Skip to content

Connecting AMYboard to Tulip over grove (I2C host) can send Tulip into a power boot loop #1026

Description

@bwhitman

Summary

Several users report that connecting an AMYboard to a Tulip CC using the Tulip's grove / STEMMA I2C header and the AMYboard's back "tulip" I2C host port sends the Tulip into a power boot loop (it resets repeatedly at/before the REPL).

This is the documented, supported hookup — from docs/amyboard/README.md: "you can directly connect your AMYboard to it via the GROVE port and the back I2C HOST connector. Tulip will power and communicate with the AMYboard!"

Leading hypothesis: the Tulip's power source can't feed both boards

When the AMYboard is connected this way and isn't separately powered, it pulls its power from the 3.3 V on the Tulip grove port. Our own docs give the numbers:

Adding ~350 mA @ 3.3 V (≈ +230–350 mA @ 5 V once you account for the regulator and losses) on top of Tulip's ~575 mA @ 5 V pushes the total toward — or past — what a marginal 5 V source can actually deliver. If the supply / cable / hub can't keep up, the 5 V rail sags, Tulip's 3.3 V regulator drops out, and the ESP32‑S3 brownout detector resets the chip. It then tries to boot, pulls the same current, browns out again → boot loop. Inrush is highest at boot, which fits a boot-loop symptom (rather than a mid-session crash).

This lines up with the usual real-world culprits:

  • weak / old USB chargers (the ~0.5 A "computer USB port" class)
  • charge-only or thin high-resistance USB-C cables
  • powering Tulip from an unpowered hub
  • a low battery

Light investigation — what I confirmed in the repo

  • The supported path is grove ↔ AMYboard back "tulip" host port; the AMYboard front-panel I2C is for accessories, the back port is host / power-in (faq.md:110).
  • AMYboard power inputs and the ~350 mA grove draw are as quoted above (measured ~2.94 V at the 4th board in a chain, per the same FAQ entry).
  • I found no explicit ESP32 brownout-detector configuration in the Tulip firmware (tulip/esp32s3), so it runs at the ESP-IDF default (brownout detector enabled) — consistent with a brownout-triggered reset loop.
  • I found no current limiting / soft-start on the grove 3.3 V rail, and no power warning in the quick-start docs about the extra draw when attaching an AMYboard.

Note: the +5 V-side current estimate and the regulator topology are inferred, not bench-measured — worth confirming on hardware.

How an affected user can confirm it's power

  • Watch the Tulip serial/boot log for Brownout detector was triggered (or repeated resets).
  • Power the AMYboard separately over its own USB-C — it takes the highest voltage present, so 5 V USB-C will feed it instead of pulling 3.3 V from the Tulip. If the loop stops, it's power.
  • Try a known-good ≥1–2 A 5 V supply and a quality data USB-C cable, plugged straight in (no hub).

Possible resolutions

  • Docs: add a power note to the AMYboard ↔ Tulip "Quick start - Tulip" section and Tulip troubleshooting — connecting an AMYboard adds ~350 mA, so use a ≥1–2 A supply or power the AMYboard on its own USB-C.
  • Bench check: measure the 5 V / 3.3 V rails at boot with an AMYboard attached on a 1 A vs 2 A supply; confirm brownout vs. regulator dropout vs. inrush.
  • If a firmware change can help (e.g. deferring high-current init, ramping display brightness at boot), treat as a firmware bug.

Asks for reporters

If you hit this, please share: Tulip hardware rev, what's powering it (charger/cable/hub/battery + rated amps), whether the AMYboard had its own power, and any boot-log output.


Filed from a maintainer investigation; numbers cited are from the repo docs unless marked inferred.

Metadata

Metadata

Assignees

No one assigned

    Labels

    amyboarddocumentationImprovements or additions to documentationtulipccTulip CC hardware

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions