The Big Think

October 12, 2015

R2 Update

Filed under: The R2 Project — jasony @ 8:32 pm

Tonight I spent four hours at TechShop drilling eight holes. Not even full holes. I had to set up the manual mill, clamp in the four curved front utility arm pieces, centerfind each hole and then drill a .045″ countersink on each one with a .50″ bit. Each hole took about 20 minutes. With setup, cleanup, and a few side trips to help other members with the WJ (as well as answer questions about the laser and software) it took me four hours.

It’s fun being an instructor at TechShop, and fun being a well-known one. But sometimes it can be distracting! It took me a full 30 minutes to get to the mill and put my headphones on (the universal T.S. “leave-me-alone-I’m-working” signal that most people seem to understand).

So now all of the frame pieces are cut, drilled, and tapped. I have a little bit of wet sanding to do (about an hour) and then I can start in on the high-gloss polishing. I bought two different polishing wheels and a whole slew of compounds to use. I’m hoping to have the frame ready to go to the anodizers in a few weeks. Then assembly and I’ll then move on to the legs.

Pics will follow soon! I try not to touch my iPhone too much when I’m in the machine shop since there are so many tiny aluminum shards flying around.

August 30, 2015


Filed under: Movies,The R2 Project — jasony @ 5:40 pm

Great time seeing the movie and catching up with good friend Giles. I was impressed with the way the screenwriters handled the potential for the movie to descend into silliness and farce. There were only a few times that I felt sorry for the actors having to deliver such goofy lines. But the sense of humor and obviously over-the-top story worked. The screenwriters, director, and actors pulled it off. Way to go!

We saw it at the Alamo, a place that, sadly, Erin and I haven’t gone to enough. To sit and watch a movie where you’re guaranteed to not have talkers or small screens in your face was nice. And sipping a Guinness in a comfortable chair really added to the experience.

BTW, Giles geeked out on R2 so for convenience I am posting a short-link to my R2 posts. Enjoy!

July 6, 2015

Skirt Gap

Filed under: The R2 Project — jasony @ 10:55 am

I got rid of the gap! I spent a couple of hours on the manual mill and lathe last night (I’ve gotten pretty good at setting those machines up and have fun working on them). I was able to create pockets on the backside of the thick piece that accept the short rods. I also shortened the rods slightly to tighten up the gap. There was a small problem with insetting the screws (the non-threaded part of the screw punches too far through the hole and keeps the rod from tightening down all the way). I’ll go back in today and taper the threaded screw hole in the short rods to get rid of that problem.

What this means is that I can now JB Weld the thick part onto the skirt surround. I’m one step closer to finishing the skirt. After the JB Weld dries I’ll apply LabMetal and let that cure (it should be here today). Then I can start making the little vertical details (this pic grabbed from online):


I have to cut the bar so that the top slope is 53 degrees and the bottom is 36 degrees, then I’ll drill 2 holes and tap them, then temporarily fasten them in the exact location onto the surround, use a center punch to mark the holes, and then drill those holes out. Then it’s a simple matter to screw the pieces to the surround from the inside. I’ll have to smear Lab Metal on the screw holes to cover them up and then sand everything to make it look good.

It’s a lot of work but each step is straightforward. I have to hand tap 24 holes with #4/40 1/4″ machine screws (hmm… have to buy some of those). By this point tapping screws scares me not a bit. I’m really good at it. I just put on a good podcast and get into a nice Zen state and a few hours later they’re done.

I’m shooting to have the skirt done within the next 2 weeks…. hopefully.

I need to start some research into what I’ll need to buy to get the finishing done on the frame. I’m going to sand and polish every piece until it sparkles and then get it all anodized. Then the lasering.

Hopefully it’ll be done by Sept 1st but I’m not going to rush it. It has to be right, not just fast.

July 2, 2015

More Skirt

Filed under: The R2 Project — jasony @ 5:57 pm

Went back to TechShop today for about four hours and recut the skirt parts. Fortunately the design work was done and I already had the free metal so all it took was about 30 minutes to set up the machine and cut (the cutting time itself was exactly 5 minutes). So at a cost of $15 I remade the surround parts.

Next, I turned to the slip roller and carefully rolled them out so that they’d match the curve of the thick piece (in the photo below). It took figuring out a little jig but I’d already done that, so even that part went fast. Unfortunately, the machine grabbed one of the curved pieces and chewed it up pretty badly. Sure, I could have made it work, but did I mention I’m redoing the skirt? I was going to just use it but didn’t want a rehash of the first time through when I kept plowing forward despite errors, only to have to start again so… more waterjet! I got the key, set up the blank, zeroed the machine, and had the part cut inside of five minutes. It took 1:18. Such a short cut, in fact, that the desk staff just let me have it free! Between the free metal and this free waterjet part, all it cost me to redo that specific part was my time: about ten minues.

I… very carefully… slipped rolled this part until it was perfect. And I mean perfect. It’s by far the best made part on the skirt surround. In fact, I was so happy with it that I grabbed the other curved part (the one that was almost perfect) and decided to run it through the machine so that it would match. I’d have two perfect pieces!

Big mistake.

Did you know that aluminum, when subjected to multiple rolling forces, slowly gets harder? Yes it does. And now I know it, too. Because when I sent the formerly-almost-perfect piece through the slip roller it got stalled and bent out of shape. Not a lot out of shape, but just enough to throw the alignment of the entire surround off. AARGH!! In a panic, I tried to correct the mistake by slip rolling it again but only made the problem worse (and the aluminum even harder). So… I stopped, took a breath, and manually re-bent the piece. I got it back to aaaalmost as good as it was before and then called it done. It’s close enough that I will (hopefully) be able to put some pressure/clamps on it when it JB Welds and then repair any small gaps with Lab Metal. If not, there’s always Skirt 3.0!

The surrounds are drying at TechShop right now. I’ll go back in at 10 or so tonight and move them to my locker after the JB sets. Then tomorrow I’ll go and spend some quality time on the mill fixing the gap issues. Here’s a pic:


Overall it took about 30 hours to design, cut, and assemble my horrible looking first version of the skirt. The second version took about 4 hours (though it’s not quite done yet). I’m glad I redid it but I’m sure ready to have this part behind me. Lots to do on it, though. Stay tuned.

R2 Skirt Redux, Redux

Filed under: The R2 Project — jasony @ 8:46 am

I used the wrong plans. The skirt is wrong. Yes, after all that work – 30 hours of work over a month or so of days – it turns out that R2’s skirt just isn’t going to work. There are clearance issues when I put the skirt on the frame that I just can’t kludge together and make work. Here’s a pic:


There shouldn’t be that gap between the skirt and the frame (keep in mind the whole assembly is upside down in this photo).

It turns out that I was using a slightly modified version of the skirt made to fit a different frame. The external dimensions are correct but the thick piece on the top of that photo above causes alignment issues. I should have used bent sheet metal that would have created a pocket where the posts rest on the skirt, thereby getting rid of that clearance.

So I’m going to have to start completely over on the skirt and remake it. Fortunately I have the waterjet files and the side pieces are correct. I also scored a free piece of metal from Metals4You that was the right size and shape from their cutoff bin. They told me to just take it (guess they appreciate all the metal I’ve already bought from them). The waterjet time will cost about $12 so that’s not a huge loss. More JB Weld cost less than ten bucks.

I will have to go back in and spend some time milling up the thick bottom plate to create pockets for the posts to remove that gap. I’ll also have to shorten the posts slightly (about .2″) to tighten the gap up. Those posts aren’t Canonical since they’re inside supports so I still have a screen-accurate droid. The important part is the shape and size of the visible parts of the skirt.

To be honest I’ve suspected I was going to have to remake the thing for a couple of weeks now. The old skirt was progressively looking worse and worse as I made mistakes and kludged together fixes that caused more errors. Now that I’ve been through the process once I’m confident that I can quickly cut, bend, and assemble the new skirt with minimal fuss. And I still have the old skirt to practice powder coating on so that’s good.

I keep telling myself that I have plenty of time and no real deadline on this (though I would like R2 to be mostly finished by my 50th birthday in Sept. 2019). This is supposed to be fun and a learning experience. I guess that means that I’ll occasionally have to go remake something as I learn my way around mistakes the first time through.

June 21, 2015

JB Fillets

Filed under: The R2 Project — jasony @ 11:24 pm

More Skirt work. I went to TechShop tonight and took a look at the skirt brackets. The JB Weld I set up a few nights ago had cured and looked pretty good. Well, it actually looked like warmed over cow manure (grey cow manure), but it was on the underside/inside of the skirt and will never be seen. As long as the brackets hold and the skirt surrounds keep their shape I’ll be good.

I squeezed a whole tube of JB Weld into a plate and mixed it up and then used a plastic spoon to shape it into fillets on the underside where the pieces met. The brackets I’d JB’d into place a few days ago were holding everything in place while the fresh adhesive filled in and supported the fillets. It was a HUGE mess on the inside but, again, nobody will see that.

So then came the clamping. When we tried to weld the pieces a few weeks ago (and subsequently burned a hole in two of the pieces), we very slightly misaligned the pieces right as the welding bead took. It’s not much (about 2mm), but it’s enough to make the whole skirt a little crooked. You can force it into flatness but it takes some pressure.

Before the JB started to set up I got some clamps and a board and flattened out the slight warping in the skirt. Just as I was tightening the last clamp — **PING!!!** — the bracket that I’d tried to braze with the Alumaloy cut loose. Complete failure.


So now I had fresh, wet JB Weld drying and a structural failure. Well, the nice thing about being backed into a corner is that there’s only one way to go. Onward! I did my best to clamp things down and align all the parts. The nice thing was that, with the failed joint, things tended to straighten out a little bit. Here’s a pic of the failure point. You can also see the great big grey mess of we JB Weld on the inside:



I let the JB set for six hours and went back in to TechShop to check it. Once I took it off the clamps I still had some minor warpage and a few alignment issues but it sat surprisingly flat.

I took out the quarter inch mounting plate and sat it on top of the skirt and, except for the opened joint and some squeeze out (which I’ll deal with later), it all looked better than I thought it would.



I’ll clean up the squeeze out and sand down the scratches before powder coating but how to deal with the failed clamp? Well, there’s a product called LabMetal. It’s a little expensive ($35 after shipping) but basically acts as a moldable, powder-coatable filler paste. You goop it into place exactly like bondo and it fills gaps and adds some structural integrity. Then you can sand and grind the dried paste until you get a good surface for coating (again, exactly like bondo). My hope is that this will not only fill the crack but also add even more stability to the joints and connection points.

My only concern (and it’s pretty minor), is that so far every single opportunity for failure on the skirt has happened, and almost every solution has either failed or added additional issues that I’ve had to deal with. It’s been a constant matter of adjustment and readjustment and fixing problems that the last solution caused. For example, once I put the 1/4″ plate on the newly JB-welded skirt pieces I suddenly realized that I have no way to attach them together except to bolt them to the undercarriage of the frame. The problem is that I want to fill the gap between the two parts and smooth it out so that you can’t see the joint. BUT this means that the parts have to be connected. Before they get attached to the frame.

