Skip to content

9.5.42 — reverted: Goggle One back on FPGA payload VA-206

Choose a tag to compare

View all tags
@github-actions github-actions released this 29 Jul 06:00
· 29 commits to main since this release
v9.5.42
047ca64

9.5.42 — reverted: Goggle One back on FPGA payload VA-206

This backs out the VA-200 experiment. Functionally identical to 9.5.40 — the Goggle One FPGA payload is back on VA-206 (77-206), as it has been since the 9.5.38 line.

9.5.41 is skipped deliberately. It was briefly published with the VA-200 payload, withdrawn with zero downloads, and reusing that number for different content would be a trap.

What happened

Record OSD: No still records the OSD. The investigation ruled out every software cause:

  • Display_Osd() writes 0x84 = 0x11 / 0x01, unchanged since December 2022 and byte-identical to the official firmware
  • The set of FPGA registers the Goggle One driver writes is identical to official v9.5.0 — nothing was renumbered
  • The recorder has no OSD handling at all; the composite happens in hardware before the encoder

That left the FPGA payload as the only variable — official v9.5.0 ships HAL 75-200 and reportedly does not record the OSD, while this tree follows upstream main and ships 77-206. 9.5.41 tested that theory by reverting Goggle One to VA-200.

Why it was reverted

The reasoning held up, but it was never confirmed on hardware, and the trade was bad:

  • RX-77 paired with VA-200 has never shipped. Official pairs 75-200; this tree pairs 77-206. Receiver and FPGA firmware are normally versioned together
  • This tree's true 90Hz DVR playback and the Goggle One green-raster fix were developed and field-verified against VA-206

Trading a field-verified playback path for an unproven OSD fix was the wrong risk to take in the release line. So it is out, and the experiment lives on its own branch instead.

Where the Record OSD investigation stands

Still unfixed, and still diagnosable at zero risk. 9.5.40 added a register readback, which is the cheap first step and is in this build:

  1. Set Record OSD: No, record ~10s with a flight controller connected
  2. Check hwlog.txt on the SD card for the line starting Display_Osd( or fpga rec start

That says what 0x84 actually held at capture time, which separates "wrong value" from "value got clobbered". If it is the wrong value, the OSD Probe button action (added in 9.5.39) sweeps candidates against VA-206 — a proper fix, rather than a payload downgrade.

The payload experiment remains available as the g1-va200-osd-test build for anyone who wants to confirm the theory on hardware. It is not recommended for general use.

Prime suspect, for the record: the payload moved VA-200 → VA-206 across three commits by ligenxxxx, one of which is "google draw dot when record hdmi in" (5392389, June 2025) — an FPGA change to what gets drawn into the recording, landing after v9.5.0 and before this fork's base.

Everything else

Unchanged from 9.5.40 — the bolagnaise 9.5.19 merge, five Record Bitrate steps, the Shutdown menu entry, the VTX Sent display options with live preview, and the OSD Probe diagnostic. See 9.5.38, 9.5.39 and 9.5.40.

Files

File Size (bytes) MD5
HDZERO_GOGGLE-77-206-9.5.42.bin 9973760 8bbfcf81ee7e1058c697df9fb3e792a7
HDZERO_GOGGLE2-77-184-9.5.42.bin 12892160 5e11468857e32d1d278e7a7777ce02d8
HDZERO_BOXPRO-77-211-9.5.42.bin 9676800 273c075cbe826a8f9ecf3ef2bbec3c63

Copy the .bin for your model to the SD card root, then Menu → Firmware → Update Goggle.

Assets 5
Loading