Isn’t it sad that the last meaningful post on this site was in February? I think it’s a travesty. A combination of perfect storm factors has overwhelmed even my blogging habits. I’m kind of like the Waffle House test of blogging – if even I stopped blogging, you know some shit went down. And I do have some very interesting news to report. In no particular order of criticality or intensity, I present…
- The extent of what I can say about Battlebots on ABC before the season premier!
- I got a new shiny thing, a MarkForg3D Mark One continuous-filament 3d printer!
- Porting (heh) 2.00gokart to the water: The making of 2.00Battleship for this summer’s SUTD program.
- When it’s not robot season, it’s go-kart season. Time to make some changes to Chibi-Mikuvan!
BattleBots; Or, How Charles Went to San Fransokyo to be a Professional Underground Bot Fighter
It started out innocently enough with an e-mail.
Four days before Motorama 2015, I received an invitation to apply for a new season of BattleBots to be aired on a Major U.S. TV Network™. I immediately forwarded this to our rogue’s gallery of builders of cantankerous machines – namely Jamison, Adam, and Dane. The underlying message was not unlike one BattleBots has told before: that they were in negotiations for a new TV show, and that we should submit a design to be appraised. It’s a story that the builder community had been told several times in the intervening 13 years between Season 5 and the present day.
BattleBots had held smaller tournaments in 2005 and 2009, but the rumors and promises of a new TV show were the subject of constant speculation in the builders’ forums and Facebook pages. Some veterans brushed it off as another rumor. I guess those of us who grew up with the show and were eager to see its return were willing to give them the benefit of the doubt again – especially, as we found out by talking to each other (since the Underground Robot Fighting Circuit is still quite tightly knit), that many of us who had been contacted were “newbies” in the eyes of Battlebots.
So we figured why not. During a break in preparing our Motorama entries, we huddled in a corner and fired off an entry that we thought was entirely bogus. Anyone remotely paying attention would clearly notice that the design was just cobbled together from our 30lbers, copied & pasted and imported piecemeal, smashed together and with cleverly scaled down etek motors to make it look huge.
ITP: Totally not Fission Product.
The entire goal was to try and get us in the door; anything we build may or may not actually resemble the entry. So Motorama 2015 came and went without us giving much thought to the application.
They bought it.
Well, crap. Now we’re screwed… After a few minutes of staring at each other in disbelief, we launched into a series of design meetings.
“Design Meetings”. For very tenuous definitions of “design” and “meeting”.
There were a few different potential design paths that we explored, based on what bots we’d built successfully in the past. Besides historically winning designs, we also tried to design for “weird” – the words of encouragement from the producers and show organizers seemed to emphasize showing how far the sport has evolved and changed in the 13 years since BattleBots was on TV and egging builders to consider modularity, unconventional/rare designs, etc.. One approach was actually sticking to our words: mashing our most prolific designs together into 1 bot. This might have actually involved physically arranging bots:
Yep, that’s 3-wheeld Überclocker and megatRON both in post-Motorama dilapidation. In addition to this genetic mixing, we figured the “old friends joining forces and designs” angle would make for a neater TV storyline. No matter what, the presence of TV will influence you a bit. We understood that this would result in a suboptimal design, but so long as the producers were looking for something out of the ordinary, and this being a wholly new game from what we had been used to, might as well try.
This project pretty saw the most whiteboard sketching and conceptualizing on paper that I’ve ever voluntarily done, outside of being forced to in classes. Here, some higher-level concepts are being appraised, such as “Which ‘bot daddy gets to pass down more genes?”
Thinking of some more armor shape and fabrication ideas…
The whole project was a great example of “CAD-as-you-go” due to the limited timeframe we had. While this lent itself to immense flexibility, there were also downsides, such as the design changing shortly after a set of parts was ordered. Oops… anyone need some 4″ ring bearings?!
There was also a whole lot of what I call “power assisted whiteboarding” – basically working in a CAD program over a projector with everyone screaming at you. It was instrumental in getting linkages and sizes down in an agreed upon range for more detail work individually.
The first parts started arriving in early March. Picture above is perhaps literally 100 bearings of various sizes. Basically, we wanted to get one thing out of the way first: the complication that was the cam-driven walking mechanism. That involves a lot of bearings:
These walkers are quite the work, if I do say so myself. They could move the robot at a theoretical 16 miles per hour – actually above the wheeled speeds of many similarly sized bots. Non-theoretically, we never clocked the robot down the quarter mile, but it was still alarmingly fast for a non-wheeled bot. They are 4-phase (90 degree cam offset) so we get less height “ripple” than the more common 3 phase mechanisms.
Some of the frame tubes cut up, cleaned, and ready for welding…
One of the hallmarks of this robot, as you’ll see, is a gigantic pincer. Here are its side plates after some welding and bending – still far from being complete.
A risk I decided to take with the weapon actuation was to mate Ampflow-150 (nee Magmotors) to Banebots P80s. If you had listened to the collective gasp coming out of the robot builders’ communities when I posted a photo of this, you’d think I was about to build a robot entirely out of dry pasta.
Banebots products at one point had a poor quality reputation in the community, and I think despite massive improvements they’ve never quite grown out of that impression. Inspecting the gear size and materials used in the new P80s, I decided it was worth a try. In fact, the pinion gear on these are through-hardened. I had to anneal it in order to rebore and broach it, then reharden it!
Only a few modifications were needed to mount the Ampflow F30-150 to the P80’s mounting blocks. In fact, the CIM motors these are designed for share a mounting circle and thread size! Granted, I wasn’t about to trust two #10-32 screws with securing the whole assembly together, so in the robot itself, external 1/4″-20 tie rods and a backing plate sandwich the assembly front to back.
Two completed walker pods mark a milestone in the robot’s construction.
Here is what the Ampflow + P80 rigs run: two massive ball screw linear actuators for the lift and clamp weapons. These are 25mm ball screws ordered through eBay, with custom-machined ends ready.
I think high-force linear actuators are underloved in the robot fightin’ world. Many people mentally associate linear actuators with slow, inefficient Acme leadscrews, so designs rarely use them. Ball screws are 90+% efficient in converting rotational power into linear power, whereas Acme leadscrews might only be 10-30% depending on speed and load. So yeah – in the best case, I could be getting 8-9 times more power through. That translates to a faster, more effective lift, or a more forceful pinch.
Now, commercial ball screw linear actuators are super bulky and expensive, so I don’t blame people. However, it’s easy in my opinion to put together your own!
One thing we became worried about as the competition got closer was competitiveness. Through official tournament documents and the rumor mill & grapevine, we became concerned that to enter the TV rounds at all with the weird design, we would need to be competitive through a qualifier round first. We decided it was strategically disadvantageous to be weird all the time if there was a risk nobody would see it. After all, we had foregone the more “winning” designs in the interest of TV, but now the TV was no longer guaranteed.
In addition, Battlebots doesn’t provide a “walker/non-wheeled” weight bonus, which it used to at one point. Events like Motorama Robot Conflict do, and so there is actually an upside to making a more complex drivetrain that is likely slower: 50% more weight for you. Without that advantage, there really is no up-side to the walker design at all. And so faced with a noncompetitive design that could potentially get a drivetrain-heavy opponent first round, we cut our (potential) losses and designed, fabbed, and tested these swap-in wheelpods in the last week before the robot had to ship. It used spare Ampflow F30-400 motors we had originally earmarked for the walker drive.
The wheelpods were designed to hit up to 20 miles per hour to give us a speed edge.
Pointy hardened tool steel pincers for the business end of the robot…
In a fortunate twist of events, a research group in the IDC had bought a Shopbot Buddy (which they had turned into a gigantic 3D printer through a collaboration with ShopBot itself). Its massive crate was just sitting on our loading dock, waiting for a new purpose.
And a new purpose it received. We jokingly called this “the apartment”.
Moving The Apartment downstairs was a whole ‘nother adventure. Notice how it IS the size of the freight elevator.
While I brought it up empty easily enough, once weighed down with robot, a dedicated tool kit, all the spare parts and materials in the universe, and a few silly vehicles, wrestling it back onto the freight elevator and loading dock was an hour-long adventure involving a stuck pallet jack wheel and subsequently all of the van service jacks and lifts I could find in the building. It was jacks on jacks moving the JACD box.
The upside was that we could (and did) take naps in the crate at the event!
You might notice there’s no pictures of the bot itself. Well, that’s because WATCH ABC SUNDAY JUNE 21st AT 9PM EASTERN / 8 CENTRAL FOR THE SEASON PREMIER OF BATTLEBOTS! I swear I don’t make any money when I say that.
Once our episodes air, you’ll probably hear the gory build season details here first. The build season ended up involving a dozen people, some of which are now firmly addicted to the sport in their own right and so you might see them “go indie” for future events. We all learned about individual build habits meshing or conflicting with big team dynamics. And we had one hell of a time at the event and in California.
