  2. Less nozzle ooze (wow!)
  3. Infill
  4. Stronger parts
  5. Much (MUCH) faster prints. I am printing 3 ounces of plastic per hour right now…way faster.
  6. Ability to do 0.5-.75 mm layers with success
  7. Fewer “stringy polymer hairs” loose between distant prints
  8. Much cleaner top surfaces (overlaps between lines look better)
  9. Bridging seems slightly more reliable.

 

Disadvantages

  1. You will loose sharp edges (slightly). This is noticeable (and may be an advantage)
  2. Minimum resolution is 1.15 mm (instead of 0.55 before) is somewhat noticeable, especially if you manage to print 0.5mm layer thickness. However, resolution is not really lost much. The reason is you can still print 0.1 mm layers. Your steppers also have 10-20 microns of accuracy, so the ability to print lines instead of individual points (like an inkjet) really allows you to potentially do 2, 3, even 5 mm diameter if you wanted, you just need thicker objects to print a small hole. Also, if really needed desperately small feature sizes, you can rely on your Z-axis for accuracy (like you do now).
  3. When you get “stringy hairs” coming off the nozzle, they will be thicker and harder to remove
  4. Anything else obvious I’m missing?

 

Ideas for new filament printing materials

Filament

Ideas for new filaments have been brewing at ArmaniCreations for almost a month, since the realization that 3D printers could make fully functional, commercially competitive products in the future, but that it would require a bundle of new materials.  To date, the majority of sellers only carry ABS and PLA plastics.  A few resellers carry PVA (a dissolvable support material) and even fewer carry polycarbonate (PC) or polyethylene (HDPE) because of printing challenges.

So, Michael posted on the reprap forums to see what the reception would be like for some new materials, what other materials are desired, etc.  Surprisingly, conductive filament and magnetic filament, which could be used to print objects like functional motors or speakers were not even mentioned after the initial post.  Reprappers were mainly interested in printing rubbery materials, heavy-weighted polymers, and alternative materials like Delrin.

For each of these materials, there are obvious technical constraints, and safety considerations.  All of the filaments proposed by ArmaniCreations assume a 1.75 mm filament and melting temperature typically between 180 C and 250 C.  This is specifically aimed at the filament deposition modeling printers, where it is possible to simply exchange one polymer for another by changing filaments or using a multi-extruder head.  The development of new materials has extremely exciting prospects for RepRap.

By contrast, printers like the B9Creator can only print one or more light-cured polymers.  While it is possible to light cure rubber, generally these materials are brittle, have little resistance to the elements, and can not be used as functional parts in a final product.  More generally, they are best suited to 3D printed models.  On the other hand, their resolution is state of the art, and you can not easily see layering lines with their prints.

Easy Prusa Mendel Leveling Calibrations, Optimization for makergear V2 Prusa mendel

Prusa

In a sept 14th RepRap Post, Michael suggests the use of paper in place of a heated bed for printing, a new way to even the printing bed base with thumb screws, modified firmware for greater speed, and a “faked small nozzle” to allow better bridging and tighter plastic bonds.

We are honored to have inspired Rick at Reprap to modify his printer based in part on our design.

 

Details from Michael’s original post are copied below:

I made some big changes to improve 3D printing out of the standard domain of ~25% fill to 100% fill. I print almost all my parts in 100% fill, which is tricky, because if the fill is off your parts can warp inward or outward. Also, as the nozzle goes over for multiple close passes, it tends to “bunch up” or collect and dislodge previously laid down plastic. Also, full fill was very slow before.

To solve these problems, I started with the base alignment.

EVENING THE BASE:
The makergearv2 comes with a screw set with springs and many nuts to help calibrate the base. This is a great start, however, because the base is wood and wasn’t threaded, it meant that turning the screw for calibrations would squish the nuts that held the screw in position around the wood. What you could do was turn the nuts holding the springs in place, but it was very difficult to get in there and turn small screws, plus it only gave you about 100 microns of alignment space because the screw was weak. To fix the base, i tapped two pieces of hardwood with 1/4-20. This is tight enough to act as a loose lock-nut – the friction prevent dislodging from stepper vibration. Meanwhile, I upgraded to fully long springs that have about 5 times more force. Next, I use thumb screws for easy access. I am able to use a piece of paper to test the friction between the nozzle and base at Z=0, and can calibrate the whole device in seconds. Also, I applied flat glass ($2 from home depot) over the PCB and uses hot glue and two paper clamps to hold it in place. The glue needs a few heat-cool cycles to equalize (otherwise the center bows upwards about 0.2mm).

