diy centrifuge for DNA extraction

bill of parts

build log

• look here for notes that i took while building the project (e.g. measurements, reference pics, etc.)

basic knowledge

• batteries
  • LiPo batteries typically listed as 1S, 2S, 3S, etc. each S represents a battery cell, which corresponds to more voltage:
    • $\mathrm{1\ S} = \pu{3.7V}$
    • $\mathrm{2\ S} = 2 \times \pu{3.7V} = \pu{7.4V}$
  • LiPo batteries also have a discharge rating (“C” rating). the discharge rating describes the maximum number of amps that can be safely pushed from the battery

cad

• the centrifuge should be able to spin these 1.5 mL Eppendorf tubes.

board

• i had to mill the board five times to complete the final project: what a time...
  • the first board i made was very tightly packed (even following the fab.dru design rules) and resulted in a board with weakly adhered traces. likely, i also weakened the traces during the soldering process; this resulted in me ripping the traces off when i tried to disconnect the JTAG programmer cable from the board
  • the second board had a major design oversight: somehow, i didn't realize that my USB header was flipped, and so the 5V and ground traces were flipped. this meant that as soon
    • i tried to cut traces and wire jumpers, but because of the way i had laid out the board, this turned out to be pretty much impossible, and i opted to redesign the board.
  • the third board was correct, but the 1/64 endmill had become dull by the time i was milling, and broke during the milling process
  • the fourth board worked — annoyingly, i also ripped a trace off when i was trying to clean up the board of stray traces to prevent shorts, which made me have to remill the board
  • finally, i got a beautifully milled board (the z-height was still a little too low, making the traces almost engraved rather than a thin layer), that i soldered correctly without problems

ESC motor

• calibrating the ESC motor. probably the most annoying yet also satisfying moments was getting the ESC motor calibrated and running.
  • the calibration protocol (standard for most ESCs, and was also described on the Amazon vendor's page)
    • power on with uC sending high pulses, wait until beeps
    • turn off
    • switch uC to send low pulses
    • turn on
    • wait for beeps
    • turn off
    • next time the motor is powered on, it will be calibrated
  • the first issue was sending the correct signal to the motor. at first i was thinking that the calibration protocol i was using was incorrect, but it turned out that wasn't it
    • in a streak of luck, i realized that the ESC probably wasn't even receiving the pulses i thought i had programmed the uC to send
    • debugging using the oscilloscope confirmed this; pin 5 that i was using to send the PWM signal to the ESC was not outputting high voltages.
    • after cutting traces, wiring jumps, and linking pin 16 to the ESC, the calibration protocol worked, and i was able to get the motor to run.
• i was also wondering where the power (5V) came from in the ESC demo board (components); i think it's not actually a used connection and that voltage comes from the external source that powers the motor.

the case

• designing the case was the most interesting part of this week, since i hadn't done any integration in the previous weeks before
  • i was a little annoyed because i had designed a case in Fusion at home, but forgot to the save the project (i didn't lose the CAD, it just didn't upload to the cloud), and i had to re-CAD it when i got to campus
  • instead of re-CADing the whole thing though, i just remade the geometry of the faces, and exported them to DXF to be cut on the laser cutter
  • acrylic has very good tolerances with respect to kerf, and so the measurements that i
• i actually made a mistake in the size of the mounting holes for the corner braces that would connect the acrylic sides together, but it was okay because we had bolts and nuts of a larger diameter that could fit into the holes
• 3d printing the corner braces actually took way too long, but i was very pleased with the results

the end result

files

• case
  • cover.dxf
  • base.dxf
  • long-side-gap.dxf
  • long-side.dxf
  • short-side.dxf
  • corner_brace.obj
  • side_brace.obj
• rotor
  • rotor.obj
• code
  • centrifuge.ino
  • interface.py