RepRap: Builders

The linear actuator stepper motor that I acquired arrived today. I slapped it on to the Tommelise 2 Big Board and it ran immediately without complaint.



What you see here is the board cycling it backwards and forwards for 240 steps with a 7 msec interval. It transitions at a rate of roughly 14.5 mm/sec at that setting.

In that it is like the other tin can steppers I've been working with. It can be made to run as fast as 17 mm/sec, but not reliably. A 7 msec interval is the fastest setting that it can operate at and deliver substantial force with no skipping.

The really amazing thing about this tin can stepper, however, is that it is dead silent. As you can recall from other videoclips I've posted the microphone on my camera is incredibly sensitive and made previous steppers and gearmotors I've taped sound very loud. This one you simply don't hear at all.

Mind, that is double what Darwin is currently achieving and means that it can print a maximum of roughly 0.8 kg of plastic in 24 hours.

(No additional content at 3DReplicators.com)

I've done a bit more work on my Gerber (PCB) plotter. It now has a GUI! The script will show all the lines to be plotted, then colour them black in real-time as they are plotted.


I've also finally made myself a Y axis so I can now lift the pen.


The resolution on my repstrap is a little limited so i'm having to use quite large tracks and pads.
More photos on flickr

Once I got the x-axis circuitry wired and going it was a matter of about ten minutes to wire up the y-axis.


Mind, I'm just meatballing the wiring, so it looks really nasty. Works, though. I guess that I'll have to run this through Eagle and do a PCB of this when I get it to working prettily.

(No additional content at 3DReplicators.com)

Working on Reprap has had a distinct picaresque feel to it. We ride off armed with our soldering irons and side cutters determined to take on and vanquish the whole factory/manufacturing establishment as we know it.


I wonder at times if this isn't the likely outcome of our efforts. At times, though, I find myself frightened that we might just manage to accomplish what we have set out to do. What we are doing has the potential to change the world every bit as much as the internet has.

(Read the whole story)

When I built the small test board for the 18F4550 I attempted to reduce the number of i/o port pins that were required to drive a stepper using the SN754410 quad half-H driver chip. If you just assign an i/o port pin to each of the equivalent pins on the 754410 you need six; two pins to control the two phases in the tin can stepper and a pair of pins to control the direction of current through each phase.

Tin can steppers don't cope well with half-stepping or full stepping. Wave stepping is about the most they can deal with. That simplifies matters in another way, though.

(Read the whole story)

Old news?
I've started building a RepRap. Is this old news? Yes, kind of. This is a summary of my first RepRap post on my own blog from a while ago. Since I have builders.reprap.org posting rights now (I hope you won't regret it ;) ), I will update you guys here. There will be more progress (I hope) blogged over the next days (with time-lapse series! Oh yeah), I'll post this summary first, so you won't miss a day of this fascinating journey :)

By the way: I chose the PIC based electronics because I'm more familiar with those and Arduino was not (yet) announced 'winner' of the 'What will be the main electronics development platform?'-contest. I'll stick with PIC based until I need to change.

What is the RepRap

[... skipped this part ...]

What is cool about the RepRap

Because of Self Replication, this project has to potential for exponential growth. If people print parts for a couple of their friends (a-la Pay it forward) they can use and improve RepRaps. Keep in mind that this is a first generation designed and built by a few people. RepRaps become better at a increasing speed as more people improve it, and there are more people that improve it because it's user base extends at a growing rate. All users will be able to keep their system updated by downloading latest designs and printing those upgrades. If you were to model this as a System Dynamics model, there are several reinforcing loops.

Growth for the sake of growth?

Growth for the sake of growth is all too common. There have to be other benefits. There are 3 aspects that make this project worthwile:

1. Learning. I think that everyone that has built a RepRap will have a more profound knowledge of mechanical construction, electronics and CAD /CAM (Computer Aided Design/Manufacturing). It is something that anyone can build, with a healthy dosis of persistence and experimentation.
2. Creative products. The produce of the RepRap is only limited by your imagination... and that of others. If you find 3D models made by someone else you can (customize and) print those.
3. Future RepRap generations. The power of the RepRap lies in its development. Over time, there will be modules such as recyclers (to reuse the plastic that you now throw away), 3D milling units, 3D scanners (so you can FAX, copy or digitize a physical object) and many other things. The fun part is that you can take part in this development. It could also make Open Source machines other than a RepRap.

