<div style="width:35%; float:right"> <video width="100%" autoplay loop controls muted><source src="https://i.imgur.com/L2L6pLc.mp4" type="video/mp4"/>Your browser does not support the video tag.</video> <legend>"Joystick control on radius stepper."</legend> </div> <h1>Project Development III</h1> <h4>Wednesday 31 May, 2018</h4> ## T-minus 14 Days So apparently there's 1 209 600 000 milliseconds left to complete this final project. That might seem like a lot, to an ATtiny, but to an old-school carbon-based life-form like myself, who spends about 30% of it's time either sleeping or eating, that's not going to leave a lot of room for 'the extra-curricular'. Infrequently asked questions (pt. 2): + <a href="assignment18.html">What tasks have been completed, and what tasks remain?</a> + <a href="project04.html">What has worked? what hasn’t?</a> + <a href="assignment18.html">What questions need to be resolved?</a> + <a href="assignment18.html">What will happen when?</a> + <a href="project04.html">What have you learned?</a> This page outlines some of the interesting things I came across while developing the project. If you're looking for project specific information you can see my <a href="assignment18.html">planning write-up</a>, and <a href="project04.html">the project as completed</a>. ###Dummy-fitting the assembly One of my favourite advantages of sleeping next to a 3D printer is the ability to pump out prototypes quickly to check fitment and interfacing with other components. Coupled with my experience in <a href="assignment06.html">printing insanely fast</a> I can have a fully mocked up assembly to base the next iteration of measurements from in only a few hours. With an ambitious project such as this one, it really helps to build and re-design fast. In my experience it may take me an average of three attempts to print something correctly, and in between the first print and the final print I may re-design a part up to 30 times to get it to interface nicely! (Don't worry, every time you re-do something it becomes 10-times faster 😉 ) Especially in order to make sure the enclosure would fit well and be comfortable to hold, I made a few test-pieces and then begun printing brackets to position the motors and lead screws. I also did this to get a sense of how much force the friction on the guide-rails would exert on the linear assemblies moving back and forth. After many iterations I discovered a method that involved putting the guide rails in the chuck of a drill and sanding up and down the length with 320 grit emery paper. <div style="width:100%; float:left;"> <img src="media/week12/testing-1.jpg" style="width:33%; float:left; padding:5px 5px 5px 5px;"/> <img src="media/week18/test-fit-rails.jpg" style="width:33%; float:left; padding:5px 5px 5px 5px;"/> <img src="media/week18/test-fit-rocket-surgeon-assembly.jpg" style="width:33%; float:left; padding:5px 5px 5px 5px;"/> <img src="media/week18/hammerhead-rocket-surgeon-2.jpg" style="width:25%; float:left; padding:5px 5px 5px 5px;"/> <img src="media/week18/rocket-surgeon-test-print.jpg" style="width:25%; float:left; padding:5px 5px 5px 5px;"/> <img src="media/week18/test-fit-hammerhead.jpg" style="width:25%; float:left; padding:5px 5px 5px 5px;"/> <img src="media/week18/hammerhead-rocket-surgeon.jpg" style="width:25%; float:left; padding:5px 5px 5px 5px;"/> </div> <legend>"The various fit tests of the assembly. It took me a few attempts at the theta bearing, and the main thing I had problems with was accessing bolts where multiple printed components were stacked up. The other main thing I was checking for was the stiffness and weight distribution of the assembly, because one of my main goals was that it needed to feel <strong>usable</strong>."</legend> <div class="row"> </div> ###Gearbox I decided to use a gearbox because by my calculations of using a 1.8deg stepper motor at an extension of 50mm I would have a resolution of well outside useful specification. I was also quite worried about the holding torque of the theta stepper, and so having a gearbox would contribute greatly to solving that issue. The main drawback was backlash, but I decided that wasn't going to be an issue I solve for the assignment. If you're planning to try this at home (I won't tell you not to because then you will even more..) then I would strongly recommend going for a machined gearbox. (How you do that is up to you!) I was lucky to find a gearbox already designed for printing and designed for the NEMA17 faceplate released by a guy named Kevin who now works for Microsoft. I made use of Kevin's STL's but simplified them and re-worked them in Fusion to allow more interface options on the planet-carrier. This brings me to my next issue - how do you edit *.STL's in Fusion 360? Well first of all, certainly don't do it on a MacBook. Secondly, I found a source of awesome up-to-date walkthroughs by Scott Hatfield a.k.a toglefritz. The two I found particularly useful were <a href="http://toglefritz.com/convert-a-simple-stl-into-a-body-in-fusion-360/">How to convert and simplify STLs in Fusion</a>, and <a href=" http://toglefritz.com/convert-a-complex-stl-into-a-body-in-fusion-360/">How to do that same thing but for more complex STLs</a>. I honestly just can't believe his instructions worked. Fusion gives absolutely no indication about how to make use of these tools, or even what a BRep is. Toglefritz has also got some really cool other blogs that he's worked on that show you how to do other stuff good to. <div style="width:40%; float:left;"> <img src="media/week18/machined-circuit-board-fusion-360.png" style="width:50%; float:left; padding:5px 5px 5px 5px;"/> <img src="media/week18/machined-slip-rings1.png" style="width:50%; float:left; padding:5px 5px 5px 5px;"/> <img src="media/week18/circuit-board-in-fusion-360.png" style="width:100%; float:left; padding:5px 5px 5px 5px;"/> <img src="media/week18/slip-rings-3.png" style="width:50%; float:left; padding:5px 5px 5px 5px;"/> <img src="media/week18/slip-rings2.png" style="width:50%; float:left; padding:5px 5px 5px 5px;"/> <img src="media/week18/slip-rings4.png" style="width:100%; float:left; padding:5px 5px 5px 5px;"/> <img src="media/week18/triple-bearing-update.png" style="width:50%; float:left; padding:5px 5px 5px 5px;"/> <img src="media/week18/triple-bearing1.png" style="width:50%; float:left; padding:5px 5px 5px 5px;"/> </div> ###Trying to design a slip-ring Back when I had more time, Neil had mentioned in a lecture that it was possible to fabricate slip rings in a Fab Lab. I figured it was possible and would be fun to explore, and would be a great feature to my project if the theta head was able to spin around continuously. So I went about a couple ways to do the rings. I worked out I needed 13 connections to the rotating hammerhead, of which later I realised as a bare minimum I could have made do with 7. There were the 3 for the brushless spindle, four for the radius stepper motor, two limit switches for radius, a homing switch for Z and theta each, a line for the laser I had planned to mount for centering the head on the workpiece, and a ground for all the signals. The first design (left) was a milled pcb where the brush would contact the outside. The limiting factor here was the spacing of the three columns that make up the shaft of the hammerhead, because the ring contact would have to be on the outside of them. This left me space for only two traces so I would need 7 of these boards stacked up. It could have worked, and I had designed a board stack where each one was identically machined but then I would be able to determine which board would perform which function by the components and soldering configuration. For example, the capacitor footprint was the same as a 0R footprint (1206) so I would be able to re-wire the board to do something different depending on it's function. I tried doing these in both KiCad AND Eagle initially, but found it much too difficult to do precise or circular operations. So I found this weird and whacked out way to "approximate" circuit design in Fusion360 using the ***web*** feature. Basically to model the traces I simply drew the lines I wanted, then "webbed" them to the thickness I wanted. It's very strange that I found this a really intuitive way to route traces. It was super fast, and super easy to make adjustments to. Most of the pads pictured were for capacitors, because I absolutely love them (ask Birita) and I wanted to see if they helped in smoothing noise from the sliding brush contacts. From this I would simply project the raised surface as a sketch and export it as *.dxf then post-process it in inkscape to convert it to *.png for machining. I bailed on this idea eventually as it became too difficult for me to understand how to assemble it - soldering each one layer by layer seemed risky and time consuming, and there would be no recourse should I need to fix anything on a layer that wasn't on the top or bottom. Next thing I tried was a radially applied brush contact, where I would wind copper strand wire around a printed core and have the cables run through the middle (second set of pictures). This became too much work and too large, and I decided there were many other design problems to solve given that the backup plan was to concede to a limited arc of motion on the hammerhead. ###PCB models from Eagle in Fusion Lastly as I got nearer to the deadline, I needed to interface the "mother-board" I had designed with the enclosure somehow. I saw that Birita had <a href="http://fab.academany.org/2018/labs/falabvestmannaeyjar/students/birita-idali/projects/project_development/index.html">done a sweeeet-ass job</a> of integrating her circuits with the 3D printed base of her turbine, so I asked her to show me how to import the boards I had designed from Eagle into Fusion. It was really simple, you basically just click the button on the left that's there all the time but you choose to ignore, and then you just sign into your Fusion account and it saves it as a Fusion part. You can then go and add it to any assemblies. It doesn't work so great yet with the Fab library, since the 3D models for most inventory components isn't included yet. But the different coloured rectangles were pretty representative of what I needed to know about regarding the enclosure design. You can see below how I added the circuit and designed some laser-cut parts to fit around the physical board. One thing that I didn't take into account however, was how far the header cables came out of the board. They didn't have enough space and so I had to print a spacer that pushed the board even further out from the body as I had hoped. <div style="width:100%; float:left;"> <img src="media/week19/wireframe-product-side-view.jpg" style="width:100%; float:left; padding:5px 5px 5px 5px;"/> </div> <div style="width:100%; float:left;"> <img src="media/week19/eagle-fusion-circuit-cad-integration.jpg" style="width:50%; float:left; padding:5px 5px 5px 5px;"/> <img src="media/week19/product-cad-integration.jpg" style="width:50%; float:left; padding:5px 5px 5px 5px;"/> <img src="media/week19/product-mockup.jpg" style="width:50%; float:left; padding:5px 5px 5px 5px;"/> <img src="media/week19/project-development.jpg" style="width:50%; float:left; padding:5px 5px 5px 5px;"/> </div>