Build log / Oct 18, 2026

First power-on

Oct 18, 2026hardwaremilestone

The first prototype boards arrived from the factory. This is the moment you find out whether a month of design work is a handheld or a very small heater.

Board powered over USB with a current meter
First power, with a USB meter in line and no screens attached.

Bring-up, step by step

  1. No smoke. Powered with a current limit first, then checked every supply rail on the test points. All in range.
  2. It boots. Linux came up, and I could log in over USB from my Mac.
  3. Both screens. The first time, I plugged the hinge cable in the wrong way round. Nothing broke. Flipped, both screens showed their test pattern with the right colours in each corner.
  4. No tearing. The screens only sent their refresh signal when it was set to line 0. With that set: 60 fps on both screens with no tearing.
  5. Touch. The touch chip came from the factory blank, without firmware. I loaded the panel maker's firmware onto it, and touch worked.
Both screens showing a grid
Both screens alive.
The workbench
The desk that night.

Numbers from the first board

TestResult
Screen refresh at 75 MHz60 fps per screen, 11 ms to send a frame
DS game, both screensabout 56 fps shown and very smooth to the eye; about 1.1 W total
30-minute stress testBoth cores at full load plus constant SD writes: no crash, 0 errors, settled at 52–54 °C
Backlightsabout 40 mA per screen

A few things only got noted down: L / R weren't fitted on this board, and the headphone jack was placed backwards. Both are fixed in the Beta boards.

The next night: a real game

The next evening, still on the bare board with jumper wires, a DS game ran on both screens for the first time.

Oct 19: the first DS game on my own board (Pokémon White 2).

That night I posted my first progress update on r/SBCGaming.

← OlderProving the chip on a dev board