Skip to content

Latest commit

 

History

History
94 lines (78 loc) · 5.47 KB

todo.md

File metadata and controls

94 lines (78 loc) · 5.47 KB

DualPanto issues and plans

Unreliable Instances

These are the issues that we observed in class last year and do not know the cause and how to fix (in decreasing severity):

  • disconnects (sometimes) after pushing hard into an obstacle
    • pcb level, sounds like a power surge issues @Rob
  • disconnects (sometimes) after turning on the power
    • pcb level, sounds like a power surge issues @Rob'
  • clipping through a wall when pushing downwards for ~2 seconds
    • god object ends up outside the corridor
  • obstacles do not render when there are too many obstacles
    • firmware or unity framework
  • having moving obstacles that enclose the godObject and pushing it makes it unreliable
    • either unity or firmware – no clue where exactly to look
  • “sliding” sensation when touching corner of a cube
    • physics part of the firmware @Rob?
  • “switch to” didn't work in some cases (might be an issue with student's unity implementation)
    • unity framework (maybe fixed by Martin, need to check)
  • some computers cannot detect/communicate with dualPanto, while the same dualPanto works with other computers
    • USB drivers do not work : how many computers (percentage) where affected by this?
  • obstacles don't clear up after switching levels
    • unity framework, is this implemented?
  • the firmware should be uploaded from the same computer running Unity. Otherwise, a user feels drag (feels like putting finger in honey) after touching a wall or force field
    • better way to solve?
  • movingObstacle in unity: when the obstacle moves in the direction of the handle and goes into it, the handle gets "trapped" in the obstacle, i.e. you feel the inside of the obstacle; and when it moves away, you're free again
  • does not connect reliably (when this happens, 40-80% of the times when Unity starts, dualPanto fails to connect)
    • tackled by Martin (checked on Mac, need to check on Windows)
  • syncing issue between dualpanto and unity sometimes when a user need to input COM port to pop up window inside of app
  • glitch
    • sliding wall when interacting with diagnal line
    • passing through wall
  • encoder value sometimes randomly increase nearly one frame (need to check it's one frame or not first)
  • don't actuate (1)when low battery, (2) out of working region
  • switchTo is sometime slow. Shohei feels it's related to coummunication layer, because when I run switchTo, dualpanto is disconect milisecond

The Faulty PCBs

We also had 11 broken PCBs and these are their current states:

  • cannot upload firmware and does not send heartbeat through serial (1/11)
  • firmware uploads, but does not send the heartbeat correctly (doesn't send “start” msg, subsequently no useful connection) (2/11)
  • firmware uploads, with other connection issues (1/11)
  • encoder not tracking correctly (drift), weak actuation, two pantographs acting significantly differently … etc (5/11)
  • blown ic chip (1/11)
  • USB port broke off (1/11)

More possible DFMA points

There are also some possible DFMA points to improve on, which will help us avoid issues with the hardware and fix them more easily

  • LED to indicate the state of the firmware/ESP32 @Rob
  • easier way to mount and solder the motors (now it requires many steps and soldering positions that are hard to reach; also, replacing motors requires everything to be taken out) @Rob
  • address the polarity of the motors (maybe use DIP switches or modify the motor to be foolproof)
  • replacing/fixing the linkages requires the whole device to be taken out of the casing
  • linkage bearing assembly: current design needs tweaking after assembly to act smoothly; it might have also led to mechanical issues with the haptics
  • ribbon cables connecting to the handles have a pair of flipped wires due to different versions of the geared motors @Rob
  • better way to zero the position of the handles; current process: move handles back to original position and then pressing reset button
  • calibrate handle orientation: Patrick proposes adding gamified calibration routines that appear to be part of the (shooting) game

Debugging Tools

  1. More debuggable PCB design (in v5, each component can be swapped and tested)
  • separate esp32 part with the other, resulting in 3 PCBs @Rob
  1. Firmware to test PCB (testing both with and without battery) to check if
  • firmware uploads
  • esp32 turns on
  • motor rotates
  • encoder reads properly (maybe a jig to rotate motor exactly 1 revolution)
  • motor rpm is in a reasonable region
  • generate forcefield
  • create obstacle
  • move handle (“switch to” function)
  • make one handle follow the other
  • more?
  1. Python script to test if the USB communication can
  • receive heartbeat from Dualpanto (already done by Martin; found bugs and need to debug firmware)
  • receive handle positions received successfully
  • send message to move handles
  • send message to place obstacles
  • more?

  1. A demo scene in Unity to test all the basic functions

  2. Maybe a dualPanto emulator to check the packets that are sent from Unity for further debugging Unity.

P.s The ESP chip (ESP32-WROOM-32(M103QH2800PH3Q0)) we were using is now at the “not recommended for new design” stage, and not regularly stocked by Mouser/Digikey

P.p.s we may want to add an LED on the PCB to help with testing/debugging/validation. Moreover, might we need test jigs?

Other non-device related things

  • The BPM team had audio queue up… is there a way to clear out the audio queue?
  • Voice over should maybe be in German, so Silya Korn can understand?