MarkForged Mark One 3D Printer
At the IDC, I’m always on the hunt for the newest small digital fabrication & rapid prototyping widget. MIT tends to spawn those pretty frequently. In a replay of the Formlabs “I know like half the people who started this 3D printing company” story, I jumped on the waiting list for the Mark One continuous-fiber 3D printer from Markforged. During the middle of the BattleBots build season, it arrived!
I was super excited for this machine since I’d been on the periphery in the gossip cloud regarding its development, and I definitely have uses for high-strength (albeit anisotropic) 3D printed parts. Customer service will also be a breeze, as I could, as easily as with Formlabs, abuse my connections and bang on their door if something goes wrong.
Inside the box is the machine itself, secured for shipping using a box that contains more of itself. Good – at this point, we are already beyond the Flash Drive and Two Nuts stage, so I’m satisfied with my purchase.
The bare machine. It elicited many praises of the form “ooh, it’s so designed!” from onlookers (read: looks like an Apple product). I could do without the Apple aesthetic, since I’m terrified of monolithic products with no visible fasteners and only bearing touchscreens. Regardless, a few people were strung on for 15-20 minutes thinking it was the new “iFab” from Apple, usually before someone else rolled up and asked “Cool, is that the new MarkForge printer?”
Ah, this is more to my liking. The mechanics of it are very solidly built (good, we’re still friends). The gantry is a strange modified H configuration where the X axis (left to right) is run by a straight loop belt but the Y axis is run from the other side using a belt loop in the shape of a T. A regular H-gantry produce “unit vectors” (directions that the head moves in if only 1 motor is powered) oriented at 45 and 135 degrees; this one has 0 and 45 degree unit vectors.
I’m guessing at some point this was an H-gantry. Hey, it works – I don’t claim to know better than them about why it was left this way.
Extruder detail and build plate detail. Like every 3D printer should be, it uses a modified kinematic coupling for locating the build plate. I firmly believe that 3D printers were made for kinematic coupling use (not the other way around), and that every build platform levelling method prior to this is bullshit.
The “box of itself” contains replacement extruder nozzles, service tools, and a copy of its firmware so it can be restored to factory if something goes wrong. I am told that babies also come with the same things, so I can’t wait to have kids!
The machine gets temporarily set up on a table for its functional appraisal.
First bootup shows that the swanky touchscreen controller runs Debian Linux. Nope, not an Apple product.
I spent about a minute wondering where on the machine the large Nylon filament spool inserts. Then I looked on my packing list.
Oh, that’s what the Pelican case is… It shipped separately, landing in the shipping pile behind some large Amazon orders, so I didn’t even see it until I went and looked harder!
Inside the customized Pelican case is a filament spool and mounting brackets. You’re supposed to put the Nylon spool in here, with its bag of dessicant, and close the case up. Nylon is hygroscopic (absorbs water), and as an operator of many 3D printers – trust me on this, any moisture in your filament is bad, and Nylon is particularly bad at it. The reason is that water in the filament boils as it’s extruded – either causing ‘foaming’ or unwanted expansion of just-extruded filament, or straight up plugging the extruder.
Then you hide the ugly industrial thing behind the Great Wall of Design™ so nobody can see it.
At first, I had some issues getting the filament feed to work properly. It seemed to take a long time, and some times didn’t even feed at all. The behavior was so inconsistent that I had a hard time relaying the symptoms to tech support beyond “well, it SOME TIMES works and I don’t know why” – words that I found very hard to say, given my inclination to teach deterministic debugging heuristics to the student population.
Luckily, I was able to abuse my Founder Connections. Team MarkForged walked down the street after the business day and had a look at this machine. The verdict?
A partially-connected connector on the motor driver board. Depending on what I last did – whether I tipped it up or moved it or put the build plate on or whatever – it might or might not make the connection and run the fiber feeder motor. Short of tearing the whole machine down myself, I would not have found this.
A few photos for their quality control team later, I was on my way again:
The Mark One has its limitations, however, like any machine, which doesn’t make it to the headlines. There is a minimum fiber length (about 16-18″) that is required for the slicer to include it in a part layer. That’s because the fiber-cutting mechanism is mounted, along with the nylon extruder, off to the side, using a Bowden Cable style arrangement. So no matter what, in order for the fiber to reach the part, it has to be that minimum length to start with.
