Fabr at Seattle Robotics Sept 20th

Just some quick notes:

Fabr will be at the Seattle Robotics meeting on September 20th. For more information about the meeting, visit http://www.seattlerobotics.org/.

I’ve been posting updates to my twitter feed instead of posting incremental updates. You can follow me via http://twitter.com/ooeygui.

I am working on a couple of major changes to the RepRap driver software, a post about using Programming Patterns with the Arduino, and more about the Fabr extruder. Stay tuned…

Stepper Motor Driver v2a

One of the side effects of using 80/20 for the structure of a 3d printer is that it weighs significantly more than the reprap design using rods and plastic joints. I became concerned that a 600ma Stepper may not drive that much weight or overcome the friction of the linear motion bearings. I started designing the ‘super driver’ which was capable of driving up to a 2.5 amp Stepper which would easily handle those issues.

After Fabr’s appearance on Hackaday, Zach from the RepRap research foundation contacted me about collaborating; we were going down parallel paths. I deeply admire the foundation’s work, and Zach’s contributions seriously rock. Needless to say I was very enthusiastic about this collaboration.

Zach was interested to see if we could use the ‘super driver’ as the next generation stepper driver board for the RepRap, and offered suggestions for unifying the projects.

The first iteration of this collaboration has been checked into the RepRap source forge project (and the fabr subversion project will be retired). It is looking pretty sweet!

Here’s the start of the new board:

Notable Differences:

  • License changed to GPL
  • Name change (no more super driver)
  • Uses the RepRap standard IDC header for communication
  • Uses the RepRap standard power supply connector (vertical instead of horizontal)
  • Min/Max end stops route through the motor controller using Headphone Jacks (to be implemented)
  • Uses the new RepRap .156″ headers for the motor connector (similar to the ones I was using, but non-removable and less expensive)
  • Stepper mode (full/eighth) is a jumper instead of software settable
  • Opto-isolated (to be implemented)
  • SR is jumper configurable (needs the diodes to be completed)

More on 80/20

Fabr is built using the 80/20 T-Slot fractional building system. ‘New Stock’ items were purchased from the 80/20 Garage Sale on eBay. The service was excellent. Just make sure when you purchase that you wait for the invoice as they combine shipping.

The Fabr frame is a 1″ profile ‘1010’, 12 total, each 14″ in length. I paid $2.75 each, for a total of $33.

I used 8 three-way nylon connectors to complete the cube – each was $2.85, for a total of $22.80.

(Similar to these 2-way connectors; couldn’t find a 3 way image at the time of writing)

$60 or so is very reasonable for the superstructure of the printer. I’d certainly use this system again.

The ‘linear bearings’ are ‘h’ shaped extrusions with a sandwich 3 ultra-high molecular weight shims between it and the t-slot extrusion.

I assumed that these bearings would glide smoothly over the t-slot. After assembly, I discovered that the bearings were very tight which required more than an acceptable amount of force to overcome the friction.

Disappointment set in. 

I attempted some minor changes to the bearing pads in order to reduce the size. Sanding nor scraping were successful (too much fuzz was generated causing more friction). However, removing a single pad reduced the friction to acceptable levels, but the loose fit could result in a loss of precision.

The bearings themselves were very expensive compared to the other components – $34 each x 3 = $102.

I also purchased an additional extrusion for the X-Axis, connectors and t-nuts.

The total with shipping was $192.90.

If I had to do it again, I would use the 80/20 super structure, but would investigate other options before buying the bearings. It seems that from time to time, the garage sale offers a ‘build your own bearing’ which may affect the final cost, but ymmv.

(naked bearing pictured here is $8. There is also ‘overstock’ UHMW material in the bearing profile for $7 at this time.)

 

I’ve been playing around with the idea of using an ACME lead screw with a traveler which not only engages with the screw, but also uses the screw as a guide rail (2 birds one stone). This would need to be fixed at both sides using a bearing block. (sorry, just idea at this point).

Fabr is a RepRap

I became intrigued by the concept of ‘renewable manufacturing’ – owning the life-cycle of everyday things. This idealism was captured by the RepRap project; who’s tag line is ‘Wealth without money’.

In the early phases of the construction of Fabr, I ran into the classic RepStrapping problem – how do you build a device intended to be built using a 3D Printer without a 3D printer? I had attempted to use the printing services at the TechShop in Menlo Park, but was unsuccessful. After that failure, I decided to designed and build Fabr using commonly available materials and few custom parts.

Over the last year, the RepRap organization has made changes independently which amusingly coincide with some Fabr design decisions, and Fabr has changed to be more like the RepRap in order to better leverage the software and firmware from the RepRap team.

In essence, Fabr is a RepRap.

Since my last post, I’ve been working on the following parts of the project:

  • The TextMate plugin was nearly rewritten in order to remove the Processing toolchain, and require the AVR MacPack from Objective Development.
  • Using the new Mill & Lathe, I’ve improved some of the articulation points
  • Implemented a multi-screw Y axis to compensate for unacceptable racking. I wish I could say that 80/20 is an asset, but the linear motion bearings are woefully inadequate and excruciatingly expensive. 
  • I smoked my motor shield while debugging a stepper problem, and switched to using 4 EasyDrivers from SparkFun.
  • Started building a ‘Super Driver’ which can drive steppers to 2.5A, as well as software configurable Full Step or Microstepping.
  • Ported the RepRap gcode interpreter to the TextMate toolchain and adapted it for the EasyDrivers (Need to unify this work with the trunk, and submit my updates to the RepRap team).
  • Implemented a Ruby gcode uploader; hopefully to be used in the Sketchup exporter.
  • Acquired materials and Building RepRap Opto Endstops for home positioning
  • Acquired materials and Building the Temperature controller for the extruder

Been awhile

It has been awhile since I’ve posted an update. I’ve been working on numerous projects, all at various stages. I figured I’d post an update across the board with no real conclusions.

Sources

Sources have been posted on:

svn co http://www.ooeygui.com/Arduino .

But now I have a Trac server installed at http://www.ooeygui.com/trac.

This server allows people to post bugs, read design docs, and see changes live. Enjoy!

TextMate Microcontroller plugin

While working on the gcode interpreter, I discovered some issues with the plugin. I’m in the process of rewriting portions of the plugin. The plugin now uses AVR MacPack instead of the Arduino packaged compilers. This isolates the plugin from breaking changes by the Arduino developers. I’ll release new packages as new Arduino packages are released.

Fabr Hardware

I had completed the Axes, but noticed a nasty binding problem. After disassembling, I discovered part of the drive assembly wasn’t square, which caused a bind when spinning. A second problem that was bugging me was related to an early cut on the X axis not being square to the Y axis… Both are being corrected as I build the extruder mount.

The extruder is completely built, but untested. I leveraged the extruder head that was featured in a previous post, added a cooling pipe and an insulated feed pipe in order to prevent premature melting of the extrusion cord. I’m currently constructing a mounting plate which will allow me to swap print heads relatively quickly (as well as allow me to mount the current one).

xstrudr

I have 25lbs of bioplastic granules; not terribly useful for a printer. I had considered building a hopper on the printer and extruding the granules directly to the print job, but couldn’t make it practical. Instead, I’m going to have a separate unit which processes granulated plastic into a cord, and yet another unit for grinding plastic into granules (‘chipr’?)

Motor Controller

After the experience with the triple axis controller, I’m starting to design a single purpose motor controller; I find that I need 4 controllers – one for each axis and one for the extruder. Here’s a preview: .

I still plan on having an Arduino Shield to tie 3 or 4 axes together.

First Extrude

It isn’t a printer until there’s melted plastic. Fabr has finally reached that milestone – An extrusion head which melts plastic:
Unfinished Extrusion Head

For this test, I was feeding by hand: I was amazed at how little pressure was needed in order to get a decent feed rate from it. The Kapton encased nichrome wire drew about .9 amps at 12 volts, and took about 30 seconds to heat up. I used screws and wire to secure the heating coil.

The Feed itself uses two pulleys, directly mounted to a motor. Given how little pressure was needed to drive the print head, this should work out quite nicely:
Feed

I also completed the y-Axis:

Next steps:
Do the xAxis – this should be the easiest of the bunch. I only need to fabricate a mounting plate, then mount it, the motor, and the flanged nut to the extrusion carriage. I already built the coupler.

Finish the extruder head – I need to cut down the bolts, and wrap the whole thing in an insulator. The fed plastic tended to mushroom at the outlet port as well as get mushy while in the barrel. I’m going to use a hollow rod (maybe graphite? Steel?) wrapped in fiberglass to insulate the feed barrel. Hopefully that will prevent it from loosing integrity before being pushed into the outlet.

Y Axis, Extruder, Controller and Software

Y Axis

Shortly after building the Z axis, I began working on the Y Axis. I was being cheap and wanted to attempt to use a single drive screw instead of two (as planned) thinking that the 80/20 would be true enough to prevent racking. Once I actually hooked up the cross beam I found the racking to be pretty extreme.  To counter this, I needed another drive screw and a timing chain instead of direct linkage; This meant another order from electric gold mine.  I had decided to use chain instead of a timing belt for cost reasons. I may revisit this in V2 as it adds quite a bit of weight and is really loud.  I should be able to get the Y Axis built this week and the X axis shortly after that.

Extruder

I had some problems attempting to use a screw to feed plastic into the extruder. I tested a feed mechanism which uses two pulleys and was impressed with the results. In order to get a screw to engage the plastic quite a bit of side pressure is needed, which in turn requires quite a bit of torque to drive the screw. With the dual pulleys, a smaller stepper motor can be used because very little pressure is needed, which means less torque is needed for the same extrusion feed pressure.I was attempting to work with Clayborn Labs to design a heater barrel. I need to follow up with them because my account seems to have gone into the ether… In the mean time, I’m going to be working with the electric gold mine “NASA heat tape”. 

Smoke test passes!  No smoke is passing right? I ran some simple tests designed to test each motor controller. I used LEDs instead of a motors because I didn’t want to risk destroying them. It was wild to see the coil animation represented by lights. Next up is testing the end stops and building an on-board gcode interpeter.

Software

I ran into some bugs in the TextMate plugin, so corrected them. You can download the new version and unzip to ~/Library/Application Support/TextMate/Plugins. The Sketchup plugin is successfully decomposing objects into voxels. I’ve begun work on a recomposer which attempts to use the voxels by using various ‘solid handlers’ to attempt to generate sub-volumes. During this process I noticed that I was trying to locate the same edges I used to generate the voxels – and lamented that it may not be the best solution. In the degenerate ‘cuboid’ case, using voxels is dumb. However, as the complexity of the object increases using voxels will deterministically reduce the original volume to sub-volumes without requiring edge walking.I had wondered if using octrees would be better for dividing the volume into cuboids, but I think it may loose some context when adding other shape composers (like finding holes or cylinders)

Bioplastics verses petroleum based plastics

I am building a 3D printer, a sister project to the self replicating printer project RepRap. A 3D printer is capable of building usable objects, such as toys, by ‘drawing’ one layer at a time, until the whole object has been constructed.

While researching plastics for this printer I came across a dilemma – Which is better Bioplastic or Petroleum based plastic? Bioplastics are derived from corn or potato starches. ‘Petroplastics’ on the other hand is derived from crude oil.

One of the goals of mine is to develop an 3D printer ecosystem which enables the direct recycling of plastic – thus keeping plastic out of the land fills. Until this ecosystem is complete, I will need virgin material to work with. So the question remains – Which is “better”?

Defining Better

Before we attempt to define which plastic is better, it would be helpful to have some metrics for comparison. Like many things there are a continuum of variables, and for each variable there is a continuum of values. For the purposes of my interests, I’m going to attempt to answer:

  • What is the environmental impact of the production or extraction of the base material for the resin?
  • How readily available are the resources needed to produce the resin?
  • How does the use of those resources for the production of resin affect the use of those resources for other means?
  • How long does the resin take to be reduced to base organic materials?

Extraction/Production

Petroplastic

Petroplastic is derived from crude oil. During the refining process, hydrocarbons are extracted from the oil to produce various everyday chemicals – octane is extracted and refined to produce auto fuel, nonane and hexadecane to produce diesel, kerosene, which are used in making jet fuel. In fact, Wikipedia lists dozens of chemicals which are used in most of the products that industrial societies have become dependent on – petroplastic is just one product derived from it.

Millions of years ago, places which currently have abundent reserves of oil or natural gas were lush rain forests with rich healthy ecosystems. The surface of the earth has changed via plate tectonics – which is why oil has been found in places like the Arctic or underwater.

As plants and animals die, they decompose by natural processes – by other animals, insects and eventually microbes. The raw material is used in a successive generation to form the building blocks of plants which animals eat and so on. If this process is interrupted, either by a change in climate or other catastrophic event, the organic material will eventually end up in underground reservoirs. Heat and pressure causes these compounds to combine into different types hydrocarbon chains. The variance in temperature, porosity, and pressure in these reservoirs determines how it achieved the state we see today – Oil Shale exists where rock is very porous; areas of high heat and pressure have abundent natural gas; and somewhere in between is where “sweet crude” exists.

