Giters

lab11 / socitrack

A wearable platform for social interaction studies

Home Page:https://dl.acm.org/doi/10.1145/3372224.3419190

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

HW: Rev-i Updates

ppannuto opened this issue · comments

commented
  • Fix D1/D2 footprint

    • During assembly we found that the footprints for D1 and D2 seem to be rotated 90 degrees. It looks like they will have a connection if cathodes (pin 1 and 2) are oriented top pad 1 and 2 on the board, but won't likely get the fill in the castellation. It's possible they may also pull to one side during reflow and not make a connection.

  • Update S2 footprint

    • The lands pitch is 0.48mm for PN SS2/SKY13373-460LF, designator S2. But the datasheet of the part indicates the leads pitch is 0.5mm, if you can update the gerber for future order it would be nice.

  • Update battery voltage divider — Internal ADCref is 1.19V, so no resolution above 1.19V input
    • BAT --> 5.1 --> BATMON --> 1.87 --> GND
    • 4.4 -> 1.180V; 3.0 -> 0.805V
  • R8 / R10 are flipped 😭
  • Update 1.8V net to target 1.9V to match Ambiq EVBs? [Verify 1.9V okay across net..]
  • Add 100pF decaps to RF Switch lines to match ref schematic
  • Add solder jumpers pin headers for primary power supply?

DW

  • Add solder jumpers 0Ω for DW power rails
  • Can we do RF switch bypass option? Can we do optional SMA connector?
  • Consider nudging 2.5V rail higher to give more headroom above RF switch V_min?
  • Fix DW Crystal value (2pF || DNP..)
  • Consider Add TXCO for DW
    • the UWB EVB is highly temperature sensitive: today the (both) EVBs starts to have radio errors, which got me insane… but they got right again after I left them cool down for a while
    • and while the EVB-EVB are having errors, the revI without UWB crystal cap could still talk to them

AP4

  • Route SWO to debug header as well
  • C48 should tie to BLE RF GND
  • Ambiq—Your RSTN should have a 1nF cap to filter spurious noise. Since it is connected to nRST make sure it’s open drain.
  • Ambiq—VDDAUDA must be powered (even if not using BLE or AUDADC). I see you have an enable triggered by a pin. Please note this can cause issues.
  • Ambiq-DNP C4/C5, should be enough in IC already for trim
  • Update BLE filter to match Ambiq EVB rather than using old filter from Nordic
  • Update AP4 schematic with I/O voltage domain info; consider emptying VDDAUDD if possible?
  • Replace AP4 Blue with Blue Plus
    • Create new package
    • Reason through any I/O updates (esp power domain checkboxes below)
    • Reroute
  • Add solder jumpers for VDD18,USB09,USB33 rails