Gracious Pipefessionalism: I May Have Gone A Little Too Far

Some of you may have heard of my 1lb fighting robot, Satan’s Segway.

I’ll put the details of the build in a later post, but after scoring an unbelievable 2nd place at 2015 Bot Blast, I decided to do the logical thing and scale it up to a weight class that would cost too much money and probably be utter junk in.

11889529_10155973232200424_3271120459747749575_n

Aluminum pipe gets swapped for black steel, spongy wheels for Hazard Fraughts, Silver Sparks for DeWalts. And the best part is? CAD says I make weight.

2 lbs to play with what what

2 lbs to play with what what

And that is the last I cared about the CAD. Seriously. I’m not modeling anything else in this, and weapon geometry is going to be determined solely by a scale and however much mass I have left. That being said, a sledgehammer and a shovel are both pretty appealing options…

Next up was sourcing all the bits. McMaster would be the easiest option, but I know better than that. Metropolitan Pipe Supply in Cambridge has every sort of pipe fitting you could ever imagine, and if they don’t have it they can tell you where to get it. Everything is NPT 2 1/2″ except for the weapon arm which is a more reasonable 1″.

Machining was pretty straightforward on this guy. Starting with a stock DeWut gear box from Charles Guan’s Equals Zero Designs, I turned down the front plate to fit the 2.5″ ID of the pipe. Interrupted cuts with carbide are fun, but with the right speed and feed can still get a lovely finish.

NCM_0040.JPG

Early on, I decided that Charles’s rear motor holder/billet wouldn’t quite work for the compact nature of the Pipebot. So, I carefully and painstakingly machined an adapter ring out of some nylon I that was lying around.

NCM_0055.JPG

Only after I had 3 of the terrible made did I realize that there was a nylon-extruding 3d printer upstairs which could do a better job. The final result bolted together pretty well, and slipped right into the pipe fittings.

NCM_0073.JPG

Next up was the power. Initially, I thought I could fit 8x A123 26650 cells in there for a total of 27V and 2.5 Ah. When I started to put that together, I realized that I was under some strange hallucination when I planned it out and those just wouldn’t fit. A quick trip to Dane Kouttron’s Magical Mystery Land got several mango’s worth of the smaller 18650 A123 cells, which I planned on making into a nice modular twin pack of 7S1Ps

NCM_0083.JPG

This was my second time putting together a battery pack, and my first dealing with 18650s and soldering tabs together. 1/4″ wide copper braid for McMaster was used, as well as the Weller 80W MEGA IRON. If you ever need to solder battery packs or wire 14 gauge or bigger, I cannot recommend this stupid thing enough.

NCM_0080.JPG

Each of these packs was a separate 7S1P hexagonal pack. The plan was to pad the crap out of them with rubber and swap them out in case of a single failure. I made a total of 4 packs, and they worked very well in pairs.

NCM_0087.JPG

I did have a little bit of hassle with the balancing leads. My charger, the illustrious 1010B+ doesn’t have a 7S or larger plug on the balance lead. I used a pair of 4S leads and removed 2 leads from the second plug. This was a mistake, as every time I plugged it into the board, two of the leads immediately shorted to each other and melted. Taking a look at the board, the last lead of the first plug and the first lead of the second plug are connected, which shorts when you have those two leads going to either side of a cell. The right way to do it is to remove the last lead from each of the balance plugs, isolating that one cell. No serious damage was done, I’m glad I did this with the high-resistance 18650 cells instead of their “sparkier” 26650 cousins.

NCM_0076.JPG

Finally getting stuff together, you can see how tight the spacing is! As it started to come down to the deadline, my priorities got a little bit jumbled. I had all of the precision-machined components finished, but no total assembly, no wiring harness, and no weaponry. With the maturity of an 11 year-old, frosh-army at my back, and 6 hours until the drive to Philly, I set about making the wiring harness while John and Andrew threw together as many bad ideas for melee weapons as we had time.

