Tuesday, March 1, 2016

Building a RepRap Wilson

I really love 3D printing. I enjoy the end results as much as I am interested in the technology behind the machines themselves. Having a background in 3D design, it was obvious to me that 3D printing was something I was going to get interested and involved in. But had you told me 5 years ago that I would be building them from scratch, I'd have a hard time believing you. Unfortunately for past me, I've done just that - since November, I have built three printers from nothing.

My journey of self built printers began last year when I started researching various Prusa i3 designs. I had a few things I needed from a printer - it had to be reliable, it had to be sturdy, and it had to have high print fidelity. For these reasons, I wanted a printer frame built from aluminum extrusion (and it didn't hurt that I had a bunch on hand already). I also wanted it to be as close to a full RepRap as possible - no acrylic or machined parts. Some people might consider this a negative, but don't forget I am an enthusiast in the tech as much as I am the finished results.

After doing some research, I settled on an Adapto printer - it had a large build volume, it was built entirely from 2020 aluminum extrusion, and all of the parts were printed. I put together the aluminum bits and started printing things, and I ran into a few snags. The first of which is that the threaded rod I had on hand was larger than what the design called for. While that isn't a big problem - it's simple enough to modify the designs - the new printed parts began having problems where the captured nut would physically intersect where the linear bearings would be. Or in simpler terms - I would have to start majorly redesigning the printer, or buy all new parts, if I wanted to use it.

I was a little frustrated with the process, and around this time, the Colonial Marines group build at Freeside began to consume all of my free time. I had set the Adapto on the shelf to pick back up, where it was untouched for most of last year. Once the group project and the various events for the year was over, I started researching some other RepRap designs. I was planning on rebuilding one of Freeside Atlanta's printers into a new usable design, and came across the Wilson TS. I did an inventory of all of the parts I had on hand, and it turns out I had everything I needed to convert my Adapto to a Wilson.

I was building The Kraken and my Wilson in tandem and the build was a breeze. The instructions online from mjrice are pretty good - his videos are for the 10mm smooth rod version instead of the 8mm ones I was building, so parts differ a bit - but having put together both my Folger Tech 2020 i3 and doing a lot of work on my LulzBot AO-101 and Freeside's AO-100, I knew my way around how to put a printer together.

I did decide to make a few modifications right away. The first of which is to help the structural support of the machine by adding in some hidden corner braces in all of the 2020 corners, and L brackets where the X-axis and Y-axis frames meet. This helps make the printer be extremely rigid and the prints that come off of it look amazing. I also added in an inductive sensor for the Z probe to enable auto bed tramming, a feature on the LulzBot Mini that I am totally in love with. Having your first layer always be perfect without much interaction is a real dream come true, even if the initial setup is a bit of a hassle.

I purchased an LCD screen with SD card support so that I can run the printer without having my laptop connected. This is an important feature for me since I am often running all of my machines simultaneously, and I don't want to have my laptop permanently attached to them. On my Wilson I am using a RAMBo board, and setting up the LCD had some pitfalls. The Arduino board library has to be modified in order for the RAMBo to properly output to the LCD. Here's the article I found on doing just that.

I added a top-mount for my filament spool, but I may actually take this off as it seems to have created a really unbalanced load. Running the printer at the speeds I had while the spool was on the table causes it to vibrate and move around a lot. I'm going to slow down the printing a bit and try to tune the feedrate, jerk, and acceleration settings. But having a really slow printing machine isn't a lot of benefit to me so hopefully I don't have to slow it down too much.

I've only been printing on the Wilson for a couple of weeks and I'm really happy with both my Wilson and Freeside's "The Kraken". For my next printer I am going to look into the Wilson 2 which improves on the current design, and given how great of a machine this already is, I think the Wilson 2 could be an outstanding machine.

