Native touch display for the AAG CloudWatcher Solo weather station and NINA and Domeshell Status, built on the Waveshare ESP32-P4 WiFi6 Touch LCD 4B (720x720).
A standalone observatory weather monitor that runs on bare metal (no Linux, no OS overhead). Polls an AAG CloudWatcher Solo, a N.I.N.A. imaging session, and an AstroShell dome controller over your local network, then renders everything on a 4-inch 720x720 IPS touchscreen in real time. Designed for remote observatory dashboards where you need a quick glance at conditions from across the room.
Screen 1 — Home Overview
- Big color-coded weather status: CLEAR / CLOUDY / OVERCAST and DRY / WET / RAIN
- Central SAFE / UNSAFE indicator for at-a-glance observing decisions
- Large NTP clock overlay (48px white) rendered directly inside the chart area, readable from several meters away
- Dual-line 24h temperature chart: sky temperature (green) + ambient temperature (red) with auto-scaled Y-axis
- Sky temp, ambient temp, and dew point values in the header
- Dome status banner with colored box: CLOSED (green), OPEN (red), or unknown (yellow) — queried directly from the AstroShell dome Arduino controller, independent of N.I.N.A.
Screen 2 — N.I.N.A. Live Image
- Displays the latest sub-exposure from a running N.I.N.A. imaging session
- Software JPEG decoder (TJPGD) with automatic downscaling to fit the 720x670 display area
- Metadata overlay: target name, filter, exposure time, star count, HFR, gain, image count, timestamp
- Manual refresh button — tap to fetch a new frame immediately without waiting for the poll timer
- Graceful degradation: shows "No active session" or "NINA offline" when the imaging PC is unreachable
Screen 3 — Sensor Details Dashboard
- 3x2 card grid: Sky Temperature, Ambient Temperature, Humidity, Rain, Sky Quality (mpsas), Pressure
- Each card shows the current value and a color-coded safety status (green = safe, red = unsafe)
- Title bar with data timestamp and safe/unsafe indicator
- Large bottom banner: "SAFE FOR OBSERVING" (green) or "UNSAFE FOR OBSERVING" (red)
Screen 4 — Historical Charts
- 6 tabbed 24h line charts: Cloud, Temperature, Humidity, Light, Rain, Pressure
- Each chart shows today (bright cyan) vs. yesterday (dim gray) as overlaid line series
- Fixed Y-axis ranges for consistent comparison across days (e.g., humidity always 30-100%, pressure auto-scaled)
- NTP clock overlay on chart area
- Dynamic Y-axis labels, current value and min/max summary below each chart
- 200-point downsampled display from ~750 raw data points per series
- NTP time sync — Automatic SNTP synchronization via
pool.ntp.orgafter WiFi connects, timezone set to CET/CEST - Direct dome status query — Polls the AstroShell Arduino dome controller at
192.168.1.177via HTTP (/?$S), completely independent of N.I.N.A. or ASCOM. Works even when the imaging PC is off. - WiFi auto-reconnect — Exponential backoff with up to 10 retries, then periodic 30s reconnect attempts
- ESP-Hosted WiFi — The ESP32-P4 has no native WiFi; all network traffic runs through the onboard ESP32-C6 co-processor via SDIO
| Data Source | Default Interval | Configurable |
|---|---|---|
| CloudWatcher current data | 60s | Yes (menuconfig) |
| CloudWatcher 24h graphs | 240s | Yes (menuconfig) |
| N.I.N.A. image + dome status | 120s | Yes (menuconfig) |
| NTP time sync | 1 hour | Via sdkconfig |
- Board: Waveshare ESP32-P4 WiFi6 Touch LCD 4B
- CPU: ESP32-P4 dual-core RISC-V @ 360 MHz — significantly faster than ESP32-S3 or classic ESP32
- Memory: 32 MB PSRAM + 32 MB Flash — enough to hold decoded JPEG images, 24h graph data (~108 KB for 6 series), and the full LVGL framebuffer simultaneously
- Display: 4" 720x720 IPS, ST7703 controller via MIPI-DSI (2-lane) — not RGB parallel, which means higher bandwