When we first received our PCBs we discovered they had problems, we began by plugging them into an Envision current source, the current source is made to not short under any circumstances so it's fairly safe to connect however in hindsight we should first run extensive testing on our PCBs to make sure they would work when plugged in. We noticed our PCB wouldn't accept 24 volts and instead only allows 1.5 volts at .5 amps, we were immediately suspicious of a short. A few minutes later we read somewhere that it could possibly be a firmware issue, where the TMC269C LA is allowing current to flow through small resistances before it is configured, we continued trying and finally were able to find the short by applying an IR gun to measure the high current flowing in the CMOS sections of the circuit. From there we discovered the sources of the CMOS were switched causing them to always be drawing current. We had backup drivers so Purab and Syler continued working on the tutorial while Aleksandar attempted to solve the problem, after one night we realized it was hopeless. The solution was to raise the legs of the CMOS's sources, then directly onto the board solder wires from the pads to each source, this proved extremely difficult and we realized we would spend 20+ hours on a PCB which might have more errors. After applying the technique above to a PCB we found that either our TMCs had died or there existed another problem in the circuit, from here we entirely scrapped the idea of fixing the PCBs and moved on to our backup drivers.

To the left we can see an original arm design we had come up with, this arm involved 3D printed axles which worked extremely well on the small scale with a rope belt system to move the linkages, when applying it to the large scale arm we found quite minimal friction at first. When correctly tolerancing the arm so the belt was tensioned well, we found the 3D-printed axles caused huge friction, this friction made it impossible to move the arm, so we had to switch. Aleksandar spent the Thanksgiving week and weekend on Onshape to reconstruct a working system, this involved a new base, arm, belt system, and axles. When constructing a reasonably sized base, Aleksandar discovered that fitting a 3rd stepper onto the base is extremely difficult, considering it must be near the linkages to minimize the tension differential of the string when the arm rotates. Realizing there was no time to work as a middleman for Syler to reconstruct a gripper he began constructing a gripper as well, using gears, a DC motor, and Bluetooth connection to a dev board, it was all able to come together in the last week to move and hold objects.

Gripper flaws, in the original design by Syler on the gripper the force was applied to a linear spring-driven system which when pulled would enact a force on 2 other axles to close the arm (to the left), one problem with this encountered was that the force applied to grip the object is proportional to sin(), this means when the gripper is nearly fully closed it actually has the least force acting to grip, theoretically if the gripper could be entirely closed such that the angle between the axle and the grippers are entirely parallel there would be 0 force acting to hold the object. In Aleksandar's design, he moved to a gear system that could then apply linear and equal force throughout the range of motion of the gripper.