Here are some links to the parts and designs used for my Wilson builds:
Read More

Friday, February 19, 2016

February Project Updates

I'm at an awkward stage for all of my projects, where I'm not quite done with some of them, and not quite started enough on others to start doing write-ups. It's weird to spend an average day working on several things, and not really feel like you made much progress or have anything to show for the work, yet actually have things accomplished.

My ongoing work on various 3D printers continues. I finished building my Wilson TS RepRap printer and brought it home to my printing fleet. I also have more or less finished Freeside Atlanta's Wilson known as "The Kraken" and it should be printing soon as well. I'll be writing up a build log and review of the Wilson once I have both of them in fully working order.

I've also been tinkering with Freeside's LulzBot AO-100 (and by extension, my personal LulzBot AO-101) to work out some problems it's having. That machine is really old and has had a lot of wear on it over the years, so I've had to really show it some love.

In actual prop news, I have a handful of space guns in the works.

A few months ago I decided to start making the Quake 2 Blaster that I've had an idea to make in the back of my mind forever. I have it cast up and ready to start painting, so this should be done very soon and I'll be posting a build log.

Speaking of side projects, I've been teaching Elliott about prop and costume building, and we picked a project to learn things on. We're building a set of Marine armor from the game Natural Selection 2, and we've been 3D printing the rifle and the pistol.

As if I didn't have enough space gun projects, I printed out the Federation laser rifle from Rick & Morty. Right now it's had a first pass of primer, so there's a lot of work left on this one.

The main event this year though is the T-60 Power Armor from Fallout 4. A couple weeks ago, Mike Vrogy came up to the space and helped Elliott and I build the stilts. We're starting 3D printing and CNC production on the rest of the armor, but that's going to be a much bigger write up!

That's all for now, keep an eye for more write ups soon.
Read More

Friday, January 22, 2016

Review: Folger Tech 2020 i3 kit

I had been in the market for another 3D printer for my fabrication fleet, and I had my eye on a few machines during Black Friday. I narrowly missed a great sale on a Wanhao Duplicator i3 for a cool $299, and instead I settled on a Folger Tech 2020 i3 kit on sale with an LCD panel.

I had done some reading on this particular kit so I knew to expect some hurdles during it's construction. The biggest complaint that the community has - and indeed I have too - is that the build manual has several mistakes and blatant inaccuracies that Folger Tech has yet to fix. There's some simple stuff like typos of bolt dimensions - using one bolt length in one sentence and another length in the next sentence, leaving you to figure out which one they really mean. These are easy to figure out. But then there's the problem where it tells you to mount the X-axis end stop on the wrong side, and if you don't understand why 3D printers are put together the way they are, you'll have a difficult time understanding why it's moving in the "wrong" direction and why it won't home properly. I highly recommend reading the manual fully before starting to make sure you know what to expect.

There is an absolutely massive thread on the RepRap forums which contain a huge amount of information and fixes. As of this writing, the thread is at 88 pages long and I've only managed to work backwards through about half of it. If you're considering one of these kits I recommend at least skimming through the forum thread on your own, but I've tried to compile the biggest issues and fixes from my experience here.

One minor annoyance I had was with the T-nuts that were included. They are very tiny and difficult to work with and I kept struggling with getting them into place whenever I had to tighten down or move an element. I think they are the ones that are meant to be used after you've put a 2020 frame together, so that you don't have to disassemble everything. But that they are so small it makes them a pain to slide into place along a stretch of aluminum. Your kit may come with regular sized T-nuts, but if I were to buy another one of these kits, I would go ahead and get a bag of regular sized T-nuts just to save myself some frustration.

The electronics setup was straight forward if you've used a RAMPS board before. The build manual tells you where to connect things, but I do recommend pulling up the RAMPS schematic so you know what you're connecting to and why. This is especially the case with the X-axis end stop, since as mentioned before the build manual has some incorrect info.

The RAMPS board I received appears to have had some damage in shipping. I did testing of it and it doesn't seem to be more than cosmetic, but I did read that others had issues with poor quality soldering as well as similar damage. The good news is that Folger Tech is good about sending replacements for parts, but I'm holding off on seeing if that's necessary since I've put about 50 hours on the machine without any hiccups so far.

