Pi dash input hat

Overdash

Open-source digital dashboard for Raspberry Pi. Most commercial Pi-based dash solutions lock you into a single ECU brand and require paid third-party software licenses. Overdash doesn't: it speaks J1939 CAN, OBD-II K-line, ELM327, CAN OBD, GPS, and GPIO natively, so it works with engine swaps, custom builds, and vehicles that commercial options simply don't support.

A Python async backend reads from any combination of those sources simultaneously and streams everything over WebSocket to a Chromium kiosk frontend. Gauges are rendered on canvas, and the layout is fully editable in-browser with a visual no-code editor. No paid software required.

Built for a TDI-swapped Land Rover Discovery (J1939 engine + CompuShift Sport TCM + ZF4HP22EH), but config-driven; no code changes needed to adapt it to another vehicle.

Features

• Multiple data sources: J1939 CAN, CompuShift TCM, OBD-II (K-line and ELM327), CAN OBD, Megasquirt/Speeduino USB serial, GPS (NMEA), analog ADC, frequency counter, GPIO inputs, and a mock source for development without hardware
• Visual no-code layout editor: drag to move, resize handles, properties panel for zones/thresholds/colors, gauge palette, theme picker, save to server or export JSON
• Gauge types: circular sweep (tach, speedo), vertical/horizontal bar, large numeric with value mapping, indicator lights, warning light strip with DTC panel
• DTC monitoring: Mode 03 (stored) + Mode 07 (pending) codes, MIL status, tap any active warning light to open the fault code panel
• Odometer: integrates vehicle speed into a persisted lifetime total and resettable trip counter
• PWA: add to home screen on a tablet for fullscreen landscape kiosk mode
• Auth-protected HTTP API: layout save/load, signal snapshot, trip reset, odometer set, per-source simulation mode

Quick start

# Clone
git clone https://github.com/outlandfabworks/overdash.git
cd overdash

# Create a virtual environment and install dependencies
python3 -m venv .venv
.venv/bin/pip install -r requirements.txt

# Test with the mock vehicle source (no hardware required)
.venv/bin/python -m backend.main configs/vehicles/mock.yaml

# Open http://localhost:8080 in a browser

For full installation on a Raspberry Pi (CAN setup, systemd service, Chromium kiosk), see INSTALL.md.

Project structure

backend/
  main.py                     Entry point: wires sources, processors, servers
  data_bus.py                 In-memory signal store with asyncio pub/sub
  config.py                   YAML loader
  sources/
    base.py                   Abstract DataSource
    can_j1939.py              J1939 PGN/SPN decoder (python-can)
    can_compushift.py         CompuShift Sport TCM CAN messages
    can_obd.py                OBD-II over CAN (ISO 15765-4)
    obd_kline.py              ISO 9141-2 / KWP2000 K-line poller
    elm327.py                 ELM327 USB/Bluetooth OBD adapter
    megasquirt.py             Megasquirt 1/2/3 and Speeduino USB serial
    gps_nmea.py               GPS via NMEA serial stream
    analog_adc.py             Analog voltage/sensor via SPI ADC (MCP3xxx)
    frequency_counter.py      RPM/frequency via GPIO pulse counting
    gpio_inputs.py            Overdash Input HAT (turn signals, lights, reverse)
    mock_vehicle.py           Simulated signals for dev/testing
  processors/
    odometer.py               Integrates vehicle_speed → odometer + trip signals
  server/
    websocket.py              WebSocket broadcast server (30 Hz)
    http_api.py               HTTP API + frontend file serving (port 8080)

frontend/

Signal flow

Hardware  →  Source._read_loop()  →  DataBus.publish(name, value, unit)
                                           │
                                   OdometerProcessor (vehicle_speed → odometer/trip)
                                           │
                                   WebSocketServer (batches at 30 Hz)
                                           │
                                   Browser WSClient
                                           │
                                   LayoutManager.update(signals)
                                           │
                                   Gauge.draw(value)  →  <canvas>

Adding a gauge type

1. Create frontend/js/gauges/my_gauge.js extending BaseGauge, implement draw(value).
2. Register it in index.html: GaugeRegistry.register('my_gauge', MyGauge).
3. Add a gauge entry in the layout JSON with "type": "my_gauge".

Adding a data source

1. Create backend/sources/my_source.py extending BaseSource, implement _read_loop().
2. Register it in backend/main.py: _SOURCE_TYPES['my_source'] = MySource.
3. Add a source entry in the vehicle YAML with type: my_source.

TDI Discovery: signal reference

CompuShift Sport CAN IDs

Default map derived from community reverse-engineering (Sport firmware ≥ 2.x). Override via messages: in the vehicle YAML if your TCM uses different IDs.

License

MIT. See LICENSE. Third-party license notices in THIRD_PARTY_LICENSES.md.