So. Extra step. I have to drill and tap more holes (probably six or eight) so that I can attach the two plates together. Once I do that and the entire thing is monolithic I’ll fill the gaps with Lab Metal and smooth everything down so that it looks nice.

If this doesn’t work I’ve decided that I’ll go back to square one and recut the parts and rebuild the whole thing. I have the waterjet files and know how to use the slip roller. All-in it’ll only cost me around $30. As a friend of mine at TechShop said, it’s not a failure, it’s just a first draft. I sure hope it works, though. I still have to make and attach the 12 vertical trim pieces but I’ll cross that bridge when I come to it. Or I’ll go back and build a new bridge.

June 19, 2015

Skirt Work

Filed under: The R2 Project — jasony @ 9:57 pm

More work on the skirt. I thought this would take me a few days but it’s turned into over 3 weeks of design, redesign, cut, recut, mistakes, and remakes. I was pretty frustrated tonight when I tried to use the Alumaloy I got in the mail. That stuff stinks. After applying it (correctly, I might add), I put it under just a little bit of pressure and it snapped. Definitely a no-go. Plus, it made a globby mess on the outside of the skirt that I had to sand with a flap sander to remove. Yuck. I did manage (after creating a big mess) to fill the hole created by the welding mistake from last week, though, so there’s that. Still, the Alumaloy was a $17 dead end. Oh well.


So I sat and stewed and aaaalmost decided to go off-plan and just rivet the whole skirt together. But at the last second I decided that I didn’t want tiny 1/8″ rivets showing in a place that nobody will every look. This has got to be perfect and if it’s not, I darn well want it to be because of some unavoidable issue, not just laziness.

So I dumped everything on a table at TechShop and sat an looked at it for a while. And looked, and looked. And finally decided to give my bracket-and-JB Weld plan a try. It was a big goopy mess but eventually I got everything bonded together.


The brackets are inside the skirt and invisible from the outside so I really don’t care what they look like as long as they’re strong! Hopefully this’ll do the trick.

I had to apologize to the employees there as I managed to use nearly every small clamp in the entire shop. But it’s drying now and should be cured in 24 hours. I bought another couple tubes of JB Weld and will run some fillets along the inside joints for strengh once the brackets dry. Then I’ll probably degass the JB in the powdercoating oven at low temp for a few hours and check to make sure the JB doesn’t bubble so that the powder coating won’t have issues when it cures.

But before that I’ll have to cut, tap, and sand 12 vertical details. Ugh. I’m ready to be done with the skirt!

June 13, 2015

NI Featured Project

Filed under: The R2 Project — jasony @ 6:47 am

R2 has his own page on the National Instruments Labview Makerhub. I’ll be periodically updating this page with info on the electronics aspect of the build.

Since the electronics stage won’t happen for a while, updating may be a bit sparse, but I do plan on listing some thoughts there about the scope of the project soon.

June 11, 2015


Filed under: The R2 Project — jasony @ 3:19 pm

I visited a local anodizer today and took R2’s frame in for a quote. It was much less than I thought to have the entire frame anodized. The only issue was that I really had my heart set on a vibrant blue color (think R2’s external blue but on the frame). The very helpful Bob at the anodizing place said I could have any color as long as it was black (or clear/natural). There’s one more place in town that I’m going to check on Monday to see if they can do a blue economically. My suspicion is that it would be around quadruple the price for them to do blue. At that point I have to ask myself if it’s worth it since the frame will be covered anyway.

One of the benefits of a black frame is that the laser etching will stand out much nicer against a black background than on blue. Black will also blend into the shadows much better if any part of the frame is exposed, though I think I would like to see the blue frame anyway (kind of like seeing contrasting brake calipers on a hot hatchback at the stop light). The place that does blue may not have a big enough tank to to R2’s 18″ rings, though, so it might be a moot point.

Black is nice, though. It’s just different from what I’ve been clearly pictured in my mind for a long time so I’m trying to decide if I can make the mental switch or not. Hmmm… maybe if I went to Dallas or Houston I could get blue done. How much do I want blue vs black? Gotta think a bit.

In other news, the skirt has been giving me fits. What I thought was going to be a two or three day job has turned into a month-long hassle. I designed, water jetted, and prepped the skirt pieces over a two week period. Then I carefully slip rolled the curved end pieces, sneaking up on the proper diameter until I got the pieces just right. They’re still just slightly off but they’re close enough that they will “spring” together under slight pressure to form the skirt. Here are the parts precut in wood and cardboard to check fit:


And here are the parts cut in aluminum after I slip rolled them:



Realizing that my TIG welding skills are nonexistent, I talked to a pro aluminum welder at TechShop and he agreed to try and tack a few parts together (welding aluminum is a pain and I didn’t want to burn through the thin metal) in exchange for a few bucks. I’m going to try and do just as much of this project myself as I can but my goal is verisimilitude over self-reliance, so a (very) occasional professional is okay with me if it means the difference between perfection and something that looks like a 3 year-olds’ crayon drawing on the fridge. Aluminum welding is a skill that takes years of practice and I don’t want to go down that particular rabbit hole right now. Unfortunately, the guy who did the welding kind of blew it:



He did his best but we determined that it wasn’t his fault but rather the result of very thin material (.063″) and a mating surface that wasn’t a true butt joint (look closely and you’ll see that only the corners of the edge really ever touch). It’s SUPER easy to blow through that kind of weldment, and that’s what he did. We only did this one joint and I called it off. So much for welding the thin skirt material. So what to do? After a few hours of research I found a product called Alumaloy. It’s a low temp (“low” meaning 728 degree) brazing rod for mating aluminum and filling aluminum gaps. It doesn’t get stellar reviews (2.5/5 stars) online but I think/hope that’s because people are trying to heat up giant heat sinks-worth of aluminum. Since I have thin material I hope to be able to get it hot enough with a propane/MAPP gas torch. Right now it looks like the only solution to my problem. The Alumaloy is on its’ way via Amazon Prime so I’ll tackle it next week.

Once it’s here I’ll need some way of bracketing the thin pieces together, so I did a quick test with some broken coffee stirrers:


Aha. now we’re talking. I found some scrap thin material in my TechShop locker and cut it out on the hydraulic shear/beverly shear:



And, after much trial and error, bent the small brackets to the correct angle with the bending brake:



Voila! I now have some small custom aluminum brackets at the proper angle that I can clamp onto the parts to make a lap joint. Hopefully this will be enough to form a decent base for the Alumaloy.

One more part I milled up last week was the main mounting bracket for R2’s center foot. I opted to go all manual mill on this piece instead of water jetting. It ended up taking me 10 hours since I had to thin the aluminum hunk from 1.125″ down to .75″ with a monster facing bit I borrowed from the TechShop machining instructor (who, in a very cool bit of trivia, worked on the Saturn V rocket booster 2nd stage!):


After milling, facing, tapping 8 holes, and creating an incredible mound of aluminum shavings I got this



Yeah, I’m really proud of that sucker. Most of the other parts on R2’s frame have leaned heavily on CNC techniques, and, while I can take credit for the design and pathing of the files, having a machine do most of the cutting work and just cleaning things up and tapping doesn’t feel as complete as taking a raw hunk of aluminum and shaping it like this. That part feels beautiful in my hand.

So that’s the state of things. My goal is to have the frame anodized and laser etched and the skirt assembled and powder coated by the end of August. Sept 1st is the “official” beginning of my 2nd year. It’s not a brick-wall deadline but I’d like to stay on track. I don’t think it’ll be a problem.

I’m also researching LabVIEW and my new myRIO that I was given (!) by National Instruments. They got wind of the project and wanted to be involved. They also want me to come update their engineers periodically once I enter the electronics/programming phase. Man, I’m excited about this, but a little nervous about the physical wiring and virtual programming/logic. With plans to pack R2 with sensors and make him a semi-autonomous robot I just hope my reach isn’t going to outstrip my grasp. I’m realizing that I’ve set myself kind of a ridiculous bar here. Can I meet it?

June 9, 2015

NI Contact

Filed under: The R2 Project — jasony @ 10:18 am

Just got my myRIO the other day and I’ve installed LabVIEW and played with it a bit. I got a wonderful support call from National Instruments just now. The rep asked if I had any questions about myRIO or LabVIEW and wanted to know what my project was. She offered to get me in touch with any resources that I might need to help me learn how to use their stuff.

See, this is how a great company handles consumers. I’m not only really excited about using the myRIO and LabVIEW in R2, but it feels like I have a good support group behind me if I have any questions. She was super excited about the project when I told her what I was making. I’m really looking forward to rolling R2 into the National Instruments headquarters someday.

June 5, 2015

There are Many Rio’s, but This One is Mine.

Filed under: The R2 Project — jasony @ 9:33 pm

Tonight I received my MyRio box from National Instruments. I had a fantastic lunch with the N.I. guy where we completely geeked out over what the MyRio box can do in the R2. Initially it was just going to be for sensor data acquisition but after talking to the rep it’s looking like the little box is going to be the main central brain.

This solves many problems: centralization, programming simplicity, power, and connection minimization. Plus, LabView looks like it’ll be just the thing for sensor/servo/motor control. It also contains some rather intriguing features regarding vision analysis, data streaming over the web, wifi/bluetooth integration, and a few other goodies.

Here’s the cool part: National Instruments got wind of my R2 project and was so excited about it that they wanted to supply the MyRio free of charge (street prics is almost $1,000)! They also want me to come to N.I. periodically and participate in their group lunches to update them on the progress. This little R2 is going places!


June 4, 2015

My MyRio

Filed under: The R2 Project — jasony @ 10:21 am

Just got an email from my contact at National Instruments. The higher-ups are impressed enough with my R2 project that they want to support it with a MyRio. Completely! Yes, they’re supplying one for free for the project!

So I guess R2 is now “sponsored” by National Instruments? That. Is. So. Cool. The only thing that my contact asked is that I write some posts about how I’m integrating it into the project. But yes, as of now (well, tomorrow when I pick it up), R2 will have a brain courtesy National Instruments. I’ll of course post info here but I’ll also be posting to Instructables and other such sites. Yeah, it’s a quid pro quo, and I’ve never really done sponsored stuff (does this mean I have to write “full disclosure” on everything?). But in all honesty I really do think the MyRio is a great solution for what I want to do. Yes, there are other boxes out there that’ll do the same thing, but at my level of Maker, learning to do the programming with a BeagleBone or similar box was going to be really limiting since I don’t want to become a programmer. The LabView/MyRio combo is perfect. Just enough graphics to make it easy and plenty of power (FPGA! FPGA!) to do some really cool stuff.

I’m still sticking to my modified 4 year plan but now I really want to start working on the electronics stuff. Can’t wait! 🙂

June 2, 2015

R2 Skirt and Electronics News

Filed under: The R2 Project — jasony @ 10:39 pm

Huge conversation at TechShop tonight. I had one of the guys from National Instruments in my class (both guys in this photo, by the way) and we got to talking (for about 90 minutes after class!) about the R2 project. I mentioned that eventually I wanted to get some basic sensors (temp, humidity, CO2, Methane, etc) integrated into him and wondered if the new MyRio box from NI would do the trick.

What followed was an hour and a half whirlwind conversation that I won’t even try to recount except to say yes, yes it will. Nicely. This is an amazing little dual core Linux box that comes with Lab View and will allow me to not only plug in some truly wonderful and hard core National Instruments/NASA level sensors, but will also do some amazingly creative and complex control of motors, servos, NEST thermostats, GPS, etc. It interfaces via WiFi and can run completely independent of a host compuer. It’ll do Kinect style motion following, record audio and video onboard (or stream to a server via WIFI), and (here’s the part that’ll make hardcore Geeks swoon) instantiate your logic as an FPGA with a single click. Great Babbage’s Ghost.

The box is normally $1,000, which is pricey, but the engineer was so excited he’s going to try and get me the MyRio and LabView (the programming environment) for the educator’s discount of $250. Seeing as the control boards for R2 alone were going to be around that cost, and were much less flexible, I count this as a complete and total win. My brain is buzzing over what I can do with this little beastie. Stay tuned.

In other news, I got the bread pans powder coated. They’re not super perfect but I think I’ll go with them for now. Barring much better equipment (which is why they were a little messed up), stripping and re-coating them wouldn’t be a guarantee of better results. So I’ve turned my energy to R2’s skirt. After a week or so of thinking and design work in Inventor I cut the parts as a test on the laser in plywood. Then I cut them tonight on the water jet. A couple of minor issues have made it necessary for me to do another piece on the WJ tomorrow (very small and minor). Then I’ll be able to weld the aluminum skirt together and start in on the skirt details. When that’s done I’ll powder coat the skirt (or, actually, ask one of the TechShop members who has a really nice powder coating system to do it for me). Then I’ll start the research into prepping the frame for anodizing. After that I’ll start designing the laser files to etch the artwork/circuitry/Death Star plans into the anodized frame. I’m still shooting for Sept 1st for my Year 1 completion. I may finish a bit early at this rate.

May 14, 2015

Tapped Out

Filed under: The R2 Project — jasony @ 9:08 pm

So here we are, nine days after my long post about finishing the basic frame and starting the 151 tapped holes. Tonight I completed all of the drilling and tapping! What a slog.

Since writing that post I’ve spent maybe 70 hours at either the lathe or the manual mill. I’ve tapped and drilled and center-finded and changed bits and adjusted X/Y zeros until I can do it in my sleep. In fact, I probably have done it in my sleep. The shortest day at the mill was 6 hours. The longest was when I got to TechShop around 6am (couldn’t sleep) and didn’t leave until almost 10.

But hey! The basic frame is DONE. And when I finished all the milling and threading I assembled it, screwed it together, and took a minute to step back and admire the work. I made this. Sure, you can buy it, but this one is all mine. I’m profoundly proud of what I’ve done so far. Also kind of surprised.




I am taking tomorrow off and not going to TechShop. The employees have (no kidding) started to joke that I should just clock in like they do since I’m there so much. One time this past week I got there about an hour before the night shift left, worked through both day shifts, and left just as the night shift came on again. Crazy! The funny thing was that the entire day was spent in a flow state that made time pass really fast. It’s so much fun to learn new things.

Next up: I’ve decided that instead of working on the dome and skins (which are ready and waiting), I’m going to pull the year 3 schedule into year 2. So once I finish up year 1 I’ll start in on the legs. Some good reasons for this that I won’t go into here, but mainly it’s so R2 can have some legs to stand on and not just be a sad little aluminum trash can shape. Before that, though, I need to figure out the “skirt” (the sloped bottom part of the body) as well as decide how to go about making the “bread pan” little cubbyholes in his frame. Once those are done I’ll polish and anodize the basic frame and start designing the pattern that I will laser cut into the anodizing. Hopefully all of this will be done at the one year mark of Sept 1st. If not then hey, it’s all for fun, right? 🙂

May 9, 2015

Ring 1, Ring 3!

Filed under: The R2 Project — jasony @ 8:55 pm

Today I got a huge chunk of the frame done. I milled and tapped holes in the “A” and “C” rings (JAG01 and JAG03). It took about six hours to index everything and slowly tap about 30 holes. I have to admit, the holes are much easier to tap than I’d thought and I haven’t broken a tap yet (knock on wood). Erin showed up at TechShop with dinner and I let her set up and tap a couple of #4-40 holes herself. She did a great job and didn’t seem that phased by it.

I also met a really nice guy working on the other mill by the name of Josh. He took quite a bit of time to watch my milling technique and give me some very much needed pointers. He even got me out a jam when I couldn’t figure out how to proceed.

I really feel like my machining skills are improving by leaps and bounds. It might have to do with the fact that I spend 14 hours in the shop today, 12 yesterday, 10 the day before, and 14 the day before that. 50 hours in four days! Malcolm Gladwell would be proud. I’ve been having a lot of trouble sleeping lately so I’ve been getting up really early and driving to TechShop to work. I never thought I’d say this but I’m glad they’re open 24 hours.

In related news: I think I need some new shoes, or at least new insoles. 🙂

I went through the blueprint book tonight and “signed off” about a dozen pages. Those parts are done! Not only did that feel absolutely great, but just looking at the parts I’ve completed gave me a big confidence boost that I can actually tackle some of the upcoming parts. The more experience and confidence I gain in the shop, the better I am able to look at a part and “see” in my mind how to go about machining it. I’m asking a boatload of questions of the other people at TechShop and everyone seems pretty excited about the project. The employees give regular tours every hour or so to prospective members and whenever they come through the machine shop they always say “And this is Jason. He’s building an R2D2”. No matter the age of the people on the tour, everybody breaks into a smile and takes a minute to watch me tap a hole, cut a part, or page through the blueprints. This project really brings people together.

Sorry for the lack of pictures in these posts. I’ll try to do better. One of the reasons is that there are a ton of tiny aluminum shards flying everywhere in the machine shop and it’s very easy to get a little bit of aluminum on the iPhone screen. I made a tiny scratch on the screen tonight trying to clean it off. So I keep the phone hidden away and use my new bluetooth ear protectors (these babies) to listen to podcasts and talk on the phone. That way the phone (and its camera) stays hidden away in my toolbox.

May 7, 2015


Filed under: The R2 Project — jasony @ 9:02 pm

Back at it today milling up the other 8 vertical support posts. This time I knew exactly the sequence of cuts and moves to cut each part. I even optimized my movements around the lathe to speed things up (you get a lot of time to think of this sort of thing when you’re doing the same procedure over and over and over again).

Yesterday I took around 7 hours to cut 8 parts. Today? 4.5 hours. Boom. I finished and though “gee, I’m glad that’s over with”…. and then discovered that I still have 4 more left.

Ah well, at least I’m good at this!

I took the parts I’d done and did a little preliminary screwing together of the frame. It looks great! I’ll do the other 4 and then snap a picture.

Next up (after the 4 remaining rods) is to build a jig to place on the manual mill so that I can locate, drill, and tap each of the holes in the horizontal “rings”. Lots and lots of those little holes. I also have to countersink, which should be comparatively simple.

I spent a few minutes looking through the blueprints for (much) later steps. Legs, skirt, ankles, etc. Then I abruptly shut the pages and returned to what I’m focusing on. Way too many things I don’t know lurking out there in the future. Best figure one thing at a time out and not freak myself out.

May 6, 2015

To Everything, Turn, Turn, Turn

Filed under: The R2 Project — jasony @ 8:54 pm

Whew, what a day. I went to TechShop this morning at 11:30 and stood at the metal lathe for almost 8 hours. In that time I managed to complete 8 of the 16 vertical support rods from the R2 frame. Each rod has a tapped hole at both ends, so 16 holes.

It took about an hour to get everything adjusted correctly, find all my tools, get advice, and carefully test out a piece. Once I felt good I just stood there for hour after hour carefully working on each piece. Here are the steps:

1. 45 degree chamfer and detail lines .33″ apart
2. center drill end of rod with a center bit (basically a short stubby bit that makes a pilot hole for the main drill bit
3. Drill a 1″ deep hole with a #7 tapping bit
4. Tap threads with tap (by hand…. no power)

It doesn’t look like much but the first hole took over 30 minutes. I got to where I was doing one rod (twice through these steps) about every 40 minutes. I forgot to eat breakfast before I went and skipped lunch (except for an apple) so by the time I finished at 7pm tonight I was starving. Luckily, Erin has a nice steak cooking when I got home.

Just before leaving I took one of the rods and hit it with a high rpm grinding wheel (a soft wheel) and removed the oxidation. I then did a polishing pass with a polishing wheel just to see what it looked like. Verdict? Amazement. It looks so much better than I thought. What was a simple oxidized rod of aluminum now actually looks like a for-real metal part to something. This is going to be pretty spectacular, I think.

I’ll probably go back up to TechShop tomorrow but I don’t know if my feet can handle another 7 hours at the lathe. I’m halfway through with the 16 rods but I think I need a break.

May 5, 2015

JAG 04 and a Milestone!

Filed under: The R2 Project — jasony @ 6:17 pm

Big day today! But first:

Last night I cut the final part of the main body on the Waterjet. It was by far the longest cut I’ve had to do (though not the most complex). Gratefully, the WJ performed perfectly with no garnet clogs or pressure drops. I did two 22:00 cuts for a total of 44 minutes of cutting. I cut two identical parts.

Here’s the FlowCut software with my part all laid out and pathed, ready to cut:


The solid lines are cutlines, the dashes are traverses (where the cutting tube moves but shuts off the water so it’s not making a cut). The colors denote what percentage of speed the machine uses. In short, if you tell the machine you’re using .5″ 6061 aluminum (which I was), it calculates the fastest possible cut speed to get through the material. You won’t get a neat or clean cut (you’ll have pretty jagged edges), but it will go all the way through the material. In order to get a neater/smoother cut you tell the machine ahead of time that you want it to go at only 80% of optimal speed, or 60%, or whatever. Then as it’s cutting you can dynamically speed up or slow down the cutter as it cuts out critical surfaces. You can see in the above photo that the top and bottom edges are dark blue (the 40% speed). This is because those surfaces will be drilled and tapped and a flat and neat cut is critical. The light blue lines are less critical and the purple lines are only there to cut out chunks of aluminum for weight reduction. Make sense?

Anyway, once I was happy with how the layout looked I did a dry run and the machine ran the 22 minute cycle with the pump off (so basically free since you only get charged for cut time). After that I clamped down the part and did the cut.

But first! The previous day I’d gotten to this part and realized that the narrow piece of aluminum I’d purchased was only just barely wide enough for the part with very little room left over to lay down the weights that I normally use to secure the aluminum blank. If I’d have executed the cut I’d have risked running the cutting tube into a weight, breaking it, and have to pay TechShop $250 for a replacement. So what to do? Why, spend 6 hours on the manual mill machining up a set of custom clamps, of course! Four pairs of brackets (5052 aluminum) with a threaded rod connecting them. The bottom bracket has threaded holes that the rod screws into while the top bracket has oversized holes so they can slide up and down the threaded rod and be captured and tightened down by nuts and washers. Today I cut a nice little lasered box for the whole system. I’m even going to powdercoat the metal pieces red for visibility (well, really just because I can). It’s a good system. Here’s the narrow piece clamped down on the water jet table:


The aluminum brackets hold the piece very securely and the blast skirt that surrounds the mixing tube (where the water comes out) easily clears them. The tube would probably still break off if it hit the brackets but this gives me a lot more room to maneuver.

So once this was all done I started the pump, hit play, and watched the machine do its thing.

For the first few seconds after the stream starts (eight or so) you can tell that the water/garnet mix is cutting its way through the .5″ of aluminum. All of that water has to go somewhere so it bounces back off the part, gets turned to (very warm) steam, and sprays everywhere. This is not a neat machine to operate. Wear grubbies. My arms got a light sandblasting from the overspray from each of the pierces. Once the stream is through the material, though, it’s smooth going. All the water exits out the bottom of the part and gets absorbed by the water in the tank.

Now that the critical JAG04 parts (shoulder plates) were done, I hit my milestone: all basic frame parts are now done! Today I went in to TechShop and precariously balanced/clamped the parts together into the frame to check fit. Looks like everything fit fine though I won’t know for sure until the holes are tapped and bolts/screws installed.




There are some curved pieces on the front that you can’t see since they’re held up by holes and screws that aren’t there yet. Speaking of holes and screws:

NEXT STEP: Now that the basic frame parts are done I’m entering once again into uncharted territory. I have to drill out (enlarge) the holes that the waterjet cut by just a little bit. I’ll clamp each piece onto the mill, locate the exact center of the hole with a “wiggler” tool I just bought, and then swap out the wiggler for the correct tapping bit. I’ll then enlarge the hole and swap out the drill bit for a spring loaded tap guide which I’ll use to keep the tap handle perfectly straight. Then I’ll tap and clean the hole so that it has nice neat threads. Lather, rinse, repeat.

One hundred and fifty one times.

Yep, most of those holes you can see on the frame need to be slightly enlarged and have threads cut into them. I also need to drill and thread the ends of each of the vertical 3/4″ aluminum rods (I’ll do that on the lathe with basically the same procedure as described above). One hundred and fifty one holes. I’ve tapped exactly five holes in my entire life. I’m pretty nervous about this. I figure I’ll be doing this at a rate of ten or so holes per day. I’ll doubtless start slow as I figure out how to find the centers of the current holes and get a feel for tapping (the last thing I want to do is break off a tap in my beautiful frame pieces) as well as get into a groove on procedures and muscle memory. I figure it’ll take me three or four weeks of work to finish this step. Boooring but important.

You know, it is possible to actually buy a frame online for about the same amount of money as I’m spending do to this myself, but it’s really rewarding and fun to learn so many new skills. I really feel like I’m stretching what I can do and adding some skills to my skill set.

By the way, I discovered something: if you assemble an R2D2 frame at TechShop don’t expect to get any work done. Little R2 was the star of the work area (called the “Hub” at TechShop) this afternoon. I probably had a dozen people come up and ask me what it was and then start offering all sorts of suggestions of what I should do. So for the record:

1. No, there won’t be a holoprojector (yes, you’re very clever for saying this… you’re the hundredth person to do so)
2. No, there won’t be a short guy inside (ditto)
3. No, it won’t be completely autonomous (though I’ll try and make it at least as smart as a Roomba) (and ditto again)


April 24, 2015

JAG 20 and Problems

Filed under: The R2 Project — jasony @ 9:39 pm

Well, I just can’t stop myself. Today at TechShop after a class I checked out the key to the FlowJet (I just love saying that so casually: “I checked out the key to the quarter million dollar FlowJet”) and pathed the JAG20 file in FlowCut. I figured I’d take the file back to the WJ to see how long the cut would take. I’ve done this before and know how it usually ends.

I loaded the file and did a virtual cut. 30 minutes for both pieces. There are 2 JAG20 parts that look like this:


The part is the body-to-leg flange joint that connects the main frame to the leg assembly. It’s solid 6061 aluminum and 1 1/8″ thick. You can see it located here between the legs and the frame. It’s visible once Artoo is finished so it has to look right:


I bought a slab of aluminum from Metals4U a few days ago in preparation of this part. This is probably the biggest, most nerve-wracking part I’ll do. Partially because it took me weeks to figure out how to design in Inventor (I finally got help from an expert at TechShop… he solved the problem in 5 minutes), and partly because it is, by far, the thickest part I’ll have to WaterJet in the entire project. The single hunk of aluminum measured 1 1/4″ thick, 11″ wide, and 16″ long. Cost me over $80! So no, I did not want to screw this up.

So I put the hunk of aluminum on the FlowJet table and zeroed out the Z axis to get the mixing tube right down on the metal. At this point you don’t want to smack the tube into any clamps or weights holding the stock down. If you do, you’ll break the tube and owe TechShop $250 for a replacement. Fortunately, I know the guy who teaches WJ (I am the guy who teaches W.J.) and we have procedures to avoid that.

So, the file was loaded, the stock was weighted down, and I even did a dry run without the pump running (the part that costs money). It was a 30 minute cut. Well, why not? I verified that the machine had been working well and went for it.

Part on table weighted down. File loaded. All ready to go!


I hit ‘cut’ and watched with an eagle eye for the next 30 minutes as the water and abrasive jet slooooowly worked its way around the part. 4 holes, inside cut, outside cut, repeat.

The first part went off beautifully and I could tell that it cut through the material just fine once it was finished. The machine started in on the second part next. The 4 holes went fine (they’re first), but sometime during cutting out the center “shamrock” shape the pump started to lose power. Suddenly the cutting stream wasn’t making it all the way through all 1.25 inches of aluminum. You can tell this is happening because instead of the cutting stream passing through the material, the water starts spraying all. over. the. place. All over you. I got blasted with a muddy stream of grungy water (seriously, I looked like a motocross rider after a muddy ride). The normal procedure if this happens is to manually slow the cutting stream down. As I was already cutting at a pretty slow pace (which gives a better finish) I wasn’t too excited about slowing down further and making the cut even longer/more expensive. But it’s better than the cut being ruined.