One problem I do have in the electronics department is that there is not a power switch for the machine. This is a pretty minor complaint, but I do like being able to power off the power supply without having to pull the plug or turn off a power strip. I found a mod on Thingiverse to both add a switch and a removeable power cable which is on my list of to-do upgrades.

Folger Tech has a version of the Marlin firmware on their site to download, but it's quite a few versions behind the current release. The community again comes to the rescue, and therippa has a fork of the latest release of Marlin with the Folger Tech settings already imported. There's also a version that enables auto bed tramming if you decide to upgrade to an inductive Z sensor as well - which I went ahead and added to my build, and in general I can't recommend enough.

Speaking of the Z-axis, the stock Z end stop leaves a lot to be desired. It's difficult to get it really zeroed in since it's attached to the smooth rod. There's a lot of options on how to fix this, and I went with a solution similar to what I've seen on other machines. I printed a part that attaches to the X motor mount points and puts an M3 bolt in the path of the Z end stop. This way you can easily tighten or loosen the bolt to raise or lower the contact point by small increments. I also printed a thumbscrew head for the bolt to make it even easier. Of course, if you are using auto bed tramming, you won't need to worry about this.

The Z-axis is driven by a M5 threaded rod, which seems pretty rickety compared to other machines I use that have M8 or larger lead screws. There is a popular mod for this machine to use M8 lead screws and there are several parts on Thingiverse ready to convert it over if you're so inclined. It seems like a great modification and I may switch to that in the future, but I'm not excited about having to basically completely disassemble the machine just to install it. If you decide to stick with the stock M5 threaded rod, there's also an anti-backlash hack if you want to get the most out of your stock parts. Otherwise, if you want to upgrade to a bigger threaded rod, plan on doing that during your initial build and save yourself some time.

The last problem I had with the Z-axis is that it is simply shorter than is advertised. Folger Tech's site says it has a 7" build height, but the machine really is only able to get about 5" with the Z-axis printed parts. I did find a mod on Thingiverse for Z motor stand-offs that regain the extra 2", but I'm really just confused about why it's advertised with that size. My printing volume with this hack is 200mm x 200mm x 160mm.

I've seen others who have moved the Z-axis steppers to the bottom rather than the top, and again that's something I may do in the future since the top of my machine is already pretty cramped. Between the spool holder and the LCD mount, the steppers would be nice and out of the way on the bottom.

Aside from all of that, I really am quite happy with the printer. The prints coming off of it are an excellent quality and I really have no complaints in that department. I've printed all of the modifications to the machine on itself, so it's a "true" RepRap printer in that regard. One of the first mods I made was a different filament mount, a top mounted bearing spool holder to replace the side mounted one. If you don't use a bearing driven spool holder, I can't recommend them highly enough, it reduces any filament drag artifacts from your prints.

The overall construction is great, the all aluminum frame and the corner brackets make the whole chassis incredibly rigid. This really shows through in the print quality. One thing I especially like is that it uses almost zero laser cut parts, the only ones included are to hold the Z-axis smooth rods in place. Everything else is metal-on-metal contact, or standard 3D printed RepRap parts.

I did go ahead and buy some cable management to tidy up the mess of wires. Back in my PC case modding days I used the nylon flex tubing "Techflex" religiously, so I used that for as many of the runs as I could. I also bought some drag chain for the X and Y axes to minimize issues with cables tugging or crashing into the print. Plus it makes the printer look a lot more clean and professional, which is nice.

The deal I got included an LCD screen, which I primarily use for the SD card reader since I don't like having to keep a computer attached to my printers at all times. It doesn't come with a mount or a bezel, but thankfully you already have a 3D printer to make one. There's a lot of options available depending on where you want to place yours. I may find a different location for mine since it's a tight fit between the spool holder and the LCD panel on the top. As always, be sure you check your firmware to make the necessary changes to activate the LCD panel and SD card reader!

