Project_raws on Aaron Silman

Mon, 01 Jan 0001 00:00:00 +0000

Bartomaton: The Bar_tending Au_tomaton

This project started in grad school as an excuse to use the 3D printer that my friend had just purchased - a DaVinci 1.0A which I modded with the Repetier Firmware. We were both excited about it and wanted to build something.

You can skip to the GitLab page here

Initial Concept

We started our design by first looking at other similar projects. We stumbled across one called the BarMixvah and decided we would start out by copying that design since it was fully open source. We fired up the 3D printer and started churning out the parts. We purchased steel rods from Blue Collar Supply in Elk Grove for cheap to use as the frame and we purchased some peristaltic pumps on Amazon. The BarMixvah was the creation of a front-end engineer, he ran a nodeJS server on his computer connected to an arduino and had some javascript libraries to send commands to an arduino which controlled the peristaltic pumps. I wanted to take my own spin on things - and have an excuse to learn something new - so I started working on a new software architecture for the system.

Choosing frameworks

I am not a web developer. I have dabbled in some small creations here and there but barely scratched the surface of any web development work beyond following tutorials. I decided this was a good chance to get my hands wet and learn a thing or two so I started researching frameworks for the front and back ends of the system. The only thing I really knew going into this was that I wanted the back end to be in Python since I know that much better than the few javascript tutorials I have done.

I ended up decided on Flask for the backend because it seemed simple yet extensible enough for my purposes. For the front-end framework I decided on React after trying out small tutorials on Angular, Vue, and React and deciding that React just ‘clicked’ with me better than the other two, Reacts model of passing props and modifying state just seemed fairly inuitive to me and I liked the amount of support that was behind it.

For the database model I decided to use SQLite with a SQL Alchemy interface since I know nearly nothing about SQL.

After following some more Flask and React tutorials I got a basic site up and running with a basic database. Upon loading React would query Flask for a list of drinks, Flask would get these from the database and pass the mup to React which would render them each in their own little boxes. Each box would have buttons to make the drink or modify the ingredient ratios/amounts.

Database Design

The databased design for Bartomaton was a little tricky for me since i’ve never worked with any databases, but I knew what kind of architecture for it I needed. There would be many drinks, each with a list of ingredients and amounts for each ingredients. These Drink Ingredients would need to know which (if any) dispenser would dispense it. I made the assumption that a single ingredient would only be loaded into one dispenser but if need be this relation can be modified to get many dispensers for a single ingredient, and choose which one to use based on some other criteria (such as some sensor or mechanism which determines how full each dispenser is). Each Drink Ingredient maps to an ingredient entry which maps to a Dispenser entry (or None/null if no Dispenser currently has that ingredient). In additon the Ingredient entries have one to many relationship that can be used to query all drinks that use this ingredient, allowing one to filter down any possible drinks based on the loaded ingredients. The Dispensers need to know which output of the multiplexer to map to so each has a gpio_id which determines where the multiplexer outputs the signal to the motor relays. This formed a relatively simple database structure:

System Architecture

I knew we wanted at least 8 liquids and driving that many motors was impossible from the Rasperry Pi directly. Instead of using 8 relays connected to 8 different GPIO I decided to use a multilexer. The drawback to this was that only one ingredient could be pours at a time.

We started out with peristaltic pumps but realized quickly that cheap peristaltic pumps are far too slow to make a drink quickly (pouring at about 2.5 mL/s), we either needed beefier peristaltic pumps or change how we delivered the liquid. My mechanical engineering friend went to work and designed a new system architecture wherein the different liquids would be pressurized with air pumps and the liquid would flow out when we turn a valve. This required us to actuate valves but increased the pouring speed by nearly an order of magnitude (at full pressure). The other drawback to this is that we needed a new way to measure the amount poured, since peristaltic pumps are very linear in the amount they dispense vs their actuation time, this air ressure based architecture would vary depending on a lot of factors. As a result we needed some pressure/weight sensors under the glass to detect how much was dispensed.

Since the movement to air pressurized cannisters with servos to actuate the valves required a small change to the database as well. Now a dispenser is by two different peripherals, one pump_idwhich defines which air pump is used to pressurize the ingredient cannister as well as another servo_id which maps to which servo t actuate to release the ingredient

Hardware

The main brain of the Bartomaton is a Raspberry Pi 3 which runs the webserver and controls peripheral devices over the GPIO. The Pi interfaces to a board I designed in KiCAD which contains four half H-Bridge motor driver ICs each with two paired outputs (for up to 16 unipolar outputs total) as well as an I2C bus to a PWM peripheral that can manage up to 16 PWM signals to control the servos.

Mon, 01 Jan 0001 00:00:00 +0000

Coropunt

Living near a trout and catfish stocked lake fishing is an activity I often do on weekends when there isn’t much else going on. Unfortunately I don’t have a boat and fishing from shore seems to yield much worse results than if you’re out on the water. When talking to my housemate one day about buying a boat the problem of how to store and transport it arised as the biggest issue. He suggested to me a folding kayak which he has seen but those seem expensive for what you actually get. After doing some research I stumbled upon https://www.portableboatplans.com, a website hosting the design plans for various portable DIY boats designed by a retired NASA Mechanical Engineer. Most of these boats are made out of a single 8’x4' sheet of Coroplast - a corrogated plastic used for large signs.

I looked through his designs but nothing caught my eye as the immediate boat I wanted to build. I like standing up when I fish and the only design he had for a boat you can stand up in is designed to be made out of wood (which would be much heavier than I would like. Eventually I decided to just email him and ask if he could suggest a design for a plastic boat that I could stand up in and have maximal room to move around.

To my surprise he emailed me back saying he would draw up new plans for a design that met me needs, I was a little dumb-founded by his willingness to design a plan to my specifications but I suppose that is his hobby and he was looking for a new challenging design.

Building the design

After a couple weeks of back and forth I had in my hands a pdf for the design of The Coropunt, a boat design based off a punt boat and made primarily from a single sheet of Coroplast.

The frame and cross-bars would be made from wood for structural stability but the bulk of the body was simply folded and taped coroplast similar to his other plastic boat designs. The boat was designed to be folded in half to fit in the back of my 2010 Honda Accord and have enough structural load capacity to hold up to 250 pounds (enough for me, my fishing gear, a trolling motor, and any fish I could catch).

The design was relatively simple and just required a lot of folding and taping. The cross-bars were designed to have bolts screw in on the outside of the frame which my housemate desiged a handle for.

Folding and taping the boat, although simple, was not a short process. It took my housemate and I the better half of a day to get it assembled and I still had to cut all the wooden frame pieces which took another good part of a sunday.

Floating the boat

I took the boat down to the lake with life jacket in hand and nervously set it on the water. My housemate followed and brought he boat design he had built (the CPB-2016 - free planes available from https://www.portableboatplans.com) which I had helped him build while my coropunt was still in the design phase. We set out on the water and the boat didn’t immediately sink! The boat was so light that any wind made me rotate and drift but the water line was only a couple inches from the bottom of the boat (which only had 8 inch side walls). I tried to stand up but it was hard to keep my balance, the highest I got was a tall crouch. There are designs for sponsoons (also made from coroplast) that attach to the side of the boat which help stabilize it which I have yet to make.

Mon, 01 Jan 0001 00:00:00 +0000

Restoring a Gibson clone

This Les Paul guitar belonged to my Uncle and afte he passed was stored in my childhood bedrooms closet, at the time I was still in high school and already had a Fender Squier Strat to not practice on so I just left it in my closet. A while back when I started picking up my guitar again I decided to take another look at and visited my mother to pick it up. When I picked it up she also gave me an Ibanez that she had laying aorund as well that needed some reapair. I opened the case and the guitar looked awful. This is the process of restoring this Les Paul and having something other than some old photos to remember my Uncle Tom by.

Assessing the damage

This guitar looked like it had been bashed against the ground at the finale of a Pete Townshend concert.

The neck was loosely attached to the body, three of the four knobs refused to turn, the bridge was rused out to hell, fretboard was slightly lifting off the neck at the nut, the damage goes on.

I knew repairing this would be some work but I also had nothing to lose so I started taking note of every little detail and got to work dis-assembling it.

When I got to the poteniometer knobs it was clear they were rusted out and unusable, I was fairly confident this is why the guitar wouldn’t make a sound if I plugged it in (later confirmed) as there was probably a ton of corrosion breaking electrical contact. Two of the potentiometers were fairly easy to get out with some WD-40 and some effort but two of them wouldnt budge. After a grueling hour of trying to loosen the last two pots I knew i had to take a different rounte. I grabbed my file and started filing down the shaft of one of them with the plan to file down to the nut. This worked great for one of them but the other was so corroded that even when i cutl through the nut it was stuck. I turned to the back and tried to snap the pot off from the inside but I was left with the casing the pot and the corroded contacts still stuck. After half hour of filing i broke through and pulled the pot out. Unfortunately I had to file so closely to the body that I had accidentally scratched the body with the file. All in all, paint can be detailed but I needed that pot out to even start this repair work.

I slowly pulled apart the other pieces of electronics and the pickups and also saw the nut was hardly attached so I carefully removed that as well.

The nut was hardly attached to the nect so I carefully pried it off and set it aside. It looked to made out of bone.

Repair Work

Electronics

If you haven’t guessed from the fact that I had to basically destroy the potentiometers to get them out, the electronics needed full repair.

Luckily there are lots of Les Paul electronics kits out there so I grabbed one and got to work installing it. This had to be one of the easier parts of this restoration as the plans were all clarly detailed online and the parts just fit well into their slots.

Bridge

The bridge on this guitar was rusted out but luckily many people are making Tun-o-matic style bridges, my only fear was that the space wouldnt match up. As it turned out the spacing did match up but with the curvature of the body there was a problem of alignment to deal with. Even though the body was curved the bridge posts need to be parallel to making adjusting the bridge height possible (not to mention just screwing them in properly).

I covered the posts in some quick-set clear epoxy I pushed them into the body, then I screwed in the top post and attached the tail piece and the bridge. My hope was that these pieces would force the posts to set in a parallel position.

I realized the next day that I had forgotten to route and solder the grounding wire to the bridge post. With the epoxy set I thought this was going to be the end of the repair process (buzzing would be very apparent without the grounding wire) but I decided to attempt a rescue mission.

I got a couple ratches and a large washer and put the screw through the washer, through the ratchets and into the post. By steadily turning the screw that goes nto the post while it is braced against the ratches which braced against the body and lifted the post up. The danger I feared would be the post holding its ground and drivign the ratches into the body but by carefully applying pressure I managed to get the post out!

After that routing the grounding wire was fairly trivial (if not just annoying) and I was able to put a new post in soldered correctly.

Pickups

Although I wasn’t sure the DiMarzios that were in the Les Paul were bad I had a set look in mind for the guitar and would either need to get chrome covers for these (and pot them with wax) or get new pickups. I decided to start out with some new pickups since I could always switch back.

Getting an new pickup ring for these was another huge pain since it was a non-standard clone. I measured the distance between the holes

I found a set of chrome pickup rings on eBay that seemed to match and ordered them only to find when they arrived that they were flat bottomed. Instead of continuing the search I decided to take my own approach. I placed the pickup upside down on two pieces of 2x4 with the center suspended and hammered out a rough curve. Through a process of trial and error I mananged to get a curvature into the ring that was very close the curvature of the body (with minimal scuffs on the top of the pickup). Since curving it changes the dimensions slightly I had to screw the ring in with the screws angled, since the pickup ring is changed so infrequently (if ever) I figured this was a minor problem.

Unfortunately the warping I did to the ring also affected how the pickup sat in the ring and it had significant lean forward which I was able to slightly correct by bending the mounting post on the pickups.

Headstock

The headstock didn’t look too bad but I definitely wanted to replace the tuning pegs since the current ones were rusty and hard to turn. I found some tuning pegs that had the same dimensions and hole placement and they just hapened to work out well. While I was waiting for the tuning pegs to arrive I polished and buffed the headstock but that did’t improve it much.

Neck, Frets, and Nut

The neck was another huge part of the repair since it was hardly attached to the body.

I started with focusing on getting the neck looking good before I tried to put it back on, it seemed easy to work with detached. I started by de-rusting the frets which was somewhat difficult due to how low a lot of them were worn down. I taped off the fretboard and took soome 0000 grade steel wool to the frets, scrubbing well. When that was done they were looking surprisingly shiny. For the fretboard I already had rosewood lemon oil to apply to it from my Strat so I applied a good coat (but not too much) and let it sit for a while.

I was able to easily see the freboard slightly pulling off the neck from underneath where the nut would sit. I used this opportunity to put some epoxy into the crevise and clamp the fretboard down to the neck. Although the lifting was slight I wanted to make sure the problem wouldn’t become worse.

With the body-less neck looking clean and healthy (for the most part) it was time to attach it back onto the body.

The screw holes in the neck were stripped out so I thought I had to use wood filler to refill the holes and then re-drill and screw the nect in again. My housemate had another idea: since wood filler is not always reliable in the long term (especially since I wasn’t sure how good it would adhered to the old, cheap wood of the body) he suggested we used a brass threaded inserts. He took some measurements of the screws that originally went into the neck and ordered some threaded inserts on McMaster-Carr along with some coutner-sunk bolts in case we wanted to switch to those.

I took the guitar parts to his work where we drilled out the holes for the threaded inserts into the neck on his mill, this was nerve-wracking as some of the holes came very close to the edge of the neck. Fortnately everything worked out well and with some effort we got all the threaded inserts in. One of them seemed to be a little angled and made a creaking noise when we screwed in the bolt but since the neck was (hopefully) never to be removed again I felt comfortable with it. I walked out of his machine shop with a guitar in one piece (but with still lots of work to do).

I ordered a TUSQ nut with the same approximate string notches as the Les Paul and I sanded it down to match the original nut and glued it into place with a light bit of glue. I may end up going back to the original nut as it looks more vintage.

Back plates, Headstock Cover, and Pick Guard

The back plates, headstock cover, and pick guard were all 4-ply plastic with black front and white underside. Althoug there wasn’t anything truly wrong with them I wanted to see if I could make better ones. My housemate helped me dimension them in Solidworks and the initial plan was to CNC route new ones out of aluminum but that plan fell through when the CNC at my housemates work turned out to not be very reliable.

I looked extensively online for parts that matched these but the dimensions were all all in strange ways. It seemed like these whoever made these plastic pieces weren’t making Gibson clones any longer.

