Nate Duxbury's Blog

Make all the things!

1 Comment

MakerFaire Pittsburgh!

MakerFaire Pittsburgh and Robot

This past weekend I attended MakerFaire Pittsburgh along with the company I now work for, WARDJet. We brought along our brand new show truck with one of our waterjet CNC machines in it! It was awesome talking to a lot of excited makers, both about the waterjet we brought with us and their own projects. I tried to take as many pictures as possible of things that I thought were particularly neat.

The WARDKit Show Truck at MakerFaire Pittsburgh

The company I work for recently built a show truck to show off one of our new offerings. What offering is that you may ask? Well, its a water jet CNC machine in the form of a kit! We fabricate the parts and then you assemble it. This isn’t meant to be a sales pitch though, but if you want that, just check out the WARDKit website. Sometime in the near future, I’ll do a post about water jets: what they’re capable of doing and how they work. That post is will be linked HERE (once its actually done). What I’d rather talk about in this post is my experience at the MakerFaire.

MakerFaire Pittsburgh was a two day event going from 10a-5p both Saturday and Sunday and boy was it busy. We drove over to Pittsburgh Friday night after work and decided to head straight to our truck to do a little more preparation. While the rest of the makers were gathered having dinner and drinks we were outside drilling holes and hammering carriage bolts into place! We wrapped up Friday night around 11p and took a couple pictures of our display, including an aerial video from a go pro mounted on my co-workers drone.

Friday night preparations.

Friday night required a little bit of work to get everything ready.

Saturday morning required an early start to get everything else at our display in tip-top shape. There was an impressive amount of stuff to take care of. We started up our intensifier pump, charged our air tank, filled our water reservoir and ran the machine through its paces to get ready for the day. We set-up a table with all the information, business cards and sample cuts for people to look at. We finished setting up right around 9:30a, which provided me with a little time to walk around before it started. This is when I got to talk to Bob, who was printing and vacuum forming chocolate molds (see more on this later in this post).

Saturday was definitely the busier of the two days. We had the water jet under motion and running in low pressure, water only mode, at about 12,000 PSI. We were running cutting demo’s every 45 minutes or so, mostly cutting out makerfaire robots out of red craft foam. We were handing these out all day as key chains, and people seemed to love them. Grab the .DXF here and cut out your own! The event opened up at 10a and really kicked into high gear right after lunch time. We had a pretty steady surge of people coming by to watch the cutting demos and it was great to see our show truck full of observers.

Foam MakerFaire Robots sitting on top of the Water Jets crossbeam.

Foam MakerFaire robots sitting on top of the water jet’s crossbeam.

The people who were stopping by ranged from children to makers to engineers. Almost everyone was excited to watch a huge machine move around and cut stuff with WATER! I got to practice my sales pitch and generally just talk to excited makers of all types. Kids went wide eyed and some people said “I have to have one” as soon as they saw it. My favorite reaction was probably a guy who saw it and looked intensely interested in how he could get one into his garage. Even better, was his wife hearing the price, shrugging, and then saying that it really was quite reasonable!

Excited onlookers watching the water jet cutting.

Excited onlookers watching the water jet cutting.

We also cut out the MakerFaire Pittsburgh logo and our logo during one of our demo cuts. We cut it out of the same red craft foam as we did for the robots. We also cut an identical copy of this out of 1/8″ Stainless back at our facility on Friday afternoon. When we weren’t cutting foam during a demo, we were running the same logo program tracing over the stainless steel one.

Logo Demo cut out of foam and stainless

We also cut a pear that we just happened to have!

Water jet cut pear!

Water jet cut pear!

Sunday wasn’t as hectic, but we had a hiccup with our intensifier pump that rendered us unable to do any cutting demos. Thinking we were set for the second day, we had showed up a little closer to opening time. Fairly quickly, we discovered the unfortunate situation with our pump and tried our best to fix it. After about two hours on the phone with tech support we found out that it was out of our power to make it work that day. It knocked the wind out of us a little bit, but we still talked to a lot of people and everyone was still excited to see it move.

Water Jet Cut Samples

Some samples we had set out, including: craft foam, carbon fiber, aluminum, granite and a large chunk of steel.

However, it wasn’t the end of the world! I managed to decipher our G-code files and determine what the M-code for water jet on/off was. Knowing this, I opened up the g-code files in Notepad++ and did a find/replace all for the on/off commands with a Z up/down move. Doing this allowed us to zip tie a dry erase marker to the water jet nozzle and turn the whole thing into a draw bot! We couldn’t cut foam, but we could at least draw on a white board and show how our tool paths worked… mostly. Draw bots usually have a spring loaded marker end to account for variation in the drawing surface. Our marker was only zip tied though, and while our machine is rock solid repeatable and square, white boards are, unfortunately, not rated for flatness. This experience proved to be pretty awesome though. I knew a little about g-code and such before this, but I had never done any post-processing manually like this. It was very rewarding figuring it out and seeing it work while working under the gun WHILE the MakerFaire was going on!

Sunday wrapped up and and we had to pack up the truck. The event ended at 5p and it took us until about 7p to strap everything down and button up the truck.

All packed up and ready to return to Ohio.

All packed up and ready to return to Ohio.

Other Highlights Around the MakerFaire

While both days were quite busy, I did get a chance to wander around the rest of the MakerFaire a little bit. Some of the highlights I saw included: a rubber band Gatling gun, a cardboard velociraptor suit and a guy printing then vacuum forming chocolate molds!

The booth next to us was RealBotics, a group of people who made a software platform for controlling robots over the web or via direct connection to a computer and an arduino. They had a spotter scope that was attached to a series of 3D printed racks to control its motion and a webcam where your eye would normally look into it. Even better, they had an air cannon that you could aim and fire! At it’s core, it looked like a simple pneumatic potato gun, but it was much more than that. You could control the pitch angle and rotation of the whole thing with their software, and it was also breach loaded (actuated by a pneumatic cylinder)! They were firing tennis balls all day and at the end of the day on Sunday, they were firing cupcakes!

Cupcake Shooting Cannon!

Cupcake Shooting Cannon!


Catty-corner across from us was the HackPGH booth and one of the makers there had a beautiful electro-mechanical clock along right beside a large rubber band Gatling gun and a wooden music box. The clock used a solenoid and an arduino to keep the pendulum moving indefinitely. Really makes me want to get to work on my own mechanical clock (which will hopefully be water jet cut out of granite!).

These projects were top notch.

These projects were top notch. The size of the rubber band gatling gun alone was impressive!

There were hoverboards everywhere. Here seen in an interesting application of moving around this makers music box.

There were those hoverboards everywhere. Here seen in an interesting application of moving around this maker’s music box.

Pittsburgh TechShop was directly across from us doing some forging! It was quite chilly on Saturday morning, I bet that little furnace came in handy!

TechShop Pittsburgh Forging in Action

The aforementioned cardboard veliciraptor: KitRex. What’s pictured below is actually a suit! Someone got in it and walked around the faire at regular intervals. Also of note, were the enormous googly eyes. Definitely the largest I’ve ever seen!KitRex Velociraptor Suit

A bad picture, but this stringed instrument robot was really cool to see in action. The four strings were selectively tensioned at different points by computer controlled carriages to create different notes and a guitar pick was turned to make music by strumming the string. Eric Singer of singer robots was the creator of this awesome, robotic musical instrument.Singer Bots Musical Robot

I forgot to take any pictures, but I did spend a bit of time talking to Bob of who was printing  objects on his 3D printer, vacuum forming them into molds (using an awesome & simple laser cut vacuum former) and then casting chocolate bars out of them! It was great talking with him and it gave me some ideas of my own. I will definitely be looking over his blog and hopefully making some chocolate molds of my own in the near future! At the very least I will be making a vacuum former like he was using. Too cool of an idea to pass up.

A sampling of Bob’s work (borrowed from MakerFaire Pittsburgh’s website).

The whole MakerFaire Pittsburgh Experience was awesome. In 48 hours I got to talk to so many excited makers and see so many awesome projects. I really enjoyed the experience but man was it exhausting. Once I finally got home on Sunday it sure did feel great to sit down and relax.

This event has got me back into the maker mood. It’s been a little while since I’ve done a project and going to this MakerFaire snapped me out of my slump. Keep an eye out for some projects in the coming weeks!


Space Time Coffee Table

Finished Product First!

Finished Product First!


Recently, I graduated from Ohio State, started a new job and moved into my own apartment. I had almost no furniture for the new place but this afforded me a great opportunity: I could make some of my own!

The first thing I decided to make was a coffee table. Months ago, someone on Reddit posted a conceptual design for a “space time table.” The idea behind the concept is to imitate the curvature of space time around a massive object (such as a black hole). I thought it looked awesome. At the time though, I was living in a shared college house and did not have the funds, reason or access to machinery to make such a thing, so I filed it away in my head as something cool I would look into in the future. Fast forward to May of this year, and I am now employed at a Water Jet manufacturer, where I am allowed to use the sales model for personal projects after hours! (possibly the best work benefit ever?!) With the tools now available and the reason to very clear, I began modifying the design to be manufactured.

If you’re more interested in just pictures, check out this Imgur Album. It’s not as thorough as the this write up, but it is a quick summary.


It was super convenient that the reddit user who posted the concept images also posted the source files AND they were in SolidWorks format! Perfect for me to load up and modify to fit my needs.

Step one was downloading the source files (Source). Step two was deciding what size I wanted to make it, such that it would fit nicely in the space I had available for it. So, like any normal person, I laid it out in CAD!

Rough Floor Layout of my apartment to get sizing.

Rough Floor Layout of my apartment to get sizing.


In the center right of the layout is the couch I have and then the square is the table size I felt was comfortable. The size I drew was about 24″ x 36″, but as I was going to be scaling the dimensions off of the concept, I could only really choose one of the dimensions and the other would be based on the aspect ratio of the existing design. In the end I was able to get 34.125″ x 23.5″ as the top dimensions.

I used envelopes to mark the bounding box of where my coffee table would fit. This gave me a great sense of scale for the project.

I used envelopes to mark the bounding box of where my coffee table would fit. This gave me a great sense of scale for the project.

Knowing the final size I wanted, I could now start modifying the design. There were four main points to this part of the process:

  1. The size of the table. The original design was 45.5″ L x 31″ W x 17.4″ H and I had to shrink the width dimension down to at least 24″. Applying the same scale down to all dimensions yielded 34.125″ L x 23.5″ W x 13.125″ H.
  2. The height of the table after scaling. A 13″ Tall coffee table would be no good. The bottom of the table was also very fragile in the original design. It terminated in sharp ends, which i did not like from a durability standpoint. The solution here was easy, I simply added material from the sharps directly down 3″. This increased the height to a nice 16.125″ and made the base less fragile. This feature was a departure from the concept, but it was a practical change.
  3. The concept model was all 3D features. The concept had sloping contours on every member and would have been extremely hard to manufacture. Again I decided to make a change for the practical, by removing all 3D features and making each part only a 2D contour. Since I had the 3D contours already from the concept model, all I had to do was pick the side with the larger features and extrude it to become the thickness of the whole part. I did this process for each part, but since I was going to scale everything anyways, all I ended up taking from each part was it’s 2D contour. Using only 2D contours made manufacturing immensely easier and still made for a very good approximation of the effect I was after.
  4. Making it possible to assemble. The concept model was a multi-body part file. This meant that the model’s bodies intersected each other, and that they could not actually be assembled as is. The fix here was to make the parts have matching slots and tabs. The parts going in the long direction got slots cut in the top half of them and parts going in the short direction got slots in the bottom half. With these interlocking slots, the assembly became assemble-able and even very easy to assemble!
The original concept on the right and my model on the left after modifications.

The original concept on the right and my model on the left after modifications.

SolidWorks render of the coffee table after modifcations.

SolidWorks render of the coffee table after modifcations.

I also did something really clever when I re-modeled the whole thing. I linked all the dimensions together with dimensions. This way, I could specify any thickness of material, any slot tolerance and any table height that I wanted by conveniently changing the value in a text file! I was unable to link a scale feature to the text files (not sure why…) but the rest of the features were now all linked and very easily changed. This allowed me to compensate for material thickness (since wood is rarely true dimension) and adjust the tightness of fit in the slots. It’s the first time I’ve done equations in SolidWorks, but I can see it being very helpful for projects like this. I could have used 1″ thick wood if I wanted and it would have been only three keystrokes to make the change.

I put together a quick explosion animation to check out how it would all assemble.


Here is the link to my  Design Files. The files are available in STEP, IGES, DXF and SolidWorks formats. Even if you don’t have a water jet, you could still use a large format CNC router or even print them out 1:1 and cut them with a jig saw.




After completing the design, I wanted to make sure everything would fit together correctly. When I had the time one weekend, I fired up my Shapeoko 2 and got to cutting! I planned to make the final table out of a 4’x8′ of .5″ plywood, so going to quarter scale translated perfectly. I had a 1’x2′ of .125″ HPDE, so I could even check my nesting. Not much to talk about here, the shapeoko cuts through HDPE like butter, and I was able to get all the parts out of just the one sheet!

1/4 scale model of the space time coffee table made from HDPE on a shapeoko 2 CNC router.

1/4 scale model of the space time coffee table made from HDPE on a shapeoko 2 CNC router.

Doing the scale model proved to me that the interlocking slots worked just fine and that I could fit all the parts onto one sheet of plywood. Plus I could bring it with me to work and explain to my coworkers what I was planning to make.


This part was a lot of fun. It was not the first time I got to use a water jet for something, but it was the first time that I ran the machine (mostly) my self. I leaned heavily on one of my coworkers for setup help, speeds/feeds, and a little bit guidance (shout-out to Ben Adams, thank you for all the help Ben!).

This section is best explained with pictures, so I will do my best to annotate each of the steps.

The first step was nesting the parts to make the part program. I used a 5'x5'x0.5" (actually real dimensions!) sheet of baltic birch plywood

The first step was nesting the parts to make the part program. I used a 5’x5’x0.5″ (actually real dimensions!) sheet of baltic birch plywood, and laid it all out with SolidWorks. Everything fit within a roughly 58″x 58″ bounding box, leaving me a 1″ border to clamp on. I exported this assembly first as a part file, then as a .dxf to feed into the water jet CAM.

Drilled Pierce Points

Kind of hard to see, but the little piles of wood are points where I had the water jet do a drill operation. The second head on our water jet has a pneumatic drill mounted on it that can be set up to drill holes in parts, or in this case, drill the pierce points for the water jet. A pierce is where the water jet initially punches through the material, and can result in delamination in woods, or just sloppy starts to cuts. Using a 1/8″ drill bit was an easy way to get clean pierce points, by both being big enough for the water jet to hit the target, and small enough to not ruin any of the areas of the parts that I wanted to keep.

Water jet cutting in progess

The water jet cut was going very nicely. The water jet uses super high pressure water (~50-60ksi) and abrasive (basically a sand made from garnet) to cut. As the company’s website says, its not actually cutting, but accelerated erosion. Basically, it is sanding it like crazy until it is through. Side effect of this is a super nice edge finish (well, this depends on feed rates, but I went slow enough to get a very smooth edge finish).

More water jet cutting

You might be thinking, “Isn’t water bad for wood? Won’t that make it warp and be generally annoying to work with?” While this is true, I did elevate the good birch plywood with a 4’x4′ of cheap particle board, which served two purposes. One, it elevated the good wood out of the water, so it wouldn’t just soak up all of the water. And two, it minimized blowout on the back side of the birch by supporting the drill bit when it plunged through to the other side. The birch still did get quite wet, but I paused and pulled it off as often as was convenient and then rinsed off the abrasive and did by best to dry it with a blow gun.

Almost done water jetting!

I could say the program ran without issue, but that would be incorrect. I ran into an issue with some of the consumable parts of the water jet that had worn out and needed to be replaced. Luckily, the night shift production water jet operator was around and helped me out. Thanks a lot Alex!

Done Water Jetting!

After all the cutting and fixing, it was done! All told, it took be roughly 4 hours of time after work one night. The above picture is how much I cut out from the 5’x5′ of plywood. I used a lot of it! There was only one minor/huge problem at this point in the night. I was locked out of the office, and I sort of left my bag and car keys in there. Not a smart move, and I had to call Ben to help me out. (seriously, thank you Ben).

Dry/Wet Fit




After that small problem, I drove home and immediately tried putting it together. I couldn’t resist! It all fit perfectly, it even had a slight friction fit to the joints. I designed in tolerance in the slots, but the wood was damp and slightly swollen, so it fit together very tightly. I really liked the result. The effect was awesome and it fit in the space I designed it for great. The next step was going to be lots of sanding and staining.



Ugh. This part was very challenging to get through. It took me a full weekend to sand everything up to 220 grit and then stain (about 4 hours per process). Again this step is best illustrated with pictures.

Parts Laid out of sanding

There was a considerable amount of surface area to sand on this project. (a little less than 50 sqft based on the starting piece of plywood). I purchased a random orbital sander for this job (oh man if I did not) and got to sanding. I started with 120 grit and then went to 220 grit. Very uneventful work, but very necessary to get a nice stain. When I finished, every surface in my work shop room was covered in a nice layer of dust.

Staining Step


The next step was staining. Being a space themed thing, I wanted to use a dark stain. I chose “Dark Walnut” stain. I think  I went too dark with the stain, but it did still look quite good. As you can see in the picture, what you need for this step is one pile of sanded table, wood stain, and a brush. I brushed on the stain, let it sit for a few minutes, then wiped off the excess with a shop cloth. I chose to stain before assembly because of the difficulty involved in getting inside all the cubbies once it was assembled.

Close up of the staining

Assembly Step

The next step after letting the stain dry was assembly. Take one pile of stained table, wood glue, and clamps, lots of clamps. I also used a pneumatic brad nailer to connect some parts.Clamping for assembly


Assembly was a little complicated, but once I got one part of it set up, it was just a matter of slotting in the parts to their appropriate receiving slots. I either used glue or brad nails (in hidden locations) to assemble the whole thing. The top perimeter employed only glue, as I did not want to show any brads. The result of wanting to do that meant lots and lots of clamps. (about 15 in the above picture). I let it dry over night to ensure the bonds were good.

Top test fit

When it was finished drying, I put in the 6″ diameter mirrored sphere ( and put on the acrylic top surface for it. Another problem was encountered here. The Problem was two fold: I had already cut the top for the un-glued table, so it was incorrectly sized, and I neglected to square the top of the table when I glued it, so it was slightly a parallelogram. I was able to fix these problems by setting up my dremel as a plunge router and cutting it to the correct size.Top clear coat


After sorting out the table top problem, I applied a clear gloss polyurethane to the entire table to protect it. I only did one coat, because I wanted to have the table ready in time for a dinner I was hosting, but i do plan on revisiting that in the near future for more coats. Additionally, the acrylic top is not the final plan. Sooner or later, I plan on getting a nice big piece of tempered glass for the surface. The original plan was to fit it down in the perimeter of the top, but since it is now a parallelogram and not a rectangle, I will place the surface on top of everything with a slight overhang.

I am very happy with how it turned out in the end. It provides functional space and at the same time looks awesome.





This project was not terribly expensive to execute, but my free access to the water jet does skew that slightly. I won’t include tool costs, I already had some and needed others. The cost break-down is as follows:

5’x5’x0.5″ Baltic Birch Plywood – $35

6″ Diameter Mirror Sphere – $19

2’x4’x0.25″ Clear Acrylic – $50

Stain and Clear Polyurethane – roughly $25

{(planned) 2’x3’x.375″ Tempered Glass – $130}

Total Cost of the Project: about $140-220.


Perhaps not cheaper than just buying a coffee table at a store, but definitely way cooler.

Leave a comment

Twisted Chess Set, Part II

Finished Product First!

Finished Product First!

A couple months later, I am finally getting a chance to finish up the chess set and make a board for it! Check out Part I of this project here.



I had a design in mind for a while using 1/4″ black acrylic and 1/8″ white acrylic, but it was going to be too big for the work area of my shapeoko 2. I also was not terribly excited about the idea of re-positioning the work piece or cutting small pieces and making it assemble. There would have been seams or inconsistencies that would have looked poorly. Fortunately for me, at my new job, we have a CO2 laser cutter/engraver. Just when I thought it couldn’t get any better with access to a water jet, I also get to play with a laser!

I scrapped my old design and decided to just use the 1/4″ black acrylic with the white spaces engraved onto it. Since I would be laser engraving it, I could easily add whatever hatching or small feature that I wanted to the engrave layer.

The chess pieces are based on hexagons, so I wanted to keep a similar theme for the board. I tried hexagon shaped spaces, but that didn’t really make sense on a rectangular board. Instead, I opted for a hexagon or honey-combed hatch in squares with rounded corners.

The board with it’s squares and border was from my original board design, which I laid out with patterns in SolidWorks. After that, I exported it as a .DXF and opened it in inkscape to add the lettering and numbering. The final step was exporting it as a .DXF yet again and then opening it in DraftSight to add the hatch pattern. After exploding the hatch and text down to lines, I was able to open it in the laser cutter software, SmartCarve.

Grab the source files here.

Screen Shot of the chess board design from DraftSight.

Screen Shot of the chess board design from DraftSight.



Laser cutting was fairly simple. I had only used it a few times, and was kind of guessing on what speeds and power to use to do it correctly. I tried 40 mm/s, 20% power for the engrave and 5 mm/s 95% power for the cuts. Unfortunately, the engrave was not powerful enough, and it only very lightly marked the black acrylic. The cut worked just fine however.

Near the beginning of the engraving. The hexagon hatches contain a LOT of lines.

Near the beginning of the engraving. The hexagon hatches contain a LOT of lines.

Later on in the engraving process. The hexagons took forever.

Later on in the engraving process. The hexagons took forever.

About 30 minutes later, the program was done. It appeared to be marked and cut just fine, so I removed it from the work area. It wasn’t until after I removed some of the masking that I noticed it wasn’t marked that well.

After the laser finished, I removed the mask on a majority of the areas. You can see the 'H' on the top, or rather, can't see it, due to a under-powered etch pass.

After the laser finished, I removed the mask on a majority of the areas. You can see the ‘H’ on the top, or rather, can’t see it, due to a under-powered etch pass.


Close up shot of the 'H' that did not etch enough to be seen. The only reason it is even semi-visible in this shot is because of camera flash.

Close up shot of the ‘H’ that did not etch enough to be seen. The only reason it is even semi-visible in this shot is because of camera flash.

The etch didn’t work, so for this one I left the masking on some areas. It provides a nice looking effect, but is not as durable as it could be. I may re-do this piece using clear acrylic to ensure a good engraving.

Since part one of this post, I also printed another half the set in black PLA. I still do not like the yellow PLA pieces, so soon I will print out a set in white PLA.





Leave a comment

Final Design Project: Staple Gun Reverse Engineering with SolidWorks

Heavy Duty Tacker Render 1

Swingline Heavy Duty Tacker #800 SolidWorks Render

Three months later, I am finally getting back to writing about some of the things that I’ve been working on. This one is all CAD based, and was a collaboration between myself and two others. We were tasked with reverse engineering a product using SolidWorks, and we chose to model an old staple gun that I had laying around.


This project was for ME 5680 at the Ohio State University. The class title is “Computer aided design and manufacturing”, and is a highly sought after technical elective at OSU. If you are an ME at OSU, you should so totally take it. It is awesome. You get to learn SolidWorks, get certified in SolidWorks AND the labs use Tormach CNC machines! Seriously, take the course.

We were required to find a product that we could disassemble, measure (fairly) accurately, and have a minimum number of parts with some complexity. The project began around the 3rd or 4th week of our 15 week semester, and we managed to keep a very steady pace throughout the time we had to work on it. My group members for this project were Caitlin Boone and Hamed AlGendy, both of whom did great work the whole semester, thanks guys! This was one of two large project groups I was a part of during my final semester and both of them were seriously great. Both projects did such a great job of pacing the projects that we barely had any late night or weekend meetings and we even finished with spare time before the due date!

The components making up the hammer sub-assembly.

The components making up the hammer sub-assembly.

The components making up the magazine sub-assembly.

The components making up the magazine sub-assembly.

The components making up the Body sub-assembly.

The components making up the Body sub-assembly.

After taking apart the stapler, we counted 21 unique parts, most of which were fairly complex sheet metal components. This satisfied the requirements for the project so we split up the stapler into three main sub-assemblies: the hammer, the magazine, and the body. My portion was the hammer assembly. After we split up the parts we proceeded to work on the project over the next 11 weeks very consistently by dedicating 3 hours a week to meeting with each other and working on our parts.


Modeling the hammer assembly was very interesting. I had used sheet metal tools previously, but this sub-assembly had some very complicated components in terms of features and dimensions.

The hammer itself is only a flat piece of steel with a couple squares punched out of it. The stops, both the steel and rubber one, were simple shapes again. Where this one got interesting is the hammer carriage, the springs, and the hammer body.

The hammer uses two coaxial springs, and I wanted to model them in such a way that they would animate when I moved the hammer carriage in the model. I accomplished this by defining the helix of the spring with it’s center line, but leaving the top end of the line undefined. By defining the end of the line with its mating component at the assembly level, the springs would recalculate each time the hammer carriage was moved, meaning animation was possible! The spring itself is a sweep feature using the previously mentioned helical sketch, with special sweeps and cuts on the ends to create the ground/flattened ends like a real spring.

The hammer body and the hammer carriage are both purely sheet metal features. This speaks to how they are manufactured, and means that the part starts it’s life out as a flat sheet of metal. Specific to these two components, the manufacturing methods involved are most likely performed by die presses to deform, punch, bend or cut the sheet metal into the final form. The important part about this concerning a 3D design is that it is important to get the flat pattern of the part prior to manufacturing so that you know how much material to use. In SolidWorks, if you model it correctly, the software can do this unfolding for you! Staying true to the manufacturing process and wanting it to be as accurate as possible, I used the sheet metal tools in SolidWorks.

The main part of the hammer body is a series of sheet metal bends and plain old cuts. This wasn’t too crazy until I got to the back end where there was a curved bend, which is a little bit of an oddity as far as the flow of material is concerned when you actually form it. SolidWorks handled it just fine and it still worked great with unfolding the feature.

The rest of the features were made with forming tools. Forming tools in SolidWorks allows you to create the solid that defines the press tool that would be used in real life, place it on your model where it would be used, and then actually execute the forming operation to deform the material and show how it looks in the model. The key to using this instead of just making it a solid extruded feature goes back to the sheet metal part stuff. Sheet metal parts have the same material thickness throughout the whole thing (except where material has flowed due to excessive deformation). This means that the forming tool features don’t add or remove thickness to the material, but “displaces” or deforms the metal, just like it would in real life.

Whew that’s a lot of text. Enjoy some renders of my parts instead, they are more fun to look at!

Transparent hammer body to allow viewing the internals.

Transparent hammer body to allow viewing the internals.

Hammer Body Render - reverse side

This render shows a cut away of hammer, showing the inner components.

This render shows a cut away of hammer, showing the inner components.


Caitlin and Hamed did a great job on their sub-assemblies as well. One of the mating parts that Hamed and I worked on separately even fit together within .05″! That is pretty impressive considering the difficulty of measuring some of the features.

Assembly Render 2Assembly RenderHeavy Duty Tacker Render 1

I took care of the final complete assembly, and animating the model for our presentation. I spent probably 2-3 hours just on fighting with the motion study feature in SolidWorks to make the animation play in the way I wanted it to. I was pretty happy with the result. The animation consisted of the stapler disassembling, reassembling and then finally shooting a staple (with all the inner workings showing and working as they really do!).


We did make a fancy powerpoint presentation that you can take a look if you are interested. I felt we did very well, and our professor must have agreed, because we got an A on the project! Working on this project with Hamed and Caitlin was one of the most pleasant project group experiences that I had while at school. We kept such a great pace and the quality of work was just wonderful.

V4Swingline Heavy Duty Tacker Presentation – Final

Oh. I also uploaded the assembly to GrabCAD just in case anyone needs a super complex staple gun model for some reason. Unnecessary levels of detail!


Leave a comment

FPV RC Rover Project Updates


MAKEathon Robot Render

Rover Project Updates



Quick update on the FPV RC Rover that I helped to make at the 2015 OSU MAKEathon a couple weeks back, there’s now an instructable and a project poster! I also cleaned up the 3D model a little bit for the instructable. A couple pretty renderings of the cleaned up model can be seen right above and at the end of this post.

While at the event, we were told that our project could qualify for a couple instructables contests, so we thought it’d be a great idea to do a write up and enter those contests! The link to the instructable is here: . Go check it out! And if you like it, vote for us in the contests that it is in.

Aside from that, one of my Mechanical Engineering professors saw me at the MAKEathon and requested that I bring the robot to the design showcase that the Mechanical Engineering department was putting on that week. To that end, I put together a quick poster illustrating our project and giving some basic information about it. The best part is probably the timeline. It gives a fairly accurate depiction of how much time it took for each step of the process. At the showcase event, there were plenty of interested passerby who wanted to give the robot a whirl. It was awesome seeing everyone enjoy our work!



First Person View, Remote Controlled Rover Poster 30x40




More Renders

MAKEathon Robot Explosion Render

MAKEathon Robot Render underside

Camera Gimbal Render 2

Camera Gimbal Explosion Render

Leave a comment

FPV RC Rover

First Person View Remote Control Rover

This past weekend I participated in the 2015 OSU MAKEathon put on by the electronics club here at OSU. The event was 24 hours of just makin’ stuff. I grouped up with a couple of other makers, and we decided to build a first person view remote controlled rover!

Our FPV RC Rover along with the RC transmitter and video receiver.

Our FPV RC Rover along with the RC transmitter and video receiver.

I packed up my 3D printer and some tools and headed onto campus for the event. I had gathered two other guys from the OSU makers club (Brent and Ryan) to work with me, but I was also lucky to run into another guy (Aaron) at the event who wanted to work with us. We accidentally ended up with a very nice multidisciplinary blend: two Mechanical engineers, a Computer science engineer and a Electrical Engineer. Between the four of us, we had a nice selection of tools and electrical components.

Buckled up and ready to go to the MAKEathon!

Buckled up and ready to go to the MAKEathon!


We came to the event knowing we wanted to make a rover and use a pair of FPV goggles on it. Ryan purchased a spektrum DX6i radio transmitter and a fatshark FPV camera and goggles for use in this project (he said it was a perfect excuse to purchase these items). We also came to the event with a pair of continuous rotation servos to use for the drive wheels. We had planned do use two drive wheels in a tank drive fashion and a third peg/roller to make the bot stable. Brent and I started doing the layout and detail design while Aaron and Ryan started working on the wiring and electronics.

We decided on putting the two drive wheels in front and the peg/roller centered in the back. We also decided that we would mount the FPV camera on a pan/tilt gimbal centered in the front of the rover. Brent and I mocked up many of the components to check the layout of everything on the rover. Using these mock-ups, we designed the frame, camera gimbal and wheels of the rover. All of these components were to be 3D printed, so we needed to make sure that all the parts had the right fit and provided the functions that we wanted. We spent roughly 2 hours modeling everything before we started printing.

CAD Render of the rover.

CAD Render of the rover.

The frame was the largest part and was designed to have a place for every component that we needed. The frame had a cutout for the camera pan servo and brackets integrated into it to mount the drive servos. There was also a hole in the rear for the peg/roller and holes all over the surface for accessories or other tie down points.

Underside of the robot.

Underside of the robot.

The camera gimbal was probably the most complicated part that we designed. The U-shaped piece of the camera gimbal contains cutouts on the bottom for the servo horn and cutouts on the side for the tilt servo. There is also a hole on the side to stabilize the camera mount. The camera mount itself is simply a rectangular piece with a hole cutout in the center, and a pocketed area on the rear to hold the camera. There is also a pocket the shape of the servo horn on the side as well as a hole that lines up with the stabilization hole in the pan mount.

Render of the camera gimbal assembly.

Render of the camera gimbal assembly.


After the detail design, I converted the files to .STL and started printing. The first components that we printed were the camera gimbal assembly. Both the pan bracket and the tilt piece printed in roughly an hour and half. After that, we assembled it with the servos and camera to test.

The camera gimbal assembly after printing.

The camera gimbal assembly after printing.

The pan/tilt function was practically plug and play with the RF transmitter and it worked extremely well. And of course we needed to test it! The grin people got when controlling it through the goggles was commonplace all night!

After the camera gimbal, we printed the frame. The frame was roughly a 5×6″ rectangle which qualifies it for the largest part that I have printed using my Prusa. A 3D printer is actually the wrong tool to make this frame, my shapeoko would have done it quicker and better, but I think that the printer was more versatile a tool to bring with me. This was easily the longest print to create this part. It took roughly four hours to complete. Even generating the g-code from slic3r took about a half an hour! This is very likely due to all the holes I had added in the frame and that the main plate was 1/4″ thick (way overkill).

Printed frame with the drive servos installed.

Printed frame with the drive servos installed.

After the frame, I printed the wheels. The wheels were 3″ diameter and 3/4″ thick, and looked roughly like hockey pucks. These took 2 hours each to print, and unfortunately, I did not take any pictures before we assembled them onto the bot. The printed wheels were a little smooth, so we wrapped them with rubber bands for additional traction.

The final thing I printed was the roller ball for the rear peg. It is simply a 3/4″ hemisphere with a hex nut trap on the flat face. The rear peg screws into the frame and then the roller ball. With it set up this way, we could adjust the height of the rear peg and thus adjust the angle that the rover makes with the ground. Originally, we were simply using the head of a machine screw as the rear roller, but we found that it caught on things and added some unnecessary friction to driving.


While Brent and I were designing and printing the physical parts, Aaron and Ryan were working on the electronics and coding to make it all work. I did not have direct involvement in this process, but I do know generally what they did. This project would not have been possible without Aaron and Ryan’s expertise with electronics and coding.

The DX6i transmitter we used had several servo channels that can be controlled by the sticks. However, being a transmitter designed for quad copters and RC planes, it didn’t work just plug and play with our setup. What we needed to do was take the servo PWM control signal from the reciever, process them on an arduino and then re-transmit them to the drive servos. After doing this, the drive servos performed as desired: the right stick of the transmitter controlled forward, backward and turning of the rover.

After we installed this system, the drive servos would just spin at idle (no movement of the stick). Moving the stick made the servos drive like we intended, so we had an issue with idle operation (considering we did not want the rover to move when given no signals). Ryan and Aaron checked the code for quite a while to try to determine the problem, but in the end it was determined that the servos had a trim pot on them to control how they function at idle.

Everything on the rover was powered by a 11.1V 3300mAh Li-poly battery. This battery was originally intended only to power the FPV system, but we tapped into it to provide power for everything. We used a voltage regulator to step down the power to an appropriate level for both the arduino and servos.


The wheels finished printing at roughly 2:30 am, and we began assembling the rover immediately. The drive servos fit right in the brackets and the camera gimbal fit right into its slot as well. The FPV transmitter and the Li-Poly battery were zip tied together on the top and bottom respectively. Next, the perfboard with the servo receiver and arduino nano were hot-glued with standoffs on the top side. With all the components installed, we wired everything up and provided enough slack for the gimbal to move easily.

The rover after assembly!

The rover after assembly!

And of course, video of it in action!


The rover performed wonderfully. It was immensely satisfying to see it work and to use it. Once dialed in, the drive system worked flawlessly and provided zero-point turns and easy to use controls. The camera gimbal just by itself was immensely satisfying to use, and once paired with the rover drive system, immensely entertaining to use.

The servos we used for the drive motors were not very fast, but they did move the rover at an acceptable speed, and were fairly easy to control. The pan/tilt gimbal for the camera provided us with roughly 90 degrees of pan and tilt in each direction, which was acceptable for driving and looking around. Our transmitter is capable of up to one mile, which we did not test (though we did drive it down several hallways and go down the elevator with it). Battery life was pretty good, we drove it around for roughly a half hour before we needed to bring it back and re-charge the batteries.

To anyone reading this from OSU, I will have the rover with me at the Mechanical Engineering Capstone fair on Friday, April 24th from 2-5pm in Scott labs. Come on by to check it out!


A component of the MAKEathon was that it was also a competition. Faculty/sponsor judges were around during the whole event observing and also went around at the end to judge every project, science fair style. Our project tied for 2nd place overall! We tied with a group that mechanized/computerized an etch-a-sketch to draw pictures. The winning team used DC motors with encoders, and a DC motor pump to make a pancake printer on top of a electric skillet.

Everything about this competition was great. Everyone involved, both volunteers and other participants, were very enthusiastic to see what everyone else was working on. The whole environment was awesome. I would definitely recommend other makers to participate in events like this.

I think our rover was a crowd favorite among the participants. Everyone who stopped by to talk with us or try it out had a constant grin on their face while talking with us or using the rover. It was hard not to while wearing the goggles and driving it around.

At the end of the event, Ryan took the rover home since he had purchased all the components for it. I will say that I am definitely jealous! This thing was crazy fun, and I may have to try to make one for myself!


FPV Rover 1

FPV Rover 2

FPV Rover 3

FPV Rover 4

FPV Rover 5

FPV Rover 6

FPV Rover 6

1 Comment

Shapeoko Spring Cleaning and Touch Probe

After all the cleaning and surfacing.

After all the cleaning and surfacing.

While the weather finally made a turn for spring here in Ohio, I decided it was time to do some cleanup work on my shapeoko 2. My waste board was good and thoroughly covered in cuts from other jobs, and my v-wheels were just covered in junk. Thus, my Sunday afternoon was spent cleaning things up and making my machine work its best again. Additionally, I implemented the touch probe functionality in GRBL, allowing for precise zeroing of the Z-axis.

Waste Board Resurfacing


Having just passed a year of owning my Shapeoko 2, my waste boards were quite marked up with over-cuts from old jobs. I’ve been using the original MDF waste board halves from the stock shapeoko bed, and had already replaced and flipped both halves as many times as I could. All the cuts and grooves made it annoying to clean, and less effective when using double sided tape.

About 5 months worth of projects on this wasteboard.

About 5 months worth of projects on this wasteboard.

My first waste board. Probably closer to 7 months of projects on here.

My first waste board. Probably closer to 7 months of projects on here.

It was kind of interesting to take a look at these waste boards and remember all the different projects I’ve worked on with this machine. I could recognize outlines from my rostock bases, parts for my E-stop bracket, and parts from my VR headset. With what I was planning to do I would basically be erasing all of this evidence.

My plan was to take one of these boards, cut it down to ~13 inches, mill in four new counter bored mounting holes and then resurface the whole thing. I ran downstairs quickly to use the band saw and cut it to size, then mounted it on top of my old waste board like it was just another piece of stock.

To make the counter bored holes, I setup two spiral drill operations in EstlCAM. I jogged the machine into position above where I wanted the hole, then ran the quick job. After that I jogged in the +Y as far as I could and then ran the counter bore job again. Then I jogged in the +X 180mm (the slots of my alum bed are 20mm center to center) and cut another hole. Finally, I jogged in the -Y to the final hole location. The result was a rectangular hole pattern that would fit onto my aluminum extrusion bed very nicely. I did make a small mistake with my first hole location, so there a couple extra holes that don’t actually line up with anything. Fortunately, after I turned the board around, they are outside the working area anyways. I then removed my old waste board and loaded in the new one.

Cut down waste board installed and ready to be resurfaced.

Cut down waste board installed and ready to be resurfaced.