NCM_0074.JPG

Here’s a good look at what I had only a few hours before take-off. With a 1″ pipe arbor, I could put pretty much anything I wanted on to the end as long as it had the 1″ NPT base.P1050643.JPG

First up was my “Triggo” plan, which ended up losing the bottom board for weight. It’s a Kevlar-wrapped E6000 impregnated 2×4 board that skirts entanglement rules as close as I’m willing to go. The name of the game is play Sewer Snake with this and run into people until their brushless E-Penors burn up from stalling.

NCM_0094.JPG

Next up was the craptastical “bent 1000 times” sword donated by my lovely roommates. I’m not sure what the strategy was here, but what is the strategy with a pipe-bot anyway?NCM_0096.JPG

Finally, a broom, to “sweep up” the competition. The plan was to use this to clean the arena between matches for an extra 5lb weight bonus, but I spent too much time repairing GP to actually use it. Other attachments include a forged-steel scraper for use as a wedge and a pickaxe for….reasons.

P1050619.JPG

I finished the wiring harness at 2 AM, threw everything John and Andrew had cooked up into the back of the ‘scoob, and rocketed off to Philly with Ian and Ciaran. Here I am, putting the whole thing together for the first time at 8:30, fifteen minutes before safety. The motor controllers, BotBitz 85A are a tight fit between the back of the two motor cans, but it works.

P1050667.JPG

The twin battery packs and wiring harness are an even tighter fit. Up to the chagrin of Dane, here I am shoving the batteries in with a rachet wrench. It might be a tight fit, BUT IT IS STILL A FIT. SHIP IT.

Right off the bat (or sword) I noticed a few major problems:

#1. The pipe legs with the motors would ‘torque’ themselves free, tangle up the wires, and pull a few connectors out.

#2. I was WELL beyond the discharge rates of the 18650s, even with them in 2P. I was also reaching the current limits of the BotBitz controllers and the DeWalts, which caused a massive thermal buildup inside of the tight quarters every time.

#3. Every single weapon I had broke.

That being said, the 2×4 was designed to break and did amazingly well against GK’s insane undercutter. Also, it was incredible to see how the driveability was improved by 100% when I lost the wheel. The “twisting out” was solved between matches by Dane once again, with the addition of set-screws. Probably the one time they don’t suck…

P1050684.JPG

10/10, would pipe bot again.

Next time? 8″ pipes, twin E-Teks, and water-jet AR400 planetary gearboxes.

Advertisements

Rage Frosh, Lost in Space, and Robots of MASSdestruction.

This last weekend saw the inaugural class in the World’s First Boston Area Robot Fight aka MASS DESTRUCTION held by the Venerable Rob Masek at Artisan’s Asylum. As this event was only fifteen minutes away instead of the usual fifteen hours, a fair amount of MITERS folks (and other Shenanigans-Filled builders) were up until half past dumb O’clock working on ROBOTS. Since my stuff was finished a week ago, I helped MIKE, ANDREW and JOHN throw together a 3lb entry from some old Rmr parts I had lying around. The result?

Photo Credit Charles Guan

Photo Credit Charles Guan

Slovania-Mania, or Ljubljana Lasagna, or Slovenia-Bot. Whatever you want to call it. The blade is Amazonian O-1 straight from the “Bad Ideas” pile of Rmr spares, the motors, controllers, escs, and wiring are all also Rmr spares (which turned out to be pretty bad when I needed to replace everything on Rmr during the fights). Wheels came from LUC-I, top and bottom plates were waterjetted out of spare aluminum/G10/Wood from Charles’s Uberclocker. The original plan had that large 3d printed body as the body, but weight concerns made us ditch that in favor of custom standoffs and spring steel.

Which worked pretty well! The first match was Lucy’s Deathcap, an extremely mean vertical drum made from Jamo’s Dominant Mode, wicked sharp S7 and titanium teeth, and Pure Hatred. John and Andrew were running scared until the first few hits, when they broke half of her teeth and flipped her over. Unable to self-right, it would have been the end of the match had I not told John how bad of sportsmanship that would be. He smacked her a few more times, at which point she got the pokey-end up and completely wrecked everything forever.

