Update on mechanical architecture

This past few days, Vlaya, Xavier and I also worked on the mechanical architecture of our Phyllo. We’ve already made several posts on this subject, but we still have doubts and questions. Your opinions and experiences are much welcome.

General architecture

The Phyllo (the printed shell of the petals) rests on a circular PVC plate of 25cm in diameter. The main PCB is attached on the top of this PVC plate. The PVC plate is fixed to the axis by a metal piece that the school mechanic will machine for us.

Then, going down, we find the two ball bearings carrying VCC. They are isolated from the axis by a PVC tube and from the frame by an 3D-printed insulating box. The role of this PVC cage is only to isolate the ball bearings from the frame, and to fix their external part, so it doesn’t need to be too thick. Those ball bearings aren’t the ones guiding the axis, this is the role of the bottom ball bearings. We plan to use two copper tubes (one inside the ball bearings and the other outside) to be able to solder the wires more easily.

Then, we have the top aluminum plate, which will be square-shaped, with a hole to let the wire carrying VCC5 through. Under this plate, we have the two ball bearings carrying the mass, also surrounded by a cage of copper (or another metal) to solder more easily.

Finally, the axis is fixed to the motor, itself fixed to the hexagonal aluminum plate at the bottom.

Here is a diagram to clarify:

As Xavier explained in his post, we have some troubles with the motors with integrated BLHeli ESC, so we don’t know exactly the distance between the two aluminum plates yet. 

In addition, on some of the RoseAce diagrams on which we relied, only one ball bearing was used for the transmission of VCC. For aesthetic reasons, we would like to minimize the space between the fixed base and the rotating PVC plate, so it would be great if we could use only one ball bearing too. Any opinion on the reliability of the VCC5 transmission through a single ball bearing is welcome 🙂

Ground and power transmission

The ball bearing system allows ground and power to reach the level of the PVC plate, but we hesitate on the best way to bring the GND and VCC to the main PCB. Here is a diagram recalling the distribution of GND and power:


For VCC, we have little choice but to pass a wire through the PVC plate and solder it on a PAD of the PCB.

However, we have several alternatives :

  • we put the main_board with the wire already soldered on it and we struggle to solder the wire to the copper around the ball bearings,
  • we solder the wire to the ball bearings as long as there is no PVC and we struggle to solder it to the PADs on main_board,
  • we solder the wire on both sides separately and we have a big loop of wire when we assemble the PVC on the axis,
  • we solder the wire on both sides separately and put a connector to avoid the big loop, but it adds weight which may weaken the connection…

In short, we do not find an ideal solution… If you have ideas, do not hesitate to comment 🙂


For GND, we have 3 main options:

  • Press the PCB against the PVC and transmit GND by screws.

The big drawback of this configuration is that the PCB is pressed against the PVC, so we have either to put all the components on the top layer of the PCB, or to drill holes in the PVC to be able to accommodate all the resistors and capacitors that are on the bottom layer (and holes for components that must protrude, such as  IR receivers and IR LEDs).

We could also use a single screw in the center of the PCB, in order to limit the number of holes in the PCB.

  • Extend the axis to the PCB and use it to transmit GND directly.

The idea behind this configuration is to raise the PCB (with a PVC belt or some PVC pads for example) so that the components can fit between the PCB and the PVC plate without having to drill holes (except for LEDs and IR receivers). 

GND transmission would be done by extending the axis through the PCB, to connect GND with a copper block surrounding the axis directly on the PCB.

However, we don’t know if it would make a stable connection. Another disadvantage: we have a big hole in the middle of the PCB, which complicates the placement of the components (we would like to put the large storage capacitors as close as possible to the center).

  • Use a (or several) banana plug(s) to transmit GND.

This is our favored option for now. The principle is the same as for the previous option using banana plugs rather than extending the axis. The advantage is that we no longer have large holes in the middle of the PCB, but one (or more if needed) smaller elsewhere (but still in a 5cm diameter disc in the center of the PCB).

However, we don’t know if the connection will be good enough, nor how many banana plugs to use. 

Any opinion or advice would be very helpful 🙂 Other posts on how we plan to fix and connect the other PCBs will arrive soon.

Leave a Reply

Your email address will not be published. Required fields are marked *