The energy density of oil is about 40 to 1 – a simple example is it takes 1 gallon of gas to extract 40 gallons of gas. This fact is why oil is so attractive as an energy source.

Thermoplastic is not one thing – there are many different types of thermoplastics for lots of different uses: High Density polyethylene (HDPE), Low density polyethylene (LDPE), Acrylonitrile butadiene styrene (ABS), Polyethylene terephthalate (PETE), Polytetrafluoroethylene (PTFE), Polyvinyl chloride (PVC) – just to name a few.

Plastics also have additives which adjust their properties – some making them shatterproof with Bisphenol-A, some stabilizing the resin with lead or cadmium.

Bioplastic

Bioplastic is derived from plant starches, usually corn or potato. The endosperm of corn, the center white in the kernel, contains the starch which is extracted by soaking, grinding and washing. This is then combined with plasticizers (sorbitol & glycerine in varying amounts) to produce a thermoplastic.

When harvested, the left over material (husks, ears, stalks) are composted and returned to the soil and supplemented with fertilizers.

It has been difficult to find information about the inputs to corn growing for the production of bioplastic; I’ve started looking into information about ethanol production. According to T.W. Patzek at Berkley, siting several studies averages the net energy production at +280 BTU. I’m intrigued by one study which also includes the energy consumption of farm equipment which concludes a net negative energy output of 33,000 BTUs! This means it takes much more energy to produce a unit of ethanol than the ethanol can provide to cars.

Environmental Impact for Production

Farm technology is currently fossil fuel based – from the fertilizers used to enrich soil, to the farm equipment used to prepare the soil and harvest the plants, to the machines that post process the plant material into plastic.

Availability

As mentioned before, oil was formed after millions of years in ideal conditions; trapped in underground fissures. But how much exists?

Hubbert’s peak refers to a study done by geophysicist M. King Hubbert in 1956 which predicted peak production of US oil in 1970; his prediction was accurate. In 1969, he predicted world wide peak oil production in 2000. In Kenneth Deffeyes’ book “Beyond Oil”, he suggests we are in a post peak condition now. Edmunds.com – yes the car buying site, as a fantastic summary with these words of warning:
1. Oil is not an infinite resource.
2. Remaining supplies of oil should be used wisely.
3. Alternative sources of energy need to be brought on line soon.

Civic Actions has a neat visualization of the remaining world oil.

Conversely, corn requires clean water, fertilizer and harvesting equipment all which require oil today. As oil becomes more scarce, corn production will become difficult if not impossible – unless alternative production techniques can be implemented.

Side Affects

Plastics resin is a byproduct of oil – its production has little impact on other uses of oil.

Does the production of bioplastic have an impact on corn or potato for consumption or feed stock? Using the ethanol analogy, it may have minimal impact. According to the renewable fuels association concerns about the impact on food are unfounded. In fact the correlation between food costs had more about scarcity of oil rather than its use in biofuels and bioplastic. It appears that much of the land devoted to oil alternatives were unused until recently.

Decompose

Petroplastic ‘photodegrades’ in sunlight. The ultraviolet light from the sun causes polymer chains to become brittle and crack under stress. Unfortunately, natural microbes responsible for natural decomposition do not recognize these polymer chains as food. Work is under way to engineer microbes which consume plastic (green goo?). A company in the UK called “d2w” has a product which enables plastic to degrade without sunlight. Without details, I presume that the plastic does not degrade to base organic compounds – it works by making polymer chains brittle without ultraviolet light.

Simply put – petroplastic will not decompose or compost – it simply breaks apart into smaller plastic pieces which are unusable to life as we know it.

UPDATE: June 2009 – This statement may not be true. Apparently a two students have bread a microorganism which can consume plastic (via mnn). I will follow this closely.

Bioplastic, being polymers of glucose (specifically amylose and amylopectin) are considered food by decomposing microorganisms – no special magic or unnatural remedies required.

Which is better?

Which is better – bioplastic or petroplastic? Given how much oil is used in the production of the crops used to make it would appear that petroplastics are ‘better’ than bioplastics.

As technologies are developed to replace oil in farming equipment and transportation, bioplastics will soon eclipse petroplastics as the environmentally sound choice.

I for one will support the development of Bioplastics.

From the Experts

A comment was posted to my polystarch post which seems to be from a pro-petroplastics company:

From D2W – Degradable Plastics

“Plastics made from crops, are at least 400% more expensive, they are not strong enough, and they emit methane (a powerful greenhouse gas) in landfill. Also, it is wrong to use land, water and fertilisers to grow crops for bioplastics and biofuels, which drives up the cost of food for the poorest people. See http://euobserver.com/9/25831/?rk=1

The same applies to growing cotton or jute to make durable bags. These rapidly become unhygienic and become a durable form of litter, but they can be made from oxo-bio plastic to last up to 5 years.

Paper bags use 300% more energy to produce, they are bulky and heavy and are not strong enough. They will also emit methane in landfill

However, ordinary plastic and recycled plastic can be made oxo-biodegradable.

This is done by including d2w additive made by Symphony UK, which makes it degrade, then biodegrade, on land or at sea, in the light or the dark, in heat or cold, in whatever timescale is required, leaving NO fragments NO methane and NO harmful residues. Oxo-bio meets American Standard 6954, and there is little or no additional cost.

Oxo-bio plastic is made from petroleum, but this is an advantage, not a disadvantage. Some people would like to stop using petroleum to fuel cars and planes, but until that day arrives it makes sense to use the by-product of oil which used to be wasted.”

From Plastarch material of North America

PSM is definitely price competitive to petroleum based resin, and in some cases less costly.
The strength may be suitable for some applications and not suitable for others.  Not all petroleum plastics are “strong enough” for all applications either.

In most landfills, NOTHING degrades, not newspapers, not banana peels, and not PSM.  Therefore, if it does not degrade, it will not release anything.  Conversely, if it does degrade, it will not release any more gasses than food waste such as a corn cob.

Unlike other biopolymers, PSM does not use corn intended for “people food”.  Our stock is made specifically for PSM and is not affected, nor will affect, the global corn prices.

The additives on the market that can be mixed in with petroleum plastics to make them “degradable” are beneficial in extremely limited applications.  In short, they break-apart (degrade) the product, but the plastic particles are still present.  For example, the physical “bag” may be gone, but your lawn is now infused with plastic particles which will take tens of thousands of years to degrade or be eaten by animals and harm them.  In many ways, it is harder to clean up the particles than the bags and bottles if they were intact.  These additives are often marketed with a time lapse slide show showing that the product “disappears”.  What they don’t show you is the toxicity in the soil afterwards.  Again, since most landfills are designed for nothing to degrade, anyone making a claim of “landfill degradation” is probably bogus and needs to do more homework on what actually is happening in landfills.

In the final analysis, it is well known that petroleum is not a sustainable or renewable resource.  The Earth will run out of it, so we need alternative technologies.  The CO2 released from petroleum products is carbon that has been stored since long ago.  The CO2 released from the degradation of PSM is carbon that has been recently pulled from the atmosphere so you have a much more healthy cycle.

The petroleum companies really have nothing to worry about because there will be plenty of demand for petroleum in the foreseeable future.

Elevator

My son and I put together the z-axis today. Watching a 4 year old use a hack saw was amusing – I kept it safe and he felt like it was the best thing in the whole world.

The elevator is supported by 4 screws, using flanged nuts connected by L shaped aluminum extrusions. The screws float between the upper and lower cross beams – RC car bearings prevent them from binding. Sprockets and chain from Electric gold mine tie it together.  In order to tension the chain, I added an additional sprocket which I can move along another crossbar.  I designed the z-axis so that I can replace the print surface. 

elevatorthumb.jpg

elevator2thumb.jpg

Extruder Head

I spent quite a bit of time the last couple of weeks working on a feeder and heater for the extruder. I figured this was going to be the hardest part of the build; and was right so far.

My first attempts at feeding the filament were with a screw. It turns out that the filament is too slippery to grip with any certainty. This may work for granules, but for filament I think I need something else. One design I’ve seen for commercial extruders uses a feed roller. From sdp-si.com, I ordered a feed and grooved idler. Hopefully the urethane roller will engage the plastic better than the feed screw.

I’ve been experimenting with various methods of heating the extruder head to melting temperatures. While I managed to not catch the house on fire or electrocute myself, I failed to achieve a decent heater.

Electric Gold-mine was offering a “NASA heating element” on sale. It has micro-coiled nichrome wire in an electrically insulated package. In the process of determining what this device actually was, I found Clayborn Labs – a company that makes heater tubes and tapes. The configuration of the heater tube is perfect for the extrusion barrel. I’ll be working to finalize specifications this week and will be ordering a 1 off prototype along with a thermostat – both together are actually less expensive than what I’ve spent so far on the extruder barrel!