2020-03-15 Meeting notes - Board Review

Date

Mar 15, 2020

Participants

  • @Aashmika Mali

  • @Liam Hawkins

  • @Pei Liang Guo

  • @Micah Black

  • @Josh Rong

  • @Ricky H

  • @Taiping Li

Goals

  • board review

  • general discussion

  •  

Discussion topics

Item

Presenter

Notes

Item

Presenter

Notes

General

 

  • include block diagrams for each board in altium

    • much easier to read

  • controller board changes maybe

    • second can bus

    • change mcu to be programmed through CAN

  • for telemetry, to make sure SD does not erase on turn off

    • PI should be shutting down before switching to aux

    • we are using the enable pins off of the 5V supplies to turn off pi at will

    • maybe we want a GPIO for shutting off the pi instead

  • for logging

    • maybe use more stm32s for file logging in SPI instead of pis to save on power consumption

    • esp32 maybe for logging and sharing of data to passengers

  • telemetry

    • CAN to ethernet using tritium

    • CAN dongle uses ~100mA

    • router uses a lot ^

    • ubiquitis seem to be pretty reliable ← not power efficient

    • XPs should work, but we ran into some issues ← they should work and take less power consumption

  • controller board issues might be because of

    • plugging them in when on

    • not using ESD protection

    • throwing them

    • seemingly did not have issues before the old firmware team

BMS Current Sense

Liam

  • added new test points for signals and vref

  • 20mA current through resistor across both terminals on op amps. was from previous rev

  • is this an issue for having a passive connect both terminals of the op amp? was from previous rev

  • is the RC correct values/ is the resistor supposed to be a smaller resistor? was from previous rev

  • firmware still has not been to validate readings from ADC, so further issues may arise

  • make the spi test points and gnd test point through holes

Lights Driver

Pei

  • add link to design on schematics

  • buck converter can work at 13.5V and drop dynamically

  • check if led driver can support the current for leds in series and parallel

    • can support 2A output, so probably 4 led in parallel

    • maybe have 2s2p leds

  • tent all the vias to mitigate shorting possibilities

  • have the input caps closer to input on buck

  • add another cap footprint on input and stuff if see more ripple on input

LED Board

Pei

  • for drl currently

  • 4leds in series

  • zeners provide alternate path for rest of leds in case of fault

  • fix the footprint for the connector, so the pegs line up with holes

  • look at distancing between leds for proper diffusing in our setup

    • test this out with piece of acrylic

    •  

Rules

Pei

  • found all rules for jlc pcb

  • import through altium, by selecting all parameters in rules

  • make rule set for 4 layers vs 2 layers

  • change to 0.3mm for via size/ largest minimum for 4layer vs 2layers so no mistakes when sending

  • make more rules for general use (i.e. text size, polygon connect, etc.)

Power Select

Josh

  • switch fuse location to right after connector

  • change layout to 4 layers?

  • change cap order to U7 to go from vbat

  • fine tune r19 and c21 when the board comes in

  • hold chip in shutdown in startup for discharge?

  • add caps in UV and OV as well

  • either protoype changes on current board to find issues, or add a lot of footprints to current board to play around with values and test

    • will do testing this week by @Micah Black

    • josh will add footprints for caps for delay time and gate drive and slew rate current limiting on four layer anyways (will help with testing)

  • datasheet mentions adding these caps with equation to follow

  • add testpoints for enable and shutdown pins on LTC so we can enable and disable easier without firmware

  • high current tespoints - use connector

  • low current - use tespoints with probe hooks maybe

  • high speed - through hole ground beside it and use probe spring?

  • protection on output for overcurrent? add a fuse? figure out if neccessary

  • look for high current/voltage connectors that can be swapped quickly

    • megafit not an option?

    • start a confluence page for this

  • add more mounting points maybe near C12 (in middle ish of board)

  • power calcs on fets for dcdc ? do we need 3 in parallel back to back?

    • there for reducing resistance

    • calculate temperature rise from datasheet

    • 20A, 1.44W = 72 deg C rise in temp <-with horn

    • 10A, 18 deg C rise in temp < -without horn

    • 17A, 50 deg C ->90 deg C if outside? <-max DCDC output

    • probably fine to leave the two paralleling set of fets (no need for 3)

  • hopefully make the board more dense

  • smaller version of automotive fuses to save space

  • fill in values for notes on aux and dcdc power selection

  • can make values of rc on output of temp amplifier, and make it 1k and .1uf

    • or test without op amp for temp sensing, because possibility do not need this

    • have a 0 ohm jumper and nostuff it to bypass op amp

    • try not to increase resistor too much, can have a larger cap

  • maybe find another LDO not QFN for easier to rework

    • needed to replace a few times, so would be nice to have easier package

    • also used on other boards, so review other boards when sending new revs

  • front/rear power have header to jumper + switch between and identical silkscreens

Timelines

Taiping

  • Any major layout changes for our electrical systems should be done now, we should look into it since we have a bit more than a year of time

  • Maybe finish our current revisioning, then look at making changes

  •  

Battery Module Connector

Ricky

  • switched connectors to add two more pins to each

  • add ground pins for thermistors 4 and 5

  • change pin layout for thermistors vs fuse + resistor lines

    • i believe the change was to make the thermistor lines on the inner pins vs. outer? ask @Micah Black

  • we have fuses on AFEs as well, which we technically don’t need that

  •  

AFE

Liam, Micah

  • we may want to change the fuse ratings or take off the fuses that are already there from the module connectors

  •  

DCDC

Aashmika

  • change number of ICL to 2

  • Ambient temperature increase inside a car as 70 deg C to 95 deg C

  • assume ambient as 60 deg C

  • change c16 to a smaller package size and put some caps in series so less likely to crack based on bending of the board

  • down to 60V

  • where would emi ground?

  • email/call vicor and send schematic

  • all ceramics on hv put in series

  • want a better way to discharge

  • look at precharge for discharging

  • test with a transient load

    • drive a fet with a function generator, or use E-load di/dt to simulate a transient load for any ringing on output

    • ask vicor for how to test for this and how to isolated

  • should look at inductor L1 for 25A or a lower DC resistance <-ask vicor for all selected components

  • maybe use a bimetal switch in series with fan to turn on or off,

    • or use thermistor with window comparator

  • fans will be used but not probably during normal driving when excluding horns and transients

    • around 2 amps for low voltage

  • output cap of 15000uf may also have to discharge separately

  • test discharge time for a large load and see if power dist. will switch to aux quick enough

  • maybe have fault pin go to power selection

    • josh add another connector to power selection for fault pin

    • optocoupler and pin on fault before going on 4 pin connector

    • make fault indicator red

  • output led for power green

steering

Aashmika/Sienna

  • move schmitt triggers to top layer so easier

  • talk to adrian about mounting

  • get rid of grounding on one pin on buttons (always pulled low)

  • power connectors to microfit