For the price, this kit really can't be beat and I'm already very happy with it. I would hesitate to say it's great for beginners, but if you're someone who has some hands-on experience with a 3D printer - using one at school or work - and you want to make the jump to buying your own, this would be a good machine to pick up.

Here's a few collections of parts that I recommend looking in to if you're considering buying this machine, or if you already have one and want to fix some of the common bugs with it.

Collection of upgrades on Thingiverse

Collection of upgrades on the RepRap.org forums
Read More

Tuesday, November 10, 2015

Review: LulzBot Mini 3D Printer

Back in March, Freeside Atlanta won a LulzBot Mini 3D printer during a hackerspace giveaway they were running. We already have one of their older machines, an AO-100, so we were very familiar with their printers and how easy they were to use. I've used several of LulzBot's printers before - I own an AO-101 myself - and I was really interested to see what the Mini brought to the table.

As I said in my Cube 3D review, I really dislike the idea of "just press go" type of machines. 3D printing is still too young of a technology for mass adoption, and pushing fickle equipment on to the unsuspecting masses will put 3D printing in a negative light.

Having said that: the Mini is probably the best printer I've ever used.

The Mini's name comes from it's generally small build platform of roughly 6" cubed. Normally this would really deter me from using it as I am generally printing large costume pieces, but the small printing volume is the only negative I can possibly say about the machine. The machine comes fully built and ready to use, the frame is attractive and everything is very well constructed. It took us about 20 minutes between unboxing and pulling our first print off of the bed.

Included is a LulzBot branded install of Cura which has all of the settings for the Mini included, so the time between unboxing and printing was incredibly fast. There are several preset quality options, and the highest detail option at 0.1mm produces amazing results. You can go under the hood and tweak all of the print options, but the default settings produce great objects on their own.

But really, the two best features are the PEI printing surface, and the self leveling bed.

PEI is a bit of a new development in the 3D printing world. It is an "aerospace grade glass fabric polyetherimide (PEI) composite" that requires virtually no prep to use in 3D printing. It replaces the usual glass print surface and is adhered directly to the silicone heated bed on the Mini. Unlike printing on borosilicate glass, which you need to apply either kapton tape or ABS juice or purple glue stick to really get large prints to adhere, PEI bonds to both ABS and PLA when heated with no additional work needed.

For example, we 3D printed the entire Xenomorph head - around 38 prints total at ~5 hours each - and didn't have anything come detached from the bed and only very minor lifting on 1 or 2 very long prints.

Once the surface cools, your prints will come loose with very little effort, and only large flat objects need to be pried off using a clam knife. Occasionally I will clean the surface with denatured alcohol to get rid of any dust or skin oils left over from the bed being touched while removing objects, but that's the most maintenance I've had to do.

However, after about 500 printing hours being put on it, the surface has started to bubble and should probably be replaced soon. I am chalking that up to having 20 or so people at the space using it, and not everyone treating it as delicately as they should. If you are using it yourself and you treat the machine with respect - and especially waiting for the bed to fully cool down so the PEI releases it's grip - I doubt you'd have the bubbling problem.

The self leveling bed is fantastic, and virtually negates the need for any maintenance and upkeep. Unlike printers like the PrintrBot Play which use a magnetic sensor, the Mini uses electrical signal sent through the print head. In the corners of the printer are 4 metal discs, and the machine uses the change in voltage to measure when the nozzle has made contact.

On my printers at home, keeping the bed in level is the biggest struggle and the cause for any real printing issues I have. When I am building a space gun or some sci-fi armor and I'm planning on putting 200+ hours on the machines for just a single project, reducing the down time between jobs is a life saver. Using the Mini, I know that the bed will always be perfectly level and my props will come out pixel accurate to the 3D models.