That means on small parts of roughly 1 to 1.5 square inch section, you’ll only get fiber if you permit many perimeters of fiber to be laid. Since the fiber reinforcement is only helpful for large parts anyway, I found this an understandable compromise.
I also had some fun getting the cloud-hosted slicer software to run correctly – namely, all the slicing backend worked, but I couldn’t see anything on the screen. At time of press, these issues have long been fixed. It was related to some specific implementation of webGL I had on the random old laptop I found to run the machine. Since then, I’ve relocated it to talk to our 3D printing workstation.
I’ve since had a lot of fun printing kevlar-covered bunnies and other useless things on it, just to keep test driving. But the 3rd? thing I tried printing was this 11″ long reamer wrench, which was used on the BattleBots entry to great effect – it was used to clean up the inner diameter of tubing after it had been welded!
The Mark One lives alongside the other small printers in the IDC like the Up Minis and Form 1.
For the past 2 summers, I’ve organized and run the “Electric Vehicle Design” (a.k.a “summer 2.00gokart”) program for the MIT-SUTD Global Leadership Program which fills up the research center I’m part of, the MIT IDC, during the summers. I still do not claim to know how building silly go-karts contributes to effective leadership, but I digress.
Consequently, since SUTD has shipped the student projects back to Singapore both years, we’ve basically flooded the new university with silly vehicle parts (GOOD!) and my former students regularly contact me with photos of the motorized shenanigans going on over there. The downside is that the program as-presented is getting a little stale – students have seen the parts and the builds and the ideas. Since the MIT-SUTD collab is all about that cutting edge, we held a few meetings during the fall semester of 2014 thinking of ways to spice up the program for 2015.
So for GLP 2015, we present…
Wait. Hang on a second now – I’ve skipped entire exams in the past because it was raining outside. Why on earth would I voluntarily build a boat? And be on water?
Well, between this and custom composite-sandwich bodywork, you could argue that I am just mentally preparing for the Chuckranoplan Revival. I’m okay with this. The real reason is because it’s the progression of the program which carried through the greatest number of its themes – the parts-getting lessons, the combination of design and prototyping and execution, and the “Yes, we’re making YOU ride it”.
Personal aircraft, we decided, was just too much for now. For now.
That, and we happen to have a river named after me in MIT’s front yard.
We decided that it was in the best interest of the summer session to run a small pilot class during the spring to test drive the process and gauge how much more difficult or easier the project is than the go-karts. So, for Spring 2015, students looking for 2.007 Electric Vehicle Design were disappointed – instead, what greeted them was 2.S993 (the S being a special class, seminar, or other nonstandard credit), nicknamed “2.00Battleship”.
The class also introduced Rhino 3D and making nice smooth curvy surfaces therein, and how to translate those to CNC cuttable profiles. I myself haven’t used Rhino since 2003, so what gives? That’s because Brian Chan is my co-instructor for this class. He knows how to make nice things. If you asked me to build a watercraft, I would be hunting for oil barrels on Craigslist and gluing foam to it. Brian knows a lot better:
That’s the “Instructor Vessel” for the spring, nicknamed the “Hack Tool Punt” after an idiosyncratic MIT motto. Brian obtained a few books on the subject, including “Instant Boats” basing this design on one of the examples. The build method is a subset of the “Stitch and Glue” construction… more like “wood screws, clamps, and glue”, but the principles are similar.
On my end, I primarily helped with the propulsion system and electrical system. The structure seen here was designed by Brian also, with a bit of sprinkling input from me.
Likewise, in the class lecture content, I handled much of the propulsion/electrical system load calculation and Brian taught building methods and hull & propeller creation in Rhino. What? Propeller creation? Yes, those white props are actually custom modeled and then 3D printed. Like in 2.00gokart, you can choose to either stay commercial & off the shelf, or go crazy, like…
…3d printing bevel gears for the outboard drive. The gears were sent out for manufacturing via Shapeways’ stainless steel & bronze service. It cost more than if we just, you know, bought some bevel gears, but was a demonstration to the students of the power of modern rapid proto & digital fabrication. They took a little… uhhh, “wearing in”, but ended up working very well.
The assembled outboard. It’s all solid bronze bushings and stainless steel shaft & collars. As it turns out, the propeller is mounted way too low and caused the boat to nose-up a whole lot when under way, so this will be the first thing to change for the summer.
The electrical system is probably one of the simplest I’ve wired up, because it’s one throttle, one ESC, and one motor. We used as many parts as we could with the keyword “marine” in them to entice the students – “marine” battery switches, fuse blocks and distribution blocks, and so on.
In the end, we were the most simple electrical system. Everybody else, unfortunately, had electrical bugs because they tried to go “above and beyond” with one-button remote contactor power switches and Arduino-based joystick mixing and the like. While I do like letting everyone run wild, it definitely put a damper on the end of the semester when a few straight days were devoted to nothing but basic debugging and troubleshooting, and the wiring was done so fast that the reliability was questionable.
Here is the finished “Hack Tool Punt” charging the night before the contest.
Both of our pilot class student teams went for a composite hull approach because of its shape/form flexibility. Only trouble is, composites are actually way harder than they seem. The entire classroom was a bit of a gooey fly trap from epoxy resin dribbles, and a few ‘redos” were necessary. Seen above is one of the student teams’ final products, a go-kartamaran of some sort.
The lessons learned regarding fabrication from the students is that we’re probably going to try de-emphasizing composite hulls, or at least making the introduction (to building methods) lecture hammer a lot harder on the fact that it has hidden difficulty and immense amounts of goo.
The other team went for a more aesthetic approach – they made it look like a swan. The catamaran body of the swan was also composite, but instead of solid foam, it was made of individual facets bonded together and sealed.
In fact, they spent a little too much effort making it look like a swan. For summer, Brian and I are instituting a strict “show it working first, then make it pretty” guideline. The students on this team basically underestimated the magnitude of the project – of making the electrical system more complex and the hull more complex to build yet also aesthetically pleasing. A completely valid outcome, in my opinion, and teacher of many lessons.
Instructor team photo before we all go in the water!
Brian taking the HTP out for a quick cruise.
And now come the obligatory “Charles on the Charles” jokes. This photo is a good example of the front riding too high. There’s a lot of optimization that could be done on this design, which was shown very clearly when we were passed by a sailing dinghy.
I think running the pilot course in the spring was a life-saver – there were many, MANY changes we (collectively, at the end of term with the students) discussed which would make the project more likely to succeed in the much shorter 8 weeks of summer. One of the students is coming onboard as a TA to relay some of these life lessons to the Singaporeans. Like with the GLP go-kart class, I think at the end of summer I’ll appraise the results and compile the new marine-oriented lecture notes we’ve made to something that is Internet-presentable.
The GLP is now under way – we’re 1 week into the program already!
Chibi-Mikuvan: PPPRS 2015 Season Upgrades and Mods
Chibi-Mikuvan has been sitting in roughly the same location since the MIT Mini Maker Faire exposition race and the Miku Expo. Well, fine – with the exception of a single winter shenanigan. I bought some parts in January/February to work on the upgrades it was slated to get, but… umm, I had an unexpected robot.
I’m aiming to make no changes to the powertrain side of things since it hasn’t gotten to a level of reliability where I can make a call on whether or not the sensorless drive & high reduction method is bogus or not. Recall that all of 2014’s fuse-blowing issues and motor power losses were caused by an unsecured, eroding connection with the motor, which I finally resolved for Miku Expo in preparation for this season, so the only deltas for the back half of the vehicle is better wire securing; I may elect to act on my desire to turn the motor’s 8mm female-side bullet sockets to six pigtails so the connector isn’t being stressed in bending at the motor.
The changes I want to make are all on the front half. To recap the season conclusion from New York Maker Faire, I want to:
- Move to proper double-supported steering knuckles – the single-supported box was easy to make and fit the geometry well, but started becoming more of a parallelogram with each race.
- Also changing each rod end (ball joint) connection to double-supported
- Adding caster angle to the steering wheels to make cornering more predictable and less reliant on rider weight shifting.
I now add one more insane item to the list:
- Making all of these replacement parts out of nylon-fiber composite printed using the Mark One.
It’s another one of those things which everyone so far has told me I’m crazy for doing. Good! Because may Robot Jesus Himself prevent me from building something normal and functional!
From an engineering and strength of materials perspective, there’s no reason why it wouldn’t work. After all, they make real PowerWheels toys from plastic, and disturbingly enough, parts of real cars.
In my opinion it’s a problem of geometry. Not only is the challenge in putting plastic where it’s needed to react against the force of handling, but I would also need to take into account the massive difference in strength between loading in the fiber layer plane vs. between layers (anisotropy). With a typical filament printed part, that difference might be only 30%, but with closed loops of fiberglass and kevlar in the mix, I think the parts will be > 90% stronger in the layer than between them*.
*This is my scientific wild-ass guess and not supported nor approved by MarkForged in any way.
The difference between these new parts and the equivalent parts in steel is mostly in bulk. Because plastics are less rigid – and 3D printed plastic even less so between layers, the parts are going to look fairly “fat”. Unlike injection molding, however, I can carefully control the voids inside the part. It’s better to think of the parts as being mostly hollow and with the skin layer and fibers taking most of the load.
I started modeling the most difficult part – the new kingpin post & combination A-arms and radius arm, or something. Go-kart parts tend to degenerate so far from real car parts that they stop having names. This is the stick that the wheel pivots on.
I modeled the interface to the frame first. The current “winglet” that holds the kingpins will be cut off and ground to shape, and this part with the alignment claws will be drilled and bolted to the area. Next, I made a series of offset and angular planes to capture the caster angle.
The interfacing shape is a solid triangular brace in 2 planes – one flat (normal facing forward on the vehicle) and one at 45 degrees.
When printed, there will only be horizontal bands of fiber running between the post and the two tall cylinders on the right. The majority of axle up-and-down bending loads will be handled by this closed structural loop. Nylon material on the top and bottom of the kingpin post will only handle compression forces.
Long bolts will capture the fiber band and transfer the load directly to the frame, and there will be huge thick washers holding onto the nylon skin on the bottom. Notice here that I’ve gotten rid of the “claws” that grab the frame rails. I’ll just align and clamp & drill manually – my concern was that the claws wouldn’t represent the non-ideal frame in real life, which is welded, painted, hammered upon, etc.
The new steering driving link (“Pitman arm” in car words) is basically an exercise in design for injection molding. There will only be 4 bands of fiber here – a few mm on the very top and bottom, and a few mm of closed loops on the flat portion of the long protruding tab. Bending rigidity is handled by the X brace inside on both halves of the tab, and torsional rigidity is… well, it’s double supported anyway, so hopefully it won’t matter!
The clamping will be done via two 1/4″-20 bolts – again, overkill by size, but necessary to take advantage of contact surface area to grip as many layers as it can.
Here is the new assembly modeled on one side.
The new steering knuckle, in light pink, will be printed flat to have large C-shaped sections of fiber holding onto the axle stub (a cut-down 5/8-18 bolt). The distance between it and the kingpin post is taken up on both sides by needle thrust bearings, compressed by a 4.5″ long bolt. Side loads on this assembly are handled by virtue of the fact that I’m rubbing large surfaces of nylon together.
The red steering follower link is a design compromise, since I needed high strength in 2 planes on this part. The proximity of the X-axis oriented long cap screw to where the ball joint mounts will hopefully aid in transmitting the steering force by relying on shear between the layers (which is a bit better than bending between them…). The closed fiber loops in this part run in the XZ plane (horizontally) to hold onto the ball joint.
The purple brake caliper mount should be seeing most of its loading in the XY plane, from braking, so printing that in the flattest configuration to get a bunch of fiber laps around the outside is easy.
I do need to come up with a mirrored brake caliper mount – they won’t be identical due to the need to position the brake cable exit. In building CMV, I’ve found that the left- and right-handed versions of the same product are in fact not identical at all in stroke/braking force. Pretty expected of shady scooter parts, so I use two “right hand” calipers instead, necessiting either a 180 degree rotation or some weird repositioning.
Enough talking about parts – how about some real parts?
Here is the new Pitman Arm fresh off the machine!
The follower links came out a bit…. hairy. The current state of the Mark One has some trouble with smaller cross sections since it relies on clever timing to start pulling the fiber through the extruder nozzle, so if it start moving before the fiber actually reaches the head (e.g. extra friction in the system, or your feed motor sleeping on the job), a few mm or … inches… of fiber might get missed, which then gets whipped around after it’s “done” depositing.
They clean up nicely, however. There may be some voids in these follower links because of the missed fiber lay, so I might mark these as emergency backups and reprint them. In the mean time, check them sitting on top of the (very successfully printed) knuckles.
The longest print is the
kingpin post and combination A-arm and radius arm thing I’m just going to call this The Banana from now on, which takes 1 day and 22 hours apiece. Once I have two Bananas, I’ll try to make a test fit of the assembly, and if I like it, the steel frame is getting modified this weekend!
Whew, that was a lot of life updates… Now back to it. Hopefully, with my surprise robot business done for now, I can return to the things I was doing. For instance, it’s time to return to the long-neglected RageBridge V2….
Oh, and don’t forget…