I bought a sheet of 3/32 aluminum (the closest approximate size of the thickness of these parts) and decided to make new parts myself by hand. I super glued the pieces to the aluminum stock and cut the rough shapes out with a jigsaw. Then began a tedious process of filing down the aluminum to the same shape as the plastic. When all was said and done I had pieces that fit well into their original slots. I drilled out the holes for thre screws and used a debur to create counter-sinks for the screws. The last thing I have to do is round the edges.

With the aluminum pieces made I decided they would look nicer with a mirror finish like the neck bolt plate. I ordered some sandpaper with the intention of sanding these pieces to a mirror finish… that plan ended quickly (read: slowly) when I spent hours polishing just on piece and not getting very far. I am still planning to find a place to polish these parts.

Stringing it up

Adjusting the truss, action, and intonation

If you’ve never set up a guitar you may never have had to deal with the truss rod or the action, but these are very important parts of a guitars playability. In short, the truss rod determines how much the neck bends in the center and the action is the height of the strings off the neck (adjusted via the bridge height).

Since the truss rod is buried in the neck it was impossible for me to see if it was corroded (who knows, maybe the guitar neck was submerged in salt water for some time) so I gently adjusted the truss screw with a hex key and found that it did turn, although it creaked with each turn and was very stiff to move. By adjusting the truss and waiting a bit I could tell that it was working, as I could clearly see and measure the neck bowing in and out with my adjustments. This was good news as it meant it was still working.

The bridge height was a little tricky to get down because of the way the posts were angled in the body, it was not easy to adjust just one side and the further out the bridge was the more tension the tail piece applied to the posts. Luckily a good action height was found close to bottoming out on the bridge.

The last step was to set up the intonation which is done at the bridge. This was fairly simple as I’ve done it a few times on my Strat and the process is the same just with a slightly harder to reach screw on the Tun-o-matic style bridge.

The finished product

With the guitar tuned up and putting out sound I was happy with how it ended up. This project ended up being about as much work as I expected (a lot) but I’m happy I put in the effort. It may not be the nicest sounding guitar, nor the highest quality, but it’s a nice momemnto from my Undle and a great learning experience.

Mon, 01 Jan 0001 00:00:00 +0000

Restoring an 1987 Ibanez 540S

My mother gave me this guitar the same time I picked up my late Uncles Les Paul clone. Complete with case and everything. It looks dirty but didn’t seem like anything that was beyond repair (nothing like the Gibson I repaired).

After a lot of digging I determined this to be a 1987 Ibanez 540S which was pretty exciting since it seemd somewhat unique, it was one of the last Ibanez guitars to feature an original Ibanez Edge bridge with Backstop. Up until receiving the Gibson and this guitar the only guitar I owned was a Fender Squire Strat so this was pretty exciting.

Assessing the damage

Generally the guitar was pretty dirty. There was rust on each of the frets and the fretboard looked very dried out. The body had scratches over it and lots of dust/dirt collected. It would need some TLC.

The guitar had three selector switches and a single knob which also had a built in switch actuated by pressing down on it. Two of the three selector switches had their stems broken off and all of them looked extremely rusty.

I plugged it into an amp and tested it out, with most of the toggle switches stuck in place I could only try a few configurations (at this time I did’t even know about the built in toggle in the volume knob). The guitar made some noise when the bottom toggle switch was set in one direction, and no noise in the other.

Before doing anything I made sure to read up on the Floyd Rose lockin temelo system since I had never seen anything like it before in person. I am happy I did because without unscrewing the locking nut you can snap the headstock off if you just cut the strings of like I would for my Strat.

I started by unscrewing the locking nut at the top and loosening the strings and cut them off, the bridge fell flat. I unscrewed the back and took at look at the internals. Apart from some dust/dirt and rusted backplate screws evertyhing seemed pretty good looking. I took the bridge out and examined the pickups. The humbucker pickup ring was bracked in two places but otherwise the humbucker just looked rusty. The single-coil pickups didn’t have any ring and just looked rusty. Unless there was electronics failure I saw no reason why this guitar wouldn’t work just fine, rusty pickups don’t usually cause failure.

Repair Work

Selector Switches

I knew I was going to need to replace the pickup selector switches so my first step was to understand why kind I needed. I took a ratchet and with some elbow-grease got the mounting nuts for all them loose. Pulling the switches out through the back I saw they were pretty standard. I took the calipers and magnifying glass to the and recorded what I could:

5.88mm toggle switch shaft thread diameter
6.2mm hole diameter for threaded shaft
Shaft length 8mm
Toggle switch stem length 11.5mm above end of shaft

Package size 10.5mm x 12mm
Package height 7mm

MTN301
3A AC 250V 8701
Three way switch

I hopped on Digikey but couldn’t find the exact switch so I decided to go with the closest one I could find.

After they arrived I soldered them on in the same manner I saw the originals, hoping these would do the trick.

Ibanez Edge

Since I had no ideahow long this thing had been laying around (from the look of the dirt it was probably ages) I decided it would be a good idea to do a deep clean on the bridge. Luckily for me there are many videos about how to fully disassemble and clean these things and the Ibanez wiki even has the manual for it complete with exploded view.

