Power Considerations - mhightower83/Arduino-ESP8266-misc GitHub Wiki

WIP

Power Considerations for an ESP8266 Project

Outline

  • ESP8266 power demands
    • Interesting post at https://www.ondrovo.com/a/20170205-esp-self-destruct/
      • TODO: reread ResetInfo->reason is 6 (REASON_EXT_SYS_RST) even on power up. I've seen this with WeMos Pro. Add this here because of the power-related issue.
      • Walkaways,
        • too large a CAP on the 3.3V rail can slow the rise time at power-up causing catastrophic failure in the ESP8266.
        • I believe the ESP8266 datasheet has required rise time specifications for PD/RST/Power.
        • I think a lot of the voltage droop he shows will go away with proper board layout and the ESP8266 being closely nestled with a nice 3.3V LDO regulator.
    • His next post, https://www.ondrovo.com/a/20170207-esp-consumption/, provides useful readings for determining LDO Regulator parameters needed.
      • again the deep voltage droops should go away with a proper PCB and LDO regulator.
      • Need to put one of the dev boards on an oscilloscope and measure this.
  • LDO Regulator operating parameters
    • Small size vs power dissipation
    • PCB is part of the heat management system.
      • This appears to be true for a lot of the SMD parts today.
    • High input voltage excess heat - failure
    • Looks like there are lots of knock-offs.
      • reference Reddit article on knockoff 4A2D parts
    • Too many 4A2D (150ma) parts in the wild and they should be 4A2D (300ma)
      • It is easy to miss the A vs A in the datasheet's table for the current specification.
      • It seems possible that 150ma may be okay when you consider average power consumption.
        • When you consider the time spent transmitting and consuming ~200ma compared to the rest of the time at 60-70ma.
        • For this to work the silicon must be the same as that which supports 300/500ma so that the peak consumption can be handled. Then, the resulting average power around a surge would need to be appropriate for the SOC-23 package.
          • I don't think the datasheet parameters cover this use case.
          • Because of power dissipation issues the input voltage cannot be too high.
          • And it cannot go too low else lose the ability to maintain a constant output at full load.
          • It may be a good idea to avoid using the Dev Board's 3.3V supply for anything off-board you want to connect.
      • The D indicates the lot number
      • For this part 4A2x most of the parts we see pictures of on the internet are from lot D
    • Some schematics will indicate 500ma parts for the LDO 3.3V regulator. When you look at the part on the board, you discover that a similar, but 300ma part, has been substituted.
    • Fast response to Transient Load changes.
      • 60ma to ~300ma
  • Cell phone chargers used as 5V power supplies
    • Designed for steady current flow for charging batteries
      • Voltage regulation is less than optimal.
      • Output voltage levels vary
    • Poor Transient Load response
    • Inexpensive and many to chose from
      • Some really cheap adapters are fire and safety hazards
        • No isolation from Mains
        • poor construction
        • No Safety Agency approvals, eg. UL, ETL, CE, etc.
          • Why consider Safety Agency Approvals as a concern?
            • Component selection in critical areas will be limited to parts with well-defined failure modes. For example, ceramic capacitors can fail open or fail short. For a CAP that connects across the main for RF interference shunting, a fail short could cause a fire. You want a ceramic CAP that fails open - never short. These exist and are a verification check box for Safety approvals.
    • A Switch mode PS takes a few cycles to catch up to a transient load change resulting in voltage droop.
      • This leaves the +5V rail unsuitable for most other devices except relays and LEDs.
      • Some IoT devices try to use the +5V rail for power monitoring Chips. From what I have seen, they tend to have occasional erratic readings.
    • Capacitor on the +5V to help with transient load
      • Not much success
      • IoT devices tend to have a 500-680uF cap on the +5V rail. Dev boards tend to have 1uF to 10uF.
        • Dev Boards might perform better on 5V chargers with an additional 500uF CAP.
        • Also, verify the Dev Board already has a .1 to 1uf ceramic CAP.
  • Computer Powered USB port
    • TI Application note Power Management - Powering electronics from the USB port
      • Maximum input capacitance 10uF
      • Maximum inrush 50uC
    • Should have a good Transient Load response
      • 60ma - 300ma would be a much smaller percentage of the overall load the PC Power supply would be dealing with.
  • Powered USB Hub.
    • Transient Load response - Unknown, it might depend on the total load already being handled. The more the better.
  • USB cablesFor
    • We use them for development to flash the module
    • We use them to provide power later.
    • The gauge of the wire and length may have an impact on our project's stability.
  • USB connector quality issues, resistive connector
    • power loss represented by voltage drop
    • power loss => excess heat
    • failure
    • meltdown
    • fire
    • Amazon reviews of melted connectors and destroyed phone jacks.
    • poor wire to plug connection cold solder, bad crimp, failed crimp from stress, etc.
⚠️ **GitHub.com Fallback** ⚠️