The Mini comes with a 0.5mm nozzle, which compared to the 0.35mm nozzle I use at home, I was worried that the print quality would suffer. But between the PEI bed and the self leveling feature, and the rock solid construction of the body and mechanical elements, the actual print quality print for print is some of the best I've seen off of an FDM printer.

I'm debating on whether or not I want to buy one for myself, but I am strongly leaning towards doing so. My only reservation would be to wait until the stand out features of the Mini make their way into the TAZ line of printers from LulzBot, which I'm sure is inevitable. Having a 300 mm x 300 mm printing surface on PEI with a self leveling bed would make them the only FDM printer I'd ever want to use again.

If you want a printer that is reliable and doesn't require frequent upkeep, then I can't recommend this printer enough.
Read More

Wednesday, October 21, 2015

Furiosa's Bionic Arm (Mad Max: Fury Road)

This post originally featured at Tested.com

Last year, I was talking with a friend of mine about some of our "holy grail" projects. I told him that it was one of my dream builds to make a T-800 Endo Arm, as an actual prosthetic for an amputee. You know the scene: in Terminator 2, Arnold cuts off the skin of his left arm to expose his robotic endoskeleton.

I told him how it would be a dream project to build an Endo Arm like in this scene, for someone who is missing a limb. I've met or been made aware of a few people over the years who used their unique body attributes in their costumes, but I never had the chance to connect with someone.
He said he wanted to introduce me to someone. This is Laura.

Laura is a left arm transradial amputee, meaning that she is missing her left arm below the elbow since birth. She's also really into cosplay, and living in Atlanta, she has been a "featured zombie" on The Walking Dead. You've probably seen her in the shambling hordes.

We met and I told her about my idea and what we could do, and she was enthusiastic. I felt really lucky because this really was one of my dream projects! She said she had done a few costumes in the past that incorporated her arm, but nothing really to the scale of what we planned. The idea was to 3D print a CAD design for the Endo Arm and possibly wire it up to an Arduino and some sensors and servos to make the fingers open and close. It was going to be a fun and really challenging build, and I was really looking forward to getting it started.

And then, Mad Max: Fury Road came out and changed everything.

You might have read Laura's blog post on her Tumblr that went viral a couple weeks after Fury Road hit theaters. To quote Laura's blog post, "If I don't cosplay this character immediately I'm pretty sure all my friends will riot."

We had a short conversation at MomoCon here in Atlanta and I asked her permission to build her the Furiosa arm as a real actual prosthetic, much like we were going to do for the T-800 Endo Arm and she couldn't have been more thrilled. Our plan was to finish it for Dragon*Con 2015, and we both couldn't wait to get started.

During our planning for the Terminator Endo Arm project, I took a 3D scan of her using an Xbox Kinnect and a software called Skanect. It allows you to easily get a rough 3D scan of a person or an object. It's not high enough detail to look photo realistic, but it's enough to get basic proportions. I use this myself to scale Pepakura files and do other digital sculpting where I need to have the proportions of a person or a thing. We tried getting a 3D scan of Laura's arm and the results were all right. It was just enough to use for scaling and "subtracting" her arm from the Endo Arm model.

When we shifted gears to Furiosa, I decided the first thing to do was to get a plaster cast of her arm, because the prosthetic would have to actually fit her, and there was no way for me to "try on" the prop myself while building it. After an afternoon at the shop, I had one of the weirdest casts I've ever made, but it was exactly what we needed!

At this point I got a lot of help from Adam Greene of Pixelbash Props, who took a higher detail 3D scan of the plaster cast, and assisted by creating the 3D model used for the build. Laura and I agreed that we should 3D print the arm to cut down on as much weight as possible. I was worried that if it was too heavy that she wouldn't be able to lift it, or she would become fatigued after wearing it for a short period.

The pieces were 3D printed on my home 3D printer, as well as the printers at Freeside Atlanta, the non-profit hackerspace that I work from. After a few long prints--a total of about 30 hours print time--everything was ready to be cleaned up and assembled.