I fully disassembled it, cleaning each part in a hot bath of water, then with rust remover, back into hot water, drying if off, then spraying WD-40 on it.

While dis-assembling I noticed some of the string-stopers (small square pieces of metal at the bottom that stop strings from poking through the entire bridge) came off when I was scrapingo out dirt from the string stopper hole. In addition one of them was completely missing. I found out that this was a common complaint of owners of these bridges (that the stopppers come off and get lost) so I measured the dimensions of a string stopper than had fallen off and cut out a piece from my aluminum ruler that happened to be very close to the same thickness as the stopper. I sanded the replacement down and I glued all the stopppers in with quick-set clear epoxy.

After re-assembling the bridge I was very happy with the result.

Neck and Frets

The neck was extremely dry and the frets had spots of green rust on them, this didn’t seem like a huge tast to fix so I ordered some 0000 grade steel wool (the finest it comes in) and I taped off the neck and pickups (fine steel wool fragments and strong magnets do not mix well) and got to work scrubbing the rust off. To get the sides of the frets I took a spuger tool and pushed it agaist the wool and up to the sides of the frets. When all was said and done the frets were looking pretty great!

I removed all the tape and applied fretboard oil (which I already had for keeping my Strat happy) to the fretboard. I was lucky that the fretboard was rosewood and I already had rosewood fretboard oil.

Pickups

The pickup screws and springs were all pretty rusty so I figured I would get new ones, in addition the pickup ring for the humbucker would need to be replaced.

Sourcing a new pickup ring for this guitar was a huge pain since it was so old. Ibanez had moved on from supplying parts and I couldn’t find anyone on eBay with it.

I measured the distance for the mouting holes to be 81.5mm width and 38mm height and set out to find a new pickup ring for the humbucker. Eventually I landed on one that fit the bill and ordered it only to find out it was just a single millimeter off! Luckily there was enough play in the system that I could still get the screws in.

The single coil pickups on this guitar were easier to deal with since all I needed for them were new screws and cone springs. Aside fro mthe awkwardness of dealing with mounting it without buckling the springs in a weird way they were pretty simple to put back in.

Getting strung up

The Floyd Rose sure is a pain in the butt to string up, but once done the guitar was looking great (well, as great as a 1987 guitar found in the dumpster can look presumably). I plugged it in and heard noise! I tested every selector switch as well as the potentimeter switch and everything seemed to be working great!

With all said and done I call this project a success, with much less effort than the Gibson clone.

Mon, 01 Jan 0001 00:00:00 +0000

Simulating a Robotic Arm

I have seen these cheap ‘6 DOF’ robot arm kits on ebay that use hobby servos to control the joints and have been curious about them for a while. I ordered one but it turns out that they include the gripper opening and closing as one of the degrees of freedom, thus making this arm truly only ‘5 DOF’ as far as positioning/orienting is concerned. Before writing firmware for it I wanted to explore the various techniques used to control robotic arms, and thus I ventured into the world of Inverse (and Forward) Kinematics

Check out the GitLab Repo here

Forwrad and Inverse Kinematics

Before jumping into the more interesting part I should clarify why Inverse Kinematics gets more attention. Forward Kinematics is the process of finding the pose of the robotic arm (and consequently the position and orientation of the end-effector) given the angles of the revolute joints (joints that rotate) and positions of the prismatic joints (joints that translate). This is actually not too hard given a decent background in trigonometry.

Going the other way, finding what joint angles/positions you should use to given a target position, isn’t as straight forward. This is what Inverse Kinematics attemptst to solve.

As it turns out, it has been proven that if a robotic arm has three joints that have axis of actuation that are parallel then it is possible to find a closed form solution for the inverse Kinematics problem.

Without a neat closed form solution one is left with numerical methods, and this was what I was more interested in.

Jacobian Method

The Jacobian method is a numerical method for honing in on the best angles to form the resultant position and oritentaiton. The only problem is that it requires a series of loops and isn’t guaranteed to converge on a solution (or it may converge but take very very long). It is similar to other numerical approximation methods such as gradient descent.

The Jacobian Method works by calculating the Jacobian of the position and orienation vector with respect to each of the joint paramters (angles and positons). This jacobian can then be applied to the error vector between the current position (found via forward kinematics) and the target position to get incremental changes to the joint parameters which bring the end-effector closer to the target.

The Math

Simulating

I wrote up a simple IPython Notebook which implemented the numerical method and stored the intermediate results. I noticed that although this method worked its convergance was heavily dependent on the starting position/orientation of the arm and the target position/orientation. When disregarding orientation the arm could pretty easily get to the target position but the Inverse Kinematics solver would get stuck trying to balance error in the position with error in the orientation.