I have no doubts that in the future, EVERYONE, will have a 3D printer in his home. One day, computers were for a few big corporations and institutes only. Right now most people cary more powerful devices around in the form of a phone, laptop, iPods, etc. The have become:

1. Much cheaper
2. Much more powerful
3. Much smaller
4. Less maintenance
5. Widely adopted

The same, I believe will happen with the RepRap. It will handle more materials, be more robust, etc. And if this takes of, all this valuable intellectual property will be public domain. No one can own and restrict ownership or distribution it for personal profit. Many printable objects will become freely available just like free MP3s are more and more common.

Getting the goods

Electronic supplies took 2 days from the US based Mouser to the Netherlands. Also, Zach Smith was pretty fast with his shipment of PCBs (also 2 days after I knew he shipped it). Doing all the ordering via the BoM (Bill of Materials) tool reduced the ordering process to a few clicks. The shipment of plastic Moulded parts for the RepRap takes a little longer because Ian Adkin unfortunately (mostly for them of course) had an ill co-worker.

All together I expect to spend about 500 euros. This is higher than the 400 euro mark, because I also ordered all the parts that someone with a RepRap could make more cheaply.

Soldering it

Soldering the first board was very easy. All instructions are very detailed. If you have experience with electronics you have to be careful not to skip the explanations and just build it the way you think you're supposed to. My advice (if you're somewhat chaotic like me) is to follow the instructions step by step: they are well thought out and you're almost certainly going to forget something somewhere...

Subsequent boards are simple as well. You need 5 controller boards (in the PIC based electronics variant, there is an alternative). 3 boards are for the X,Y and Z axis, 2 for the two extruders/toolheads that you can use. All boards are the same up to a certain point. At this point the X, Y and Z boards are finished as stepper controllers with a synchronization link between them, and the extruder controllers to make temperature measurements, regulate a heating coil, control a fan and to regulate extrusion speeds (a DC motor). Stepper motors are different from the well known DC motors. Look at wikipedia if they are unfamiliar.

Important tip: If you've clipped off the end of wires, make sure they don't fall on anything that is connected to the power. Clean up before you put powered PCBs on top of those things! If you don't you might end up damaging things. Most ATX power supplies have a short circuit protection and turn themselves off. You need to unplug the motherboard short (ground/black to green wire in my case, but check to be sure). On the right is my High-Tech ATX PSU conversion device^TM (also called a wire) that enables the PSU without a montherboard attached.

All added up: It took me about a full day to finish the electronics and test it. Depending on experience this may vary from 0.75 to 3 days.

The entire RepRap is using a maximum of about 50 Watts:

• including my ineffcient ATX PSU (Power supply unit) and
• with all stepper motors on (with hold 'torque' currents on them).
• With a case fan on
• With the nichrome coil heating a little bit (but it will not be fully on during normal operation).

This kind of power usage is not optimal, but certainly acceptable for now. Better tracking of the X, Y or Z axis positions would allow the hold torque to be minimized. If the steppers would skip this is not a big problem since the software would know about this.

Programming the PICs

Comms are working bidirectionally!

I created a programming cable to do ICSP (In Circuit Serial Programming). Now I can use the PicKit 2 from Microchip.com to program everything. I use "pk2cmd" to use this programmer on linux.

Software

The software was a bit of a hassle. As always, this was not the software's fault. I am especially good at breaking and then fixing things again. It turns out I had no working GLX support, my nvidia driver didn't work well under linux. I've had it working and fail again over the last few ubuntu distro's and kernel updates. I think my own fixes or workarounds to problems interfere with things that the Ubuntu team improves. A nice extra is that I've got my fancy (and arguably useful) compiz features back (see movie for an impression and to see why windows vista is "soooo 2007")!

