2016-08-29 Electrical Meeting Notes

Owned by Karl Ding
Last updated: Sept 20, 2018

Attendees

Electrical Debrief from MSXI ASC 2016

  • Our boards were not very easy to debug at the race, because each board was rather specific and had certain nuances
    • Goal is to unify our hardware and inventory
    • We spent a lot of time debugging specific issues on boards
  • We had no documents describing the connections between boards, making it difficult for members to get a full understanding of how the system works. Altium allows you to create wiring diagrams - we really should be doing this in the future - Ryan

Themis

  • Difficult to build because of all the crimps required
  • Moving LCD display to a panel, like an airplane
    • Might simplify the layout and design

Batteries

  • Building the batteries was difficult and a pain
  • Drew too much power from the auxiliary battery, which caused it to die really quickly at ASC
  • Possibly having a slightly more granular power distribution system

Kill Switch

  • Kill switch should not be software. Make kill-switch in hardware

Connectors

  • Unnecessary to use M12 connectors everywhere. Move to thinner more flexible can connectors and split to M12 when exiting a box

Enclosures

  • Battery box should have most boards in it except for lights, etc
    • At the very least, contain all HV in one enclosure - we were penalized on this - Ryan
  • Minimize external connections ("spaghetti wiring")

Modularization

  • Modularize our boards, and determining the responsibilities of each board
    • This makes wiring the car easier
    • MSXI was a struggle to wire

Other

  • Change our footprint library for electrical components
    • Had problems with solder mask clearance

GitHub

  • Move hardware to GitHub instead of SVN
    • You can treat GitHub as a SVN repository, and still keep Altium's SVN integration
  • What is the plan for mechanical?
    • Are they planning on using a different source control?

Connector Board

  • Board used to test soldering and how well connectors connect
  • Connect to an e-load to test how efficient it is
    • Basically build an eval board for testing connectors, regulators, and how well it is made

Controller Board

  • Moving to a STM Cortex 
    • Cheaper than MSP430
    • Integrated CAN controller
    • Has hardware MULT/DIV

Mechanical Update

  • By next month, mechanical will have a high level design of the car
  • In that time, they will identify projects and form a small team of new recruits 
    • We should to do something similar for electronics
  • They will identify projects that we can work on to "practice" things on MSXI to get it working

Electrical Projects

  • Have new members create eval boards that does routing for that component
  • We have a lot of members who are eager to do work (and hardware stuff) but no real projects
    • How-to-solder does not really cut it
    • There needs to be closer mentorship and individual one-on-one interactions with new members
  • Most of electrical is now off-stream with Mechanical
    • We need to recruit 8-stream students
    • We need to find someone to take over the off-stream after I leave - Ryan
  • People need to be able to design boards that are actually buildable, which is why knowing how to solder is important as well (see Plutus' crystal)

Timeline

  • Goal is to use MSXI to test things (we will not try to fix things to get MSXI to work properly)
  • We have an eval board that Titus is playing around with right now
  • Controller board should be pretty simple (it essentially is an Arduino, but for our form factor)
    • Kevin will own the controller board (ELEC-1)
    • We will aim to have the first revision prototype in our hands by end of September
    • By next weekend (September 9th), we will have the board laid out and design finalized and ordered
    • 1 week after that, we will receive the board

Battery Pack Balancing

  • ICs that include cell-balancing functionality are expensive
  • If we build a pack that is actually matched, we don't REALLY need the cell-balancing
    • We replace the pack after each race anyways, so it's not a big deal
    • Ensure that we actually attempt to match capacities of each individual cell
      • Our pack was unmatched af

Things to do on Confluence

  • Write articles on how to set up the toolchain for the STM processors
    • We will start people on the MSP430, because it is easier to modify registers
    • Ensure that setup on Windows, OSX, Linux is covered
  • Document the car, so that new members have things to look at
    • document external components, how the boards fit together

Work

  • Does Ryan have time to work on things?
    • Probably, as long as it's something I want to do I can make time - Ryan


  • Controller boards will be in our hands preferably by the 16th.
  • Connector board sent out by Wednesday
  • We need to figure out solar things
    • How is mechanical doing aerobody without solar panel designs?
    • We should get smaller modules so we don't get wrecked by compound curves like MSXI

  • Purely mechanical battery will be unreliable and a lot of work. Use a mix of spot welding and connectors
  • bigger fans for battery pack. We need to seriously consider cooling this time
  • We need to charge and discharge batteries and get accurate capacity.We need to build a pack that is matched
    • To pre balance cells, connect them all in parallel and then build the pack in series.
  • Need to integrate BPS testing into Plutus