To resolve this I implemented a heuristic that adjusted the target orientation after the arm went through a certain number of loops without making meaningful improvement to its position. The orientation would either be adjusted higher or lower depending on the position of the target and the orientation that the arm was getting stuck in. This allowed the arm to dynamically approcah from a new angle when it was getting stuck, trading off the impportance of approaching from the correct angle in favor of actually getting the end effector to the target position. The current heuristic still has some bugs to work out but for many targets it correctly dynamically adjusts.

In addition I implemented a method intelligentStartAngles that took the target positon/orientation and started the numerical descent at a position that was based off the target, this allowed faster calculations and eliminated some non-convergence issues when the arm always starts with angles that cause the arm to point start up.

Mon, 01 Jan 0001 00:00:00 +0000

Homemade American Flatbow

This is a project log for the flatbow that I made. I began the project in March 2017 and had a working bow by the end of the month, though I didn’t add all the finishing touches to it until August 2017.

Why I Built a Bow

I have always had some interest in archery as a child though I had only shot a bow once in my life. My aunt took me to the San Francisco archery range when I was about 10 years old and taught me the very basics on her light pound compound bow - although she also owned a recurve we didn’t use it at all. That one instance of archery is the only time I had shot a bow until the completion of this project.

In February of 2017 I went on a business trip to Albuquerque (where Sandia National Labs has its main facility) and met a colleague who worked in the Labs over there, he was working on a different component of the project I was on. After work he took our california team to one of his favorite resturants and chatted over dinner. During this time he spoke about a few of his hobbies - of which included crafting bows and classical guitars. Note that this man was much older than I and somewhat of a luddite, he did not own a cell phone and lived on the other side of the Sandia Mountains on a decent amount of land. He spoke a lot towards the current state of being always connected and he encouraged everyone to take more time to get in touch with nature.

I went home wondering what exactly it took to craft a bow so out of curiosity I started googling around and found poorfolkbows.com, a website dedicated to crafting bows on a budget written by a man named Sam Harper who does the hobby out of his apartment. I read through his most basic tutorial on crafting a /selfbow from a single piece of red oak and was surprised that it didn’t look very hard at all to craft a simple bow, I don’t know why I decided to persue it but I decided to give it a go.

The Bare Minimum

Although I had decided to build a bow I was definitely wary of the cost of getting sucked into a new hobby. I decided I wanted to try and keep expenses as low as possible. Here is the complete tools and materials list I needed to use to build this bow:

1" x 2" x 7' Red Oak board from Home Depot - Following Sam Harper’s guide to find one with the absolute straightest grain I sorted through about 30 planks and picked out the one with the most straight grain I could find that also had the growth pattern shown on his website. Cost: $10, Home Depot
Shinto Saw Rasp - This rasp was extremely valuable to me and was the main tool I used for forming the bow. It has one coarse side for getting a lot of wood off and a fine side for more detailed work. Admittedly I should have used a different tool - such as hack saw or a draw knife - for the initial forming of the wood but if you have more time than money this tool will do almost everything you need. I heavilty debated getting either this or the Fourrier Rasp (another common tool used for making bows) but decided on this due to the reviews. Cost: $14, Home Depot
Draw Saw - This tool was mainly only used for cutting the 7' plank into a 6' plan (for the main bow) and a 1' plank (for the riser). It was cheap and got the job done and I still use it to this day for small wood cutting tasks. Cost: $8, Harbor Freight
File Set - This file set had everything i need for some detailed filing of the bow and for foming the nocks. It was cheap and got the job done. If you are looking for specific files the small rat tail file is the most important one here for forming the nocks. Cost: $10, Harbor Freight
C Clamps - Needed for clamping when you are sawing and rasping the wood. I got three but only really ever needed two. Cost: $3 each, Harbor Freight
Tightbond III Wood Glue - Used for gluing on the backing as well as the riser I did a lot of comparison between Tightbond II and Tightbond III, in the end I think you can use either but Tightbond III is water-proof whereas Tightbond II is only “water resistant.” I have found many uses for the glue after the project was complete Cost: $10, Amazon
Linen Fabric - This isn’t strictly necessary but it helps reinforce the back of the bow from splintering - as I will explain later. Just got a big sheet of fabric in a natural brown color. Cost: $10, Amazon
4' Metal Ruler - for various Measurements Cost: $5, Harbor Freigt

As well as materials for the bow I also build a “tillering tree” which is necessary for gettign the proper curvature. I will explain more below but for that it required the following:

Spruce 2x4 - Made the basic frame from some cheap 2x4s. Cost: $X, Home Depot
Pulley - Necessary for drawing the bow from a distance to observe the curvature. Cost: $4, Ace Hardware
Bike Hook - Used to secure the pulley to the 2x4, you can do this in any fashion that works. Cost: Had it lying around, but originally from a 6-pack of bike hooks on Amazon, $12
50 ft Paracord - Used to draw the bow from a distange Cost: $5, Amazon
Luggage scale - This is used to measure the weight of the bow when it is drawn out to different distances. Cost: $10, Amazon
Drill - For drilling holes in 1" spacings Cost: Housemate had one
Assorted Nails and Screws - For assemply Cost: ~$5, Ace Hardware

Starting out

As I said I picked up the straightest grain plank from Home Depot I could find with the proper growth pattern. You want to look on the end for large amounts of “dense” growth which is darker than the the porous growth.