So I slowed it down (which you can do mid-cut with a slider in the software) and managed to get the center cut done. On the last cut (the outside perimeter), the jet was still having issues. The solution was to slow it down even further. I think by the end I was cutting at somewhere less than 50% of the max speed and still unsure if it was making it all the way through.


Here you can see the layers of garnet that get blasted onto the parts during the cutting process. It’s an incredibly messy machine to operate.

Once the machine shut off I checked the parts and discovered that the first part to be cut (the one on the left in the photo above) came out brilliantly. Great finish, cut all the way through. Unfortunately, when I lifted out the part on the upper right I discovered that there were sections of the middle that didn’t get cut all the way through. Fortunately, though, it did lift all the way out so at least it got free. Unfortunately, the finish on the outside was terrible as a result of the gradual pressure loss.

Here’s a pic of the “good” part. You can see the smooth edges with just a hint of the typical jet “lag” on the bottom of the outside curve:



For the “bad” part, you can see not only where the stream didn’t cut all the way through, but also where the water jet “lag” on the outside is much more pronounced, resulting in a much more ragged and bad looking surface:



And here’s a side-by-side of the good part and the messed up one (zoom your screen or click for a closer image):


So, while I’m really happy to get these parts behind me, I am somewhat disappointed in the results, considering that for about the same amount of money I could buy an aluminum part that’s shiny and perfect.

I’m still cogitating on how I’m going to fix this. The good news is that the final piece needs to be 1.125″ thick (1 1/8″). The rough piece in the pictures is 1.25″ thick. So I should be able to reduce the thickness of the piece and “face off” the thin little bridges of aluminum that keep the inside section attached. This will also have the effect of making the faces of the part nice and shiny. I think I can do this on the lathe at TechShop but I’ll have to ask the pros about the best way to go about this.

As far as the outside of the donut shape, I need to see if the part can be a smaller diameter than the 6.375 listed in the drawings. I think it can. If I can safely shave off a few thousands I think I can get rid of those marks on the outside and make the part shiny.

Sorry to be so wordy. This experience was a relief in that the part is somewhat workable but frustrating in that I just don’t feel like I can completely trust that machine. I need to improve my skills on the manual mill and lathe but that takes decades. The WaterJet, while expensive, can let a newbie machinist cut any part they can design in the computer. That’s huge.

The final major part for the frame is the pair of side panels where JAG20 mounts. They’re .5″ thick aluminum (thank goodness) but also contain a lot of little holes that’ll make the water jet turn on and off quite frequently. I may wait a few weeks until the pump is rebuilt to attempt those. They’re only 20 or so minutes each but I really don’t want to mess those up. At least not the small little precision holes.

April 20, 2015

JAG 01!!

Filed under: The R2 Project — jasony @ 10:08 pm

Exciting news on the R2 front. JAG 01 is cut! JAG01 is the part number for the topmost ring in R2’s aluminum body. It’s by far the most complex part of the main body (along with the leg support panels).


Each of those holes (accurate to within +/-.005) mean that the FlowJet has to cycle on/off each time. In my experience teaching the tool, there is most likely to be a problem during a start/stop cycle. Yikes!

I’ve been putting off JAG01 for a few months now since it’s such a tough part and I didn’t want anything to go wrong. I laser cut the file in thin plywood using my Inventor .dxf file and it seemed right, but there’s nothing like actually water jetting a part to show you your errors. The FlowJet has been down for a few weeks at TechShop and they finally got it working again. I happened to be there when they did and the employee said they needed a test file. Would I like to run JAG01 in some scrap (too-thin) aluminum? YOU BET!

So we took about 12 minutes to run the file in some 18ga aluminum and I took the part and laid it out against my plywood version. Perfect fit! So I know the file worked fine. Now all I had to do was clamp the 1/4″ slab of 6061-T6 aluminum onto the table, take a deep breath, and go for it. I did that tonight and kept an eagle eye on the tool as it cut (not that I could salvage the part if anything went wrong). Fortunately, it worked! 18 minutes later I had a beautiful part resting on the FlowJet water surface and I could breathe again.

Below is a picture of all the pieces I have milled up in the past 6 months when I started this crazy project. Each part is accurate to within a few thousands of an inch. JAG01 is in the upper right.


And here’s the frame so far balanced precariously together.


Year 1 is the frame and I’m more or less on track to have it done in 4-5 months. I have to do two copies of JAG04 (the shoulder mounting plate), and then two copies of JAG20 (the shoulder flange). For JAG04 I’ll use the WaterJet again in combination with the manual mill. Then I need to go back in to a bunch of the holes that were WaterJet and either tap them with threads, or countersink a chamfer on the edge so the screws sit flush. For JAG20 I’ll be learning a whole new tool: the Tormek CNC Mill. That’ll take a month or two. I could buy the part online for about what it’s going to cost me to make it, but this project is about testing my boundaries and learning to do new things. I will have to buy some parts for R2 (the skin and dome), but I’m trying to do as much as I possibly can myself, and use the most difficult material (aluminum as opposed to wood). If I hose JAG20 I’ll consider buying the part, but I want to try first.

Once all the parts are done, drilled, tapped, countersunk, and fit together correctly, I’ll sandblast everything to remove milling scratches, then sandblast again using baking soda. Then it’s off to the anodizing step to give the frame a beautiful coat of anodized metal. Then I’ll laser etch a pattern into the frame.

There’s a long, long way to go.

« Newer PostsOlder Posts »

Powered by WordPress