My plan was to take roughly .125″ off the top to create a smooth, and perfectly level (to the spindle of the machine anyways) surface. I did this by taking two 0.05″ passes, and then a .02″ finishing pass. Due to geometry of the machine, I couldn’t reach all of the space, even with the 3/4″ router bit I was using. As a result, I did the first two .05″ passes in the orientation shown above, then cut a little deeper on the front edge and right edges (which will be out of the working area after rotating). After that, I rotated the whole board 180 degrees and then cut the final .02″ surfacing pass. The result was a nice, smooth and level surface. Additionally, I cantilevered the board over the edge a little and then edge cut it with a 1/8″ end mill to clean up that edge too.

The new waste board after surfacing.

The new waste board after surfacing.

The board was fairly smooth at this point. You can sort of see some lines already in the picture above, and I believe that is due to my spindle being slightly tilted. I cleaned it up with some 1000 grit sandpaper, and considered it “good enough.” At this point, the M5 x 12mm  screws I was using were too close to the surface for comfort. Luckily, I had some low profile M5 x 10mm screws that worked just perfectly.

The final operation I did to the waste board was to engrave a 9″ x 12″ work area with .5″ grid spacing. I chose to use a V-bit do do this, and only used a .025″ depth of cut. I was very happy to see the bit cut over the entire surface, meaning the bed is quite level (at least within .025″). The lines are rather light, but they should still work. The origin is actually at X=0, Y=.2, but that is an easy thing to jog to for each job start.

9" x 12" work area with 0.5" grid markings. I wrote in the dimensions by hand because it was taking forever to add them in the CAM.

9″ x 12″ work area with 0.5″ grid markings. I wrote in the dimensions by hand because it was taking forever to add them in the CAM.

The new board opens up a couple more of my bed slots, which should allow for better work holding placement. Also, since it is relatively flat, I should be able to enter the exact height of stock now and get little to no over-cut marks!

I also took a time lapse of the whole process, check it out:


V-Wheel Cleanup



My machine was quite messy from some recent jobs (specifically cutting blue foam) and as a result the v-wheels would bind up. My standard method for cleaning the v-wheels was to use my fingernail in the vee while moving the gantry. This sort of worked, but did not really clean the wheels enough. Previously, when my machine was torn apart, I used a wet wash cloth to clean them off, but I wasn’t about to take apart my whole machine every time I needed to clean my wheels. What I ended up doing was using an old toothbrush to scrub the wheels clean. This worked surprisingly well! The bristles were stiff enough to scrub away the gunk in the wheels but soft enough to get in there still. Suffice to say, I will be keeping a toothbrush near my shapeoko now.

Dont forget to brush your v-wheels!

Dont forget to brush your v-wheels!

The top wheel has been brushed, the bottom wheel has not.

The top wheel has been brushed, the bottom wheel has not.


Touch Probe



The final part of this post is about how I implemented a touch probe to zero the z-axis. The basic idea is that an aluminum block is wired to ground, and the end mill is wired to the probe input via an alligator clip, then the z-axis slowly descends until the bit contacts the aluminum block and thus pulls the probe signal to ground. When this happens, the z-axis stops and then you can tell the machine that the current position is a certain height (the height of the plate). This gives you a nice easy way to zero the z-axis, as long as your touch plate is suitably flat and consistent. Luckily, as a student in mechanical engineering, I have access to a machine shop! The student machine shop here even has some scrap, which is all I really needed to make this touch plate. As it turned out, it was easier to ask a friend to machine it for me, since it has been a little while since I used a Bridgeport. I ended up trading some 3D printing services to him for machining services. Thanks John! The block he machined for me was fly cut on two sides to make it very flat, and he managed to make it very close to the target height of .500″ (I measured it at .501″, but my calipers are fairly cheap). He also drilled and tapped an M5 x 0.7 threaded hole to attach the ground lead.

Touch Probe in usage position.

Touch Probe in usage position.

The positive probe lead (red wire) is connected to the A5 pin on the arduino, and the negative lead (black wire) is connected to ground. I routed a wire from the arduino around the back of the shapeoko up the the front left corner and the zip tied it in place. There is enough slack in the line at the end to probe anywhere on my work area. When not in use, the plate fits in the small space in the front of my machine. My plan is to replace these moving wires with a retractable headphone cable to make it wind up when not in use.

To use the probe, a fairly simple G-Code command is used. G38.2 is that command.

G38.2 Zxxx Fyyy

As I understand it, there are two parameters to use with G38.2 to make it function. The Zxxx is by how much it will move before it’s declared a “probe failure.” The Fyyy declares what feed rate the probe action will be done at. The exact parameters I use are:

G38.2 Z-0.5 F1

This starts a straight probe operation in the Z-axis, probing no more than .5″ down at a feed rate of 1 inch/min. (or .5 mm down and 1 mm/min if you’re in metric units). After probing, you need to set the Z-height to the height of the touch probe and then raise the z axis to remove the probe plate and clipped lead.

I use the straight probe command in a UGCS macro to automate the process even more. My macro starts a straight probe, then sets the z-height to that of the probe plate and then finally raises the z axis .25″.

G38.2 Z-0.5 F1; G92 Z0.501; G91 G0 Z0.25 %Straight Probe, Set Zero, Pull up quarter inch.

After running the macro, unclip the end mill lead and move the touch plate away. If X and Y are already zeroed, simply starting a cutting job should now yield a good job.

GRBL also now supports TLO – Tool Length Offsets. This means that you can enter the lengths of tools relative to each other and after precisely zeroing with the touch plate, tool changes don’t require re-zeroing, just changing the physical bits. Doing this does require that you can load the end mills with the exact same lengths each time, which does pose a problem with the style of collet I use. Some of my end mills have stop collars on them, which does provide this consistent level, but not every tool I have has this feature. My plan of action to implement this is to mill out some stop collars out of HDPE with a slightly undersized hole, then heat this collar in boiling water and then install onto the shaft of the end mills. Hopefully, this will create a tight fit and a reliable reference point for my end mills that don’t have them already. More on this to come.