Red Oak Plank

As you can see the grain runs down the boad straight from end to end. This will make it stronger and reduce the possibility of splintering when the bow is drawn back.

The first step was to taper the sides in. I followed Sam Harper’s guide marking out where the riser will be and where the side tapering will begin and spent an afternoon rasping away with the coarse side to form the taper. It took a long time but eventually I got one half of the bow tapered.

If you look closely at the second picture you can see the taper has a noticeable kink in it. It is more noticeable on the other side of board. The following picture is the second side but flipped to see the other side of the board.

This will create a point of higher pressure when the bow is drawn so I decided to extend the taper of all sides to go all the way to the riser, this will result in a lower overall draw weight but I did not want to risk the bow snapping from pressure on the kink. I finished up the side tapering of the the second half which took all morning.

Then it was time to start tapering the belly. The belly tapering was all done with the fine side of the rasp because it affects the draw weight of the bow about four times more than the side taper does, thus it is important to get right. I carefully followed Sam Harper’s guide and tapers the belly. The belly tapering took the rest of the afternoon.

After the tapering was complete I used the fine side of the rasp to smooth out the corners so they had a small radius on them, this reduces the strain on the corner of the bow when drawn.

The final part of the roughing out the shape of the bow is called “Trapping” (from the word Trapezoid) which means to radius the belly more than the back so that it very vaguely has a trapezoidal figure (the difference actually isnt super large but it is noticeable), this is because when the bow is drawn the back is in tension and the belly is in compression, if you have the same amount of material on the belly compressing against itself it can cause fractures or create knots (called chrisels) which degrade the structural integrity of the bow.

Note: No tapering or radius-ing should ever be done on the area of the belly that the riser will be glued. The back of the bow is okay to radius, but the belly should be a sharp square.

About 10 total hours of rasping and the rough shape of the bow was complete.

Forming the Riser

Once the tapering was done it was time to form the riser. I first cut a 10 inch riser and made sure it was square, then I glued it on and clamped it down and let it set overnight.

After the glue is dried roughly form the riser by drawing out the shape to rough away. The riser should contact the bow at as little of an angle as possible so that its very thin where the bow limbs beging; effectively the risers edge should be cut to resembl a slow exponential or a sigmoid shape where it contacts the bow (it’s up to you how to shape the other side when its closer to the center)

This is just a photo of it after roughing it out, after this photo was taken I carefully took more material off to make the far ends of the riser narrower.

The reason it must be thin is because this allows the riser to bend with the limbs, if its too thick it will not bend and the riser will pull away from the belly.

Backing the bow

“Backing” a bow refers to putting material on the back (the side facing away from you) in order to increase the strength and prevent splintering. Since the bakc of the bow is in tension it is common for weaker bows or weaker material to pull up an splinter. In additon to backing the bow you can round the edges slightly to prevent splintering but if you round it too much your trapping on the belly may be less effective.

Traditionally bows are backed with animal sinew, since I didn’t happen to have any of that on hand and it seemed a lot more complicated for a first bowyer’s attempts I went with linen fabric soaked in water and glued on with Titebond. Another common backing material is sheetrock tape (also called drywall tape) and silk. Essentially all you need for a good backing material is something that will stretch very little, adhere to the wood, and bend when its drawn.

Forming the nocks

To form the knocks I took the small rat tail file from Harbor Freight and measured about 2 inches down from the tips. I put the file at an angle so the knocks point out towards the back and tip of the bow and i filed in about an 1/8 inch. I rounded out the knocks so it formed smoothly into the belly and back. I was pretty nervous I had gone too far in but it ended up working out, when in doubt you should stop filing early and just try to put the long string on, if you can put tension on and it stays put you’re probably fine.

Tillering Tree

After getting the rough shape I needed to set up my tillering tree. Tillering is the process of progressively drawing the bow out starting from very small draw length and measuring the curvature in order to get even pressure along the length of the bow when it is drawn. A bow with “perfect tiller” will look similar to one half of an ellipse and the pressures on each limb should even out so that there is almost no pressure in in the area where the riser is, this is important for /selfbows where the riser is often glued on so that the bow doesn’t pull away from the riser.

There are many ways to make a tillering tree, the basic concept is a piece of wood with marked increments of 1 inch. You place your bow on top of it and draw it out to a marked increment and observe/measure the curvature:

If one part looks like it is too flat you take wood off the belly from that area to get it to curve.
If one part looks like it has a steep change in (similar to a kink) you take off wood the belly from the surrounding areas so that they will curve more which takes pressure off the kink.

As I mentioned earlier the belly thickness contributes to the overall draw weight of the bow about four times more than the thickness of the sides, thus tillering should be done in a slow and controlled way as to not accidentally take off too much wood. If you do accidentally take off too much you will form a kink in the tiller of the bow and be forced to take off wood from the surrounding area to even the pressure, this will cause significant change in the draw weight of the bow.