Labels: begin start introduction, Darwin, PIC

In the long term the ideal way for reprap to produce circuity is by additive processes (laying conductive substances embedded in plastic objects). These processes are still being developed so in the meantime it would be pretty handy for RepRaps to be able to plot PCB masks onto blank boards with a pen so they can be etched.

I've been working on a Gerber (universal PCB format) plotting script for the last couple of days and finally have the basics working. The picture below was printed by my repstrap today and measures 32mm by 26mm. The thin lines are there because I still have no Z axis to lift the pen.


(I know it's mirrored, my X axis is inverted, needs re-wiring)



The original file in KiCad:



The script creates a photo plotter trace by plotting the straight line between start and end points, then by drawing the loci of the aperture movement from the start point to the end point at several intervals from the central line (diameter 0), up until the diameter of the aperture.
Good summary on gerber format

At present the script controls the RepRap directly over the serial port though I plan to add g-code export functionality.

svn

I had a few hours to continue building up the Big Board for Tommelise 2.0, today.

Since I am by no means settled on a final design to the point where I am brave enough to undertake printed circuit board designs, use a Eurboard stripboard for building my boards. You can get them on the web for $5-10.


My concept with the Big Board is to make it with as few parts and as simply as possible


The PIC 18F4550 does that quite nicely. That chip lets you go directly to USB connectivity with your PC without the drama of MAX232 chips to convert what PIC understands to be serial comms to standard RS232 protocols. Similarly, the PIC 18F4550 doesn't require a conversion chipset to get USB signals like the Arduino boards do either.

(Read the whole story)

Finally, I managed to do some prints about which I was very happy. YAY!

Following Demented's success I got excited and went down to my garage to do some prints. I was not brave enough to try the minimug and wine glass so far, was only testing widebar, circles and 10mm cube. I am using CAPA with a 0.6mm diameter nozzle. This works fine, my only problem is dumping in the corners where I would need some more nozzle speed up... I used a few trays to print on like HDPE, MDF, foam and mousepad. Foam and mousepad would be fine as the deposited filament sticks to them easily but they were not rigid enough to resist the warpage of the material. IKEA HDPE board did not work in my case at all. Capa just would not stick and it made a big ball around the nozzle... The original MDF board worked for me the best, the only thing is that it will wear away soon... I will give the balsa a try as Adrian did, see how it goes...

I finally finished my machine and as proof, I have a minimug!

This picture (above) shows the minimug printing the last few layers.

The print is finished here and the machine is going back to it's start position.

As you can see I have something taped to my MDF stage. It's actually a grocery bag (HDPE) and I'm printing in ABS. The black bits are electrical tape. I've since switched to an all electrical tape stage which works a whole lot better. I get really good adhesion from this setup. No problems with objects detaching themselves from the stage mid-build.

There were some alignment issues to deal with. Mainly this was a matter of getting my prefs set right in the host and aligning my head more fully with the stage. Can't have one side higher than the other. I fiddled some more with that and called it good.

Upon successful first printing--and after the minor celebration that ensued which only involved me and my friend Noah--we immediately started looking for other things to print. Well, Vik's wine glass looked nice and friendly so we gave that a go.


Minimug for size comparison for those of you who have printed a minimug.










Things were going well up until this last photo at which point we started noticing that the extruder wasn't extruding. We had lost pump preasure due to the bolts loosening--I never installed lock nuts on my pump body, stupid. We frantically fixed this while the machine continued to print and got it back on track...sort of. We were left with a ring about a half cm high of less than bonded material. No problem. The shape is still there and we continued to the end.


Sorry that's sideways. Couldn't figure out how to rotate it in Blogger. Feh... Anyways, the form came out great and I'm really happy with how the machine performed. I used a 1 mm extrusion tip on all of these prints. My longest build time was 1 hour, 20 minutes for the wine glass and 30 minutes for the minimug. I've subsequently printed several failed minimugs--one successful one too!--and another failed wine glass.

