this week the assignment was to design something big with bonus credit if we don't need to used fasteners or glue.

i've been waiting for this week because i still need a desk in my bedroom. i started by mocking up a desk in cad by defining parameters, creating the outline of one leg, and then copying it over and removing the shape of the interlocking tabs so that we are left with two legs of the same shape and size with complementary cutouts.

here we had to be careful to define the extra space for the tool, called dog-bone cutouts. i parameterized this since we aren't sure exactly what tool we might be cutting with so that it can be dynamically changed later without messing up the geometry of the design.

i assembled the table in cad and then i realized that i just didn't like how visible the dogbones were in the final design and i also didn't like how the surface of the table would need dogbones too if my plan were to assemble the legs as they look in the renderings below.

i wanted to challenge myself to make a design that hides or minizes the cutouts, so i started over with some hand sketching interlocking concepts that reduce the number of interlocking joints required, and/or better hide the ones that were required.

some of these ideas included having access to a router table that i could set to the width of the table and cutting around the edges of the tabs, and i just don't think we're equipped to do this at the harvard reef, so i went with a slightly different approach and mocked it up in fusion360.

i made sure to parameterize every value and then left these sketches 'messy' (including the non-utilized parts of the circles) because it would allow me to change any of these dimensions later (based on how big the cnc and/or material and/or space i want to fit the desk into, is). after i did drawings of both leg 'types' i tested out the parameters by changing all of them sequentially to make sure that the geometry could be shifted without breaking, and it worked!

then i copied each leg twice, joined them, and started to design the top.

i wanted to get rid of the issue in my first design, so i decided to simply route out half the width of the material so that it sits snuggly on top of the legs without exposing the legs on the desk surface.

i did this by placing the table top directly in the middle of the legs (this is also parameterized, so that if the width or length of the table changes that the top will still sit directly in the center), and then i extruded the top of all the leg pieces up by the length of 1/2 * width_of_material. this way, i can ensure that this cad design will work for any tool diameter, any width of material, and any size desk! it should be 100% scaleable.

the last thing is to make the final dogteeth on the table top surface (underneath) and add a little bit of tolerance. since we are working with the cheapest material (OSB) i didn't want to assume perfection so i added a small gap so that we can still fit the table top on with some minor imperfections (warping or chipping) on the sides of the OSB.

when we combine these parts, we have a table i'm reasonably pleased with!

finally after several weeks, the cnc at the reef was ready to be used. i booked a session with christine to learn about the machine and run my files, but she informed us that it was still buggy and could only take svgs, so i converted my files to svg.

the machine was set up with carbide create, a 1/8" double flute upcut bit, and the osb we used measured in at 0.61". i imported my svgs to the carbide create software, and selected the files, and calculated speeds and feeds. there is a helpful guide for the shopbot machine here.

we ended up using a pass depth of 0.625", a spindle speed of 18000 rpm, and a feed rate of 1.125"/second.

after a first pass, that took around an hour and a half, my 4x4' sheet was not cut all the way through for most of the sheet. it looked like something was uneven because there were cutaways around the outides of the bed but not the center, like the table was sagging in the middle.

you can see in the photo below

you can see that in the photo on the left that the cuts were made through on the top right and bottom left, but not in the middle. we tried again with a little extra depth (an extra .2"), but it still didn't cut fully through. we took it off and started measuring the table and testing to see what was off.

we found that the center of the bed was sagging almost 0.5" lower than the outsides, which we found almost unbelievable. we built some props for the cnc to prop up the center.

by this time my time with christine had expired, we had spent almost 5 hours working on the machine. other students who followed had cnc cuts that worked, but leo suggested that i just document the narrative for this week, and spend the rest of the time on my final project!