The goal of this project was to work with my partner, Hanna, and build a model of a hand-powered windlass for a well. We were given several constraints, including:
- A fixed well diameter of 12 cm
- A crankshift that could not be located over the well
- A minimum height for the bottle to reach
- A time limit for elevating the bottle to the minimum height
- A cap on the Delrin sheet, rod, and string we could use in our final design.
When we were first sketching out designs for our windlass, we took all five of the constraints into account.
We started by measuring the height that the bottle needed to rise to, and we found that to be 13 cm above the table. Giving ourselves a 2 cm allowance, we decided that our windlass needed to be 15 cm tall. This gave us the height we needed to place the axel at and the height for our triangular mounts.
There was some debate over whether our mounts should be 45-90-45 triangles or 60-60-60 triangles, but we finally decided on the 60-60-60 to save material. Using some trig, we found the inner dimensions of the triangle to be 17 cm, and decided to stick with our 2 cm allowance trend and make the mounts 2 cm thick. Each side of the triangle is 21 cm long, and each was fitted with two pegs, one at each end. These dimensions determined the length of the base plate.
In response to the crankshift placement constraint, we decided that our windlass would have a crank that protruded forwards as opposed to out to the side, so both triangular mounts would rest on both sides of the table and span the gap, as opposed to one mount on either side.
All that was left to do was decide the width of the base plate. We knew that we wanted to use the Delrin rod as an axel and part of the handle, and we decided that we would use it as horizontal struts that ran parallel to the well gap and perpendicular to the mounts. Two of these struts seemed like enough, so we divided the rod roughly 55% for the two struts and 45% for the axel and handle. This resulted in a base plate width of around 15 cm, the length of the horizontal struts. Now that the length and distance between the mounts was finalized, we could give the base plate a width and height, leaving room for the slots.
The remaining Delrin was used for making a spindle for the axel, a handle piece with a wider radius to increase torque, and bushings to fix the axel and handle in place. The spindle was designed to be as close to square as possible, and it was left open on two sides to save material.
Foam Core Model
With our design concept finalized, we needed specific dimensions for the pegs, slots, and bushings, so we created a test plate of several slots/bushings and several pegs. We fixed the peg width at 2 cm, but changed the depth, and we made several slots that were all .02 cm different in length and width. Here are our ideal dimensions:
Press fit: peg = 2 x 0.52 cm; slot = 91.98 x 0.48 cm
Press bushing: diameter = 0.62 cm
Loose bushing: diameter = 0.64 cm
The joint between the mounts and the base plate were all press fit slots/pegs because we wanted the object to be rigid and perpendicular, but we didn’t want the button from thermal press (we wanted it to sit flat on the table). The joints holding together the spindle are press fit slots/pegs for the same reasons. The joint between the horizontal struts and the mount were press fit bushings because we wanted the structure to be rigid. We had considered thermal pressing them once we finalized the design, but that ended up being unnecessary. The axel sits in a loose fit bushing so it can freely rotate within the bushing. The spindle is fixed to the axel with (very tight!) press fit bushings so it would rotate as one object. The joint between the handle is also press fit for the same reason, but we ended up thermal pressing where the handle met the axel to ensure that it wouldn’t spin out.
After we modified the dimensions slightly with our new bushings and slots/pegs, we cut our windlass, assembled it, and tested it out our first iteration!
Though the windlass worked as we had expected it to (~4 seconds to pull up the water bottle), we noticed several features that could have been improved. First of all, the axel had a tendency to pull out of the far slot, so we needed a press fit bushing that was bulky enough to keep it in place. We solved this by making a fairy shaped bushing, since wells reminded us of wishing wells and fairies. Also, the windlass needed to be supported with one hand and cranked with the other, which wasn’t ideal. We noticed that most of the shifting/instability was side to side, so we thought that if we braced the windlass against the table, it wouldn’t need support from a second hand. The 12 cm gap actually worked to our advantage here, so we could fix the distance the four braces were from each other. We cut a base plate with four additional slots, as well as four braces, and thermal pressed the pieces together. The size of the pegs and slots are directly copied from the Strong Joints Toolbox. Also, the order in which they were pressed is abundantly clear.
(versus) 
This is the windlass we successfully demonstrated on Friday! The new bushing served its purpose well (we added one to the handle too), and it was able to be cranked rather quickly with only one hand.
If I had more time, I would re-cut the base plate. Our second base plate ran into some issues with kerf, because the laser cutter shifted it at one point, so one of our mount pegs required some filing and eventually coaxing with a mallet from Larry, and another one was rather loose. I would also consider different spindle shapes, perhaps something completely square or maybe even triangular or a more complicated polygonal prism that was closer to circular. I also liked how one group used two strings and two axels to distribute the weight, but I believe I would require more string and definitely more rod to accomplish that. Besides those changes, through trial and error, I would try and see if I could reduce the amount of material we used without compromising the windlass’ structural integrity. Still, I am overall proud of the windlass’s performance and aesthetic.
A formal count of the material used puts us at 490.5 cm^2, with 9.5 cm^2 to spare. We used all 50 cm of the Delrin rod as well.