My next task is to install springs and lock nuts to the pump so that I can print without loosing pressure. After that, I'll get an aluminum body for my pump to replace the cast one and then I'll begin printing a set of parts. I'm really stoked that Vik replicated finally and I really would like to follow close behind him. I've got about 5 people who want a set and I would like nothing better than to give them one for a birthday or for Christmas.

Regards to all
Demented

Labels: finished, minimug, wine glass

While I was digging around yesterday for the parts I needed to build up my full-sized controller board for Tommelise 2.0 I unearthed the extruder polymer pump that I printed with Tommelise 1.0 last year.

From Chris' (nophead) work subsequently, I know now that my delamination problem came from not extruding the HDPE hot enough. Even so, looking at it I decided that it would be interesting to see how the warping that I encountered compared with what Chris experienced with similar (50%) fills.

(Read the whole article)

I had my youngest daughter and son at my place for Easter a few weeks ago. As usual I cleaned up rather quickly the day that day before my daughter was to arrive.

"Cleaning up" for me means filling whatever box comes to hand with all the loose bits that lie around my place and setting them aside. That is quick and quite efficient. Unfortunately, it can make finding what you "cleaned up" quite difficult afterwards.

In spite of that, I managed to find all of the parts that I need to build the full controller board for Tommelise 2.0.



You can see the prototype board above the full controller board. I haven't exactly decided how I want to lay out the hex inverter and stepper driver chips yet. It looks like there is going to be plenty of room, though.

While I've got all of the connectors to drive the three steppers, the extruder motor and the extruder heater in this photo, I haven't put in the connectors for the limits detectors in this picture. They are going to be in the upper, left-hand side of the board.

(No additional content on 3DReplicators.com)

I finally got everything to do with Tommelise moved over to and running onmy new XPS 720 H2O so I can start doing some useful work again.

Some weeks ago I hit a little bit of a wall trying to buffer USB data coming into my 18F4550 microcontroller in RAM. While the memory banks were arranged in 256 byte banks, dimensioned variables couldn't be any bigger than 128.

(Read the whole story)

I'm an engineering student, and one of the prospects of this idea that excites me most is the potential released by exploiting mechanical branches or "mutations" that individual end-users will create. Many end-users will want to tweak their machines and add cool features, and seeing as this is open source, I think that incorporating the most successful of these tweaks into the communal "gene pool" is not only legal but laudable.

In the interest of facilitating these mutations, I think one thing that would help immensely for some of us is to have a copy of the CAD used to design Darwin available. This allows more involved modifications to be digitally prototyped, directly printed, and then tested (I love how much easier this whole process becomes when we have access to 3D printers!) Currently there resides in the Subversion directory a copy (I'm not sure how up-to-date) of the CAD files in the Solid Edge format that Darwin was originally designed in, as well as copies of all of the RP parts in .STL and some of them (I believe) in .AoI.

I suggest that we augment these files with a complete copy of the machine plans in either the IGES or STEP file formats (both are open and ISO approved). This will allow the more CAD inclined of us to do our thing. :)

Embedded below is a copy of the Solid Edge files that someone who owns the program exported for me a few months ago as an IGES. I haven't worked with the files (much) since then, so this is pretty much the state of the design. There are some pretty obvious omissions. Unless someone on the dev team has more up-to-date files and can provide me with them, I'm trying take this model and by digitally following the assembly instructions, turn it into a fully modeled copy of Darwin. At that point, I will export a new IGES file and either find somewhere to host it online or place it in Subversion. My plan is to continue updating and re-releasing the model as the project goes through revisions.

For everyone's information, the software I'm using for this is Autodesk Inventor. I imported the IGES file, did a tiny bit of cleanup, then re-exported the files as a .dwf. I then uploaded that file to Autodesk's Freewheel service, which allowed me to embed the model here. It's quite a nifty feature, actually.




(direct link)

I'd like to put out one final call for more up-to-date files. These will work, but will take significantly more time to collate and rebuild than a more recent model would.

In addition, if anyone has Inventor and is interested in helping with this endeavor, let me know. I could always use some fellow monkeys. ;)