> DATE: 2026-02-28
Finished the first full version of the structure holding everything together. This is the version I will initially print with PLA and PETG. The goal here isn't to use this in the final bicycle (though that would be nice), but to essentially push it to its limits to see what parts need to be reiterated upon and what parts can actually be used in the final version.
There are quite a few "interesting" design choices made and honestly I am not too happy with them. They are "interesting" because they have a lot of complexity that is often necessary - a few exampls:
- (Image 1) I am having quite a bit of difficulty actually measuring the angle of the longest bike tube, and my previous attempt to mount an end of the flywheel shaft to a fixed piece has failed. I am therefore trying a shaft mount with an adjustable angle for now, but this ~could~ introduce a lot of stress to a few small parts, but it may also not, as the shaft should (theoretically) not actually experience that much force - it should mainly come from the flywheel's mass and be offset by the pulley tension (which is in the opposite direction).
- (Image 2) I have opted to try mounting all the electronics at the top despite the fact that this will increase the center of height and make angle recovery harder. This is due to a practical reason, that at the currents I am anticipating to need, longer wires that stretch from the bottom to the top of the bike could literally melt. Also something about the LiPo battery dropping from that height and making this project into a self-balancing IED. A benefit here is that all other wires (IMU -> ESP32 -> VESC) are shorter as a consequence (the IMU should be as high as possible on the bicycle to measure the angle as accurately as possible, so it's also at the top) and I am dealing with an application so sensitive that having longer wires could measurably introduce lag to the control system and make it significantly harder to operate.
My yee yee ass 3D printer also seems to have lately been hyperfocused on ragebaiting me, with every single (PLA) print I attempt failing due to one of two reasons. At first, it seems that prints' lower solid layers were successful and they only detached from the base when starting infill. This is probably becuase I tried experimenting with one of the "ideal" infill patterns, Cubic on PrusaSlicer, which has intersecting paths in layers.
After that, I tried switching to Gyroid (which should be close to isotropic) and doesn't have this, but I am now seeing bed adhesion problems from the first layer despite leveling it. I will try another print later today with Gyroid after cleaning the bed and leveling it again.
-----------------------------------------
> DATE: 2026-02-22
Starting this log (and whole website) today. The current state of the project is:
I have soldered everything together that needs to be soldered (Lipo -> XT-90S -> VESC -> BLDC motor). It seems to work fine for now, but I have only been able to test it at low currents (5 amps for now, theoretical limit is 100) because I do not have a mount and I do not want to hold an unloaded motor with 2.2 kW of power. I have also managed to connect an ESP32 to the VESC which is now able to read from it and write commands to it using the VescUart library.
The structure meant to hold everything together is technically ready to print in CAD, though the goal here is really to just break it as soon as possible so I can reiterate on it.
For context, the parts I am currently using are:
Control: ESP32
VESC: Flipsky 75100
Motor: Surpass Hobby C6354 BLDC (250 KV)
Battery: Spektrum G2 6S 4000mAh 50C LiPo
Bicycle: Giant Peloton (unknown model)
As for the physical structure I am planning to 3D print as much as possible, most likely with PETG (at the moment). I will try to lathe the reaction wheel into existence but this strongly depends on whether my uni will allow me to use their equipment. If not, I will order it as laser cut parts but this would be fairly expensive.
-----------------------------------------