The process of cleaning up a 3D printed prop is pretty simple: Rough sand the surface to get rid of some of the print lines, then (in the case of an ABS print) use "ABS sludge"--a thick mixture of acetone and ABS--to coat the surface. This acts like a body filler and will help fill in the remaining gaps, but as the acetone evaporates, the ABS bonds to itself, so you have a single rigid object. The part is then sanded again with a finer grit sandpaper, and coated in a thin layer of spot putty to fill in any remaining pits or print lines. After that dries, the excess is sanded off, and then primed for painting.

Once the 3D printed parts were cleaned and roughed together, I designed and laser cut the mesh screen for the fingers. It was cut out of 3mm acrylic and heated with a heat gun, and then bent to shape around the fingers. Since building this I've discovered the actual prop likely used a motorcycle exhaust baffle, but the acrylic worked out great for us as it was lightweight and readily accessible.

There are two wrenches in the arm, one attached to the "pinky" finger and one lashed to one of the forearm pistons, that needed to be fabricated. I rummaged through the autobay in the shop to find a couple of wrenches that were of suitable size, and then molded them in Mold Max 30, one of the molding silicones that Smooth-On produces. I wound up casting these out of Smooth Cast 320, but my original plan was to use a light and flexible foam. That turned out to be unnecessary since the resin ones were small and light enough to not cause issues.

The hand and the finger grilles were hit with a primer, and then a base coat of matte silver. I then did a light dusting with a darker metallic paint for the the lowlights and to bring down the "shiny and chrome" factor. After all, Furiosa's arm is a functional piece of equipment and has a lot of wear and tear from being out in the Wastelands!

Laura came in for a test fitting, and for us to size and finish the strapping system for the arm. Scrap leather and spare belts were cut down to size and riveted together for the harness. Laura sewed the shoulder pad which goes underneath the pauldron, and we attached those together.

For the support pistons, I used some 6mm fiberglass rod laying around the shop, and 3D printed connecting joints for them. Those were then bolted on to the 3D printed arm, giving the wrist a range of motion. In other words, Laura will be able to pose the wrist.

The two wrenches came out really well. The small one was attached to the pinky finger, and the other was wrapped tightly to one of the support rods with some leather cord. Fun fact: I believe the leather cord is there to cover up the manufacturer of whatever wrench the prop crew cast off of, because it's placement is exactly where you would expect to read "SNAP-ON" or something. So, I followed suit and covered up the name with the leather.

Then it was on to weathering, which is my favorite part of any project. I did a few light washes in black and various tones of yellow, orange, and brown, but I wiped most of it away to keep it looking metallic. The movie has an orange filter applied to most of it, so I relied on B-roll and behind the scenes photos for true to life colors. The arm isn't rusted as much as it is worn down and dirty, but I did apply some light rust around the bolts that connect the finger grilles and the other hardware attachments.

The shoulder was designed in CAD from looking at stills from the film, then laser cut out of EVA foam. There is what appears to be a model plane engine on the front, so I grabbed a random DC motor from the shelf and glued that in. The pull strap I quickly 3D printed based off of product photos for a weed wacker. Then the whole pauldron was weathered as well.

There are three cables connecting the shoulder to the arm: a braided metal hose, a clear/yellow tinted tube, and a brake cable. I got similar looking things of each and bolted them on to the arm, and attached them on to the shoulder end.

She also wanted me to make the belt buckle emblem, which I quickly 3D printed up, and cleaned and weathered. I grabbed this model off of Thingiverse, which you can download here.

I met up with Laura at Dragon*Con and delivered the prop to her in her hotel room. We did some final fitting of the prop on Saturday so she could wear it to the costume contest, and to make sure everything was 100% for the big Fury Road photo shoot on Sunday afternoon.

That about wraps it up! It was an incredibly fun and rewarding build and I'm happy I can scratch one of my dream projects off of my list.
Read More