Photo Cred Charles Guan

Photo Cred Charles Guan

At which point, Mike, Andrew, and John rapidly threw all of the electronics back into the 3D printed body and covered it in tape and old armor plates from Rmr, as well as a weaponized drill bit ram horn and impact-resistant tire armor. Which almost worked, until it didn’t.

Charles Guan again. Will Bales and Herp-a-Derp on the kill though

Charles Guan again. Will Bales and Herp-a-Derp on the kill.

Not that I did too much better.

Credit to Jamison Go for both the photo and the hit. Despite evidence to the contrary, he's weaponless now because of it.

Credit to Jamison Go for both the photo and the hit. Despite evidence to the contrary, he’s weaponless now because of it.

A year of late nights and dumb ideas comes to an end. RIP in Peace Rmr, IT’S ABOUT TIME. Onwards and upwards, but I’ll save that for later in the post. All in all, my favorite robot fight I’ve been to with fast-pacing, lost of TKOs, and an incredible arena. Thanks again Rob!

Charles takes a picture of ALL THE FUN WE'RE HAVING BREAKING STUFF

Charles takes a picture of ALL THE FUN WE’RE HAVING BREAKING STUFF

Like a reasonable person, I woke up the next morning and went to a car sho-

NOT A CAR....?

NOT A CAR….?

Well, it was a something-show hosted by MIT. I helped MITERS shuffle over Silly Electric Vehicle Armada, which was a good deal of fun, and we got great looks from the crowd and a serious talking-to by the police for “Why didn’t you guys bring your even more dangerous electric vehicles?!”.

They see me Doofin'...(Charles Guan)

They see me Doofin’…(Charles Guan)

Doof-Kart was hotter than Baked Alaska at an Antarctic Research Station though, making me think that an autonomous and a commercialized version might be in the works soon. We still need to “finish it” also known as give “Charles his Vyper back”, but it is really really hard to justify taking apart something that just plain works. It was also a big hit with the little kids, who really had fun playing around with it. One of them happened to be the son of Tim Hirtzil, of “Boat Bot” fame, and he decided to have a few words with us.

The ‘boat’ elements of Boat Bot have seen far better days (foot-long cracks up and down the gunnels!) but the ‘bot’ part is as shiny and chrome as the day it arrived in Valhalla. And now it’s in my basement, awaiting A Good Cause like any apathetic boomer afraid of what his family has become. I’m currently feeling “Unnamed MITERS All-Terain Amphibious Assault Driver” or “U-MAAAD?” but if anyone has better suggestions, please send them along.

After all that, like a reasonable adult, I immediately went back to building battlebots.

Master of Yards

Master of Yards

Yard Master was a 3lb Silent Spring clone I had planned for Bot Blast. Those sneaky frenchy sons-of-whatevers down at McCaster-Barr decided that shipping metric stock to people dimensioned as SAE but with really really wide tolerances was a cheap and easy way to improve sales, and they’re right. Unfortunately, this means I had to mill down all of the plastic stock I got in the mail by about .2″. I was being lazy and accidentally Did The Thing while doing that though…

Nooooooooooooo (Rob Masek)

Nooooooooooooo (Rob Masek)

So I gave up, despite having all the parts except for a new sheet of Tivar. Enter Team Rage Frosh, looking to build a new 3lber for Franklin Institute in two weeks. John was very interested in taking over the design, but had some concerns about my “magnetic clutch” design.

NCM_0012

See, the blade is full-hard A2 off of Rmr and has tapped 6-32 holes for that hub. The pully for YM is nowhere near that big, so I had to design a really crappy adapter plate that I just didn’t like. So I decided drilling holes in everything and filling it with magnets was a legit idea.

Science seems to agree with me. It also does the “little give at the start” thing that I’ve been playing with for mechanically soft-starting sensorless brushless motors. The other idea is that it’ll also preserve my everything in case of a big hit, but we’ll see how effective that really is later. In the mean time:

NCM_0016

This is a post-Pipe-bot 30.0 idea, but I’m starting to like a big circular blue “Boom-ba” with all brushless drive on 4s as a handicap. Otherwise, it’d feel a little too much like mixing 3lb Touro with Weta, and I’d feel a little unoriginal doing that.

Next post I’ll talk about either my new Jerb or Gracious Pipefessionalism, stay tuned to find out!

Christmas 2013

I’m a bit of a miser when it comes to birthdays and Christmas. If I can get away with making something cheap and novel that gives the impression of caring instead of buying something, I probably will. More often than not, I end up actually caring and going way too overboard on my creations. This Christmas was no exception, and the addition of a Sherline mill and lathe at my work only encouraged my habit.

NCM_0089

Target One: My Roommate and her Boyfriend

A Japanese major and her mechanic boyfriend, I tried to think of something that was both simple, seemed nice, and the two of them could enjoy together. Seeing her near-obsession with Japanese food, I made her a lovely set of turned aluminum chopsticks for two.

NCM_0069

6″ length with a distal end taper and a proximal end round, I think these came out looking pretty nicely, especially after a scotch-brite pad. Ultimately, I was limited on length due to the size of the lathe, but they did serve as a nice early test on the length abilities of the Sherline.

NCM_0071

The four of them sit nice and tightly inside of a piece of aluminum tube extrusion, and the large proximal ends force them to “jam” against the sides of the tube and keep them firmly in place. Tied up in a clean rag with a pretty bow and it made a nice project.

Target Two: My Dad

A bit more difficult, but I had this one in mind for a while. Being a professional and fan of simplistic and easy to use items, I figured an aluminum pen with a Zebra cartridge would go pretty well with him. Messing around with the lathe for the first version though made me realize the necessity of a prototype, so two versions were made: a shortie for my friend back home, and the second that would actually go to my father.

NCM_0083

The pens were of a simple cap style, nothing fancy. There is an inner tube, front taper, cap, rear ‘weight’, stock pen cartridge, and a clip stolen off of a few pens. The first, smaller version used an O-ring to retain the cap on, but the second used some simple colored wire banding.

NCM_0084

NCM_0088

Target Three: My Girlfriend.

A mechanical engineer like me, I figured she’d appreciate a more “raw” present. She’s also a pretty big Americana fan, especially with religious iconography. This, coupled with my ongoing exploration with a Replicator 2 3D printer lead me to want to make an aluminum Christmas tree decked out with a new 3D printed ornament for every day of December. This coincided with Thingiverse’s contest on a similar subject, making finding designs easy. 

NCM_0053

 

The trunk was made up of the hollow aluminum tubing seen in the chopstick set. Coupled to a lab-stand base made it stable and easy to attach to any surface.

NCM_0056

 

Alternating 90 degree holes evenly spaced up the length held press-fit reducing lengths of 1/4″ aluminum rod (noticing a trend with my materials here?). 22 gauge wire was wrapped around the outer points of the branches in a double-threaded spiral, and 28 gauge was used in the opposite direction on the lower branches to make a second spiral. 

1472945_10152470869078012_961191402_n

 

Finally, to throw it all together, I used a laser-cutter to cut out a nice little star to top the tree. Using some acetone and c-clamps, I welded it together and threw it on.

Needless to say, she was pretty happy about the whole affair, and I was pretty pleased to get to call messing around with engineering tools a Christmas shopping spree. 

Straight Outta the portfolio again: Unnamed PVC ROV

0811081121In retrospect, I probably should have named this one. It started as a Sea Perch Project that MIT’s SeaGrant Lab gave me to keep me busy and entertained the summer of 2008, and quickly escalated into my Extended Essay for my high school’s IB program the next year. It was my first experience using a lathe and a mill (yes, I did shoot the chuck key into the ground) and other than the shafts I made for it, is entirely COTs parts. After I brought it home with me, I added on a camera, water bottles for adjustable ballast (instead of the buoys) some fishing sensors, a claw arm, and a water-sampler. It also is the second use of my “control” system that was seen in Steve-V1-Der and seemed to dominate my high school aversion-to-programming career.

IMGA0301I think that control scheme is small sticks for big thrusters and elevator, red buttons for panning thrusters, and black DPDT rockers for the arm.  This whole project was really important to me because it taught me a very important lesson I ended up using heavily in my Undergraduate Thesis: Waterproofing Things is Hard.

0805080934

I based most of my design off of the methods and plans laid out in Build Your Own Underwater Robot and Other Wet Projects by Harry Bohm and Vickie Jensen. Initially, I made the construction out of cemented PVC pipes and fittings, plastic rods, and small buoy floats, changing only the chassis design. Instead of going with their standard rectangular prism with non-parallel motors, I chose an upside-down pyramid with parallel motors. This allowed the ROV to descend much quicker than a previous design. I also ended up swapping out the foam buoys for empty water bottles, a far more readily available substitute. The bottles, while not looking quite as nice, allowed for adjustable buoyancy as well as greater overall buoyant force. In addition, I also changed the book’s recommended wax-potted motor for several modified and pre-waterproofed bilge pump motors with custom drive shafts and larger propellers.

IMGA0297

Other, more off-the-book modifications I made included adding in a few cheap off the shelf sensors, a remote fishing camera off of eBay, and a custom-made grabber arm. The off the shelf sensor was an inexpensive fishing tool designed to get both luminosity and temperature. The fishing camera came with its own monitor and displayed a laminated black and white picture. The arm was made of a long single piece of PVC tube filled with expansion foam, and was hinged on a waterproofed window motor allowing for easy articulation in the up and down direction. A simple waterproofed servo in an off-the-shelf grabber hand sat at the end of the arm and there were custom brackets with a test tube that could be attached to obtain water samples.

IMGA0296

The added power of the motors made a huge difference in the mobility of the ROV right off the bat. Additionally, the new shape of the ROV made descending much quicker at the cost of greatly-slowed rising speed. The sensor set worked perfectly at telling luminosity and temperature in the murky waters of the test lake, but the camera was of poor build quality, leaked, and broke. The actuators for the arm performed aptly, but it was difficult to use without a camera. The write-up for it in my Extended Essay received a passing grade nonetheless.

IMGA0304

 

 

 

Straight outta the Portfolio: Steve-V1-Der

Steve-V1-der was the first “robot” I ever built on my own.  With uthermal, of course, but still, it was a lot on my own. A very simple design, it was made with saws and drills and things that could be bought at home depot or radioshack (the exception being the Solarbotics wheels and motors).

 

Mid-Stage Steve-V1-Der

The National Science Olympiad Robot Ramble competition for 2008 was to collect a variety of objects and place them into a square box. The objects included tennis balls, D-cell batteries, and a notecards. Additional points were awarded if the 3 balloons in the box were popped. A time limit of 3 minutes was placed, and tethered robots were allowed.

After watching previous years designs fail due to complexity and being limited by expensive wireless controllers with 4 channels, my friend and I decided to go for a simple, tethered design with zero ICs or complex circuitry. Non-toggling rocker switches were wired in such a fashion to allow us reversible current flow through the motors, and they were directly wired to the battery pack and motors. These switches were in turn placed inside of a simple plastic project box from Radio Shack that had the battery packs velcro’d onto the back. There were 3 total switches, 2 for control of the drive train and the third for the actuator.

1208071556

The robot itself was equally simple in design. The frame was bolted-together 1” angle aluminum stock, cut with a hacksaw to the sizes we needed. The motors were cheap Solarbotics Gear Motors with integrated wheels, with 4 wheels being dedicated to driving the robot. The scoring device was a simple Plexiglas ramp with 2 modified drawer sliders spanning the length of the robot, from the lower front edge to the top rear edge. The drawer-sliders were connected via a c-shaped cart, from which a spring-loaded “door” was hinged. The far end of the door had 2 lengths of high-strength fishing line attached to it which ran to a powered spindle underneath the ramp. When the spindle was fully out, the cart would be at the bottom of the ramp, with the door extended out, allowing objects to pass into the C-shaped cart. When the spindle was pulled in, the door would swing in, push the collected objects into the cart, and begin to pull the cart, door, and objects up the Plexiglas face. Gravity would carry the cart back down when the spindle was released; allowing us to create a very simple robot that only required 3 points of actuation. To pop the balloons, we had pre-loaded the cart with Styrofoam balls filled with T-pins (pointy end out). Our first move would be to drop these “spikey balls of death” into the box, popping the balloons effortlessly without any need for another actuator. This robot we called Steve-V1-Der (Stevie Wonder) in honor of our team mentor, Stephen Orr.

0313080034

For our next version of the robot, we based our strategy off of one we had seen online. The robot had collected all of the objects together outside of the box, then picked up the box and placed it on top of them. We kept the same tethered rocker-switch control scheme for this one, and upgraded all of our motors to try to gain more speed and power. The new robot also had the bolted aluminum angle stock frame, and now had the drawer sliders up front and vertical, resembling a fork-lift. With this design, we were easily able to pick up a Plexiglas scoring box and shuffle all of the objects together. We named this robot Steve-V2-Der (Stevie Tudor).

The regional competition for Steve-V1-Der had a very non-standard field, which consisted of a heavy wooden scoring box and shag carpet. Many of our competitors had used VEX systems or other heavy metal frames, and found themselves unable to move in the carpet at all. The light aluminum frame, simple design, large wheels, and off-board batteries of Steve-V1-Der were able to easily maneuver the shag carpet, collect all of the objects and place them into the scoring box. None of the other 3 competing teams were able to score at all, making Steve-V1-Der win by a landslide.

Steve-V2-Der was a redesign for the national competition, and proved to be less-successful than his predecessor. The national competition also had a non-standard field that also had a heavy wooden scoring box that Steve-V2-Der was unable to pick up (few of the other competitors were able to pick it up either). It placed 36th nationally, whereas the simplicity of Steve-V1-Der would have fared far better.

The Excelsior; the Latest and Greatest in Hat Technology

The Excelsior; the Latest and Greatest in Hat Technology

Every year around graduation time, the students of MIT are encouraged to come out of their celebratory drinking comas just long enough to design their own mortar board hats, making the bland black squares a reflection of their artistic abilities that have been suppressed in the name of science for the last four or five years. The sea of inky shifting rectangles comes alive with the brilliant colors of rhinestones, Legos, laser’d acrylic, and foofy ribbons and bows. The more mechanical and electrical minds can be seen with gears and LEDs on theirs, blinking and spinning a chorus of controlled chaos in the cool cacophony of Commencement.

My art is not one of lights, cameras, or action, but one of practicality and hacking things together that probably shouldn’t go together. Like a Chinese Voltron toy, I shoved, strapped, stripped, soldered, and Amazon’d together my latest masterpiece: The Excelsior, named after the speech given by Drew Houston during the festivities. Appropriate for its namesake, it was both literally taller than all of the other hats due to its size and better than all of the other hats, truly rising above all. Also like Drew Houston’s speech, it was made up of a bunch of things that I found on the internet stapled together.

The hat had every feature that one could want out of a hat, and then more. An umbrella shielded me from the elements, while a foam dome of coke cans provided me with a constant stream of refreshment. A miniature fan blew cool breezes across my face, ensuring a comfortable temperature no matter the conditions. A small spinny propeller was also planned to top my mortar board, but was unfortunately damaged during assembly and was not included.

More unfortunate occurrences came with the line up to the march in, where I discovered that post-Marathon-and-Mooninite Boston really doesn’t like exposed circuits. Or Coke cans. Or umbrellas. Or fun, student health or well-being, but I knew these already. The headgear was confiscated and destroyed by Hat-Fascists masquerading as event security, another victim joining the sad ranks of Quadcopterhat, Vinylhat, Alsoumbrellahat, and Rockethat (to be fair, I knew that last one wasn’t going to fly at all, either by security or later when we actually tried to launch it).

MIT, so long and thanks for all the Pneumonia from sitting out in the rain exposed for 5 hours straight.

(Photo Credit: Bryce Vickmark, a nice man who stalked me through the Infinite prior to the hat’s destruction)