PATTERN BEGINS IN PAPER:
Next, I recognized that no matter how hard you try, sometimes getting a pattern started requires tricks, temperature changes. Instead of tricking the print to extrude more in the beginning, or modifying temperatures, I use a piece of paper clamped to the base. This allows all extrusion to embedd the paper, which sucks up the cheap chinese PLA like a sponge (even at 185 C). Also, because the paper absorbs by wicking/surface tension, it keeps the nozzle head extremely clean and it doesn’t bunch up or collect old prints for the first layer or two.

MODIFY FIRMWARE (5:1 SPRINTER)
The basic idea of 3D printer now is keep the nozzle cool, and print slowly, otherwise you’ll pickup and distort previous prints. To prevent this, and ooze from forming, you also retract the filament, and Z step up and then back down to prevent ramming into/crossing over a perimeter from the same layer. What a pain!! And so much slower. So in my design, I speed up the firmware so it accelerates very quickly – 4 times faster than the default settings. I also ramp up speed to 350mm/s between printing areas. This allows you to turn up the print speed so fast that between extrusion areas, the nozzle doesnt have time to ooze. That means if you go over an existing perimeter, it doesn’t have time to scrape it off while it is hot. It also means hotter temps are okay, and thus, welding of plastic fully. I also eliminated the temperature checks to allow the Printrun software to set the temps and the firmware ignores any settings. Also, set max speed to 350mm/s, print at 60mm/s. My default layer height is 0.1mm, and don’t worry, its FAST enough.

FAKE SMALL NOZZLE:
Use slicer 0.9.1 to tell the nozzle it is 0.4mm OR set default extrusion width to 0.4mm. This way, you get multiple passes. At low extrusion temps, the nozzle puts out a line thinner than the 0.5mm width of the nozzle. Therefore, faking a smaller distance closes the gap and allows better overhang formation.

 

The Future is even cooler than you imagined! The 6 “Holy” Pillars of 3D Printing!

screen-shot-2012-08-03-at-1-13-29-pm

The origins of RepRap, as suggested by the wikipedia article, are to make a machine that can mostly print itself.

Mostly print itself?

So why stop there?  Well, three of us got to thinking and quickly realized that, while it is far fetched, there is no reason why a printer couldn’t fully reproduce itself.  However, the requirements are fairly staggering.  Today, only structural plastic is printed, and that is good for some prototypes and functional parts; however, products like cars, houses, bikes, and toys, need more function elements like non-porous metals, conductive wiring, printed electronics, magnets for motors, rubber for tires, and more.  In addition, today’s printer resolution would need to be improved to allow for the re-invention of micro-circuits to be printed in an additive process from raw parts.  Also in addition, full color printing would be desired.

The 6 Pillars of a full 3D printer include:

  • Structural Plastic
  • Stuctural Metal
  • Structural Magnets
  • A Soft, Rubbery Material
  • A Conductive Printable Material
  • Full Color Coatings (possibly including ‘clear’ to simulate glass)

From this list, one could see that it is possible for printer technology to grow, organically, into something that can print VERY rudimentary but fully functional parts.  Take a stepper motor for example; it requires a heat-dissipating shell, a strong structural core and shaft, conductive low-resistivity wiring for electromagnetic forces, insulation for the conductive wiring, and solid magnets to interact with the electromagnets.  However, surprisingly, only two materials are needed to print this: structural plastic, which also acts as insulation, and conductive polymer, which can be printed with ease within the part.

A 3D printed stepper motor would be weak; because of the high resistance of the wires, large cross-sectional areas would be needed and maximum hold strength would be weak.  However, even early models might boast extremely low cost and high power-to-weight ratios that may be desirable.  It is then possible to see that with moore’s law type scaling in the industry, improvements in materials, speed, and costs could make this “all inclusive” 3D printer extremely competitive with existing manufacturing practices.

In fact, if you could print a car for example, you would be able to have a better product, not an inferior one; for example, you can create dimpling in the side panels of the cars tht only special racing cars use to reduce drag today.  In this way, in the future, many consumer products, such as houses, cars, and toys will become 3D printed, once the speed is fast and costs are low.  This would be a revolution bigger that the internet and computer itself.

If that scares you, never fear – an idea is emergent, and its going to happen whether you want it to or not!  While this technology creates many concerns, such as protecting jobs, patents, and dangerous weapons from being made, it is also extremely liberating, allowing people to do more with less.  In fact, in the future of full 3D printing, the dominant highly-paid trade might become artistic engineering, because the only distinguishing factor between manufacturers is the design of a product, and not so much its cost or functionally.

Also rest assured, we are talking decades.  However, we feel that this technology applied as manufacturing for the masses, a “replicator” in effect, is inevitable.  Oh, and its already happening:

A (mostly) 3D printed car
A 3D printed house and prototype
A nearly fully 3D printed bicycle! -see video at the end.

For additional discussion, please post to the related reprap post.