Setting up the Tillering Tree

To set up the tillering tree I first screwed a small piece of 2x4 into a full 8' 2x2 to form a tiny L-shape. Then I lashed the long side to a support beam on my balcony using some of the paracord. I screwed a small piece of 2x4 a about eye-level and then to that piece I screwed in another but offset - this formed a “ledge” that i can place the bow on.

Then I used my ruler to mark off draw lengths from the base of the leddge. I started with 6" and marked all the way to 34". Then I drilled a hole in the center of the tillering tree at each marking, this hole will be used to hold a screw so that i can keep the bow drawn without needing to hold it.

At the base of the tree I screwed in my bike hook and put my pulley on it, then tied the luggage scale to the paracord and threaded the paracord through the pulley.

You can’t see the markings and drill holes in this photo because I actually did them after it was set up.

I ended up taking off the second ledge 2x4 piece and replacing it with a thinner piece of wood (a leftover redwood plank used for making fences) so that it was easier to get the bow on and off.

Tillering Tool

An amazingly generous user (Eric Krewson) over at primitivearcher.com sent me a “tillering Gizmo” he created to help get the curvature correct, it’s design is amazingly simple yet effective.

It simply consists of a long trapezoidal piece of wood with a hole going through it. In the hole is placed a threaded insert. The threaded insert allows you to screw a pencil into it and have it poke out at varying depths. Then when the bow is drawn out on the tillering tree you drag the Gizmo along the bow with the long edges will touch the bow creating a in the center if the bow has curvature. If the curvature is not strong enough to create a gap larger than the amount the pencil is sticking out the pencil will mark that section of bow thus the stiff areas of the bow will be marked. Similarly if you have a kink in the bow the areas on each side of the kink will be marked since they will be significantly stiffer. You simply start out with the pencil sticking out a very small amount (to mark the stiffest areas) and as you tiller progressively make the pencil stick out more to tiller until you get the curvature you want.

Tillering “on the long string”

When ready to begin tillerng you are supposed to tiller with a “long string”; i.e. a string that is longer than the final string length so that it doesnt bend the bow much when its on. I wanted to use paracord for this but the folks at Primitive Archer forums told me that although paracord can work as a long stirng ideally you want something that will not stretch, essentially you just want bow string that is longer. Another amazingly generous user (Loon) offered to handmake me a long string out of B-55 Dacron (a common bowstring material) and he even included some hollow string which is used to wrap the ends and center of a bow string for the nocks and arrow nocking.

As you can see one side has a loop called a Flemish Loop, when one gets a bow string often both sides a flemish loops as it builds the structure of the loop into the string so it can’t be undone. Loon sent me a string with one half flemish loop and one half open ended so I can progressively shorten the long string using a (also called an archers hitch) and tiller with different length strings.

Putting on the final string

When i was confidently pulling the bow out so tha the tip-to-belly depth was over 6 inches i put the final string on. The final string was ordered from eBay and has a double flemish loop design so it cannot be adjusted. Tillering on the final string is the same as tillering on the long string. I proceeded very slowly tillering until I could draw the bow out 30" of draw length.

Finishing the bow

I tested the bow at Redwood Bowman in the Joaquin Miller park in Oakland. After happily shooting (very innacurately) with it over 28 targets I had decided the bow worked and all that was needed to do was to finalize it.

I orderd some Linseed Oil and sanded the bow with very fine grit sand paper. I applied several thin coats, letting it dry completely each time. When it was done I was happy with the darkened hue it had taken on that accentuates the woods natural grain pattern.

Other finishes I considered were Tru-Oil and Tung Oil but after looking at comparisons online I decided I like the look of Linseed Oil the best.

Project_raws on Aaron Silman

Bartomaton: The Bar___tending Au___tomaton

Initial Concept

Choosing frameworks

Database Design

System Architecture

Hardware

Coropunt

Building the design

Floating the boat

Restoring a Gibson clone

Assessing the damage

Repair Work

Electronics

Bridge

Pickups

Headstock

Neck, Frets, and Nut

Back plates, Headstock Cover, and Pick Guard

Stringing it up

Adjusting the truss, action, and intonation

The finished product

Restoring an 1987 Ibanez 540S

Assessing the damage

Repair Work

Selector Switches

Ibanez Edge

Neck and Frets

Pickups

Getting strung up

Simulating a Robotic Arm

Forwrad and Inverse Kinematics

Jacobian Method

The Math

Simulating

Homemade American Flatbow

Why I Built a Bow

The Bare Minimum

Starting out

Red Oak Plank

Forming the Riser

Backing the bow

Forming the nocks

Tillering Tree

Setting up the Tillering Tree

Tillering Tool

Tillering “on the long string”

Putting on the final string

Finishing the bow

Bartomaton: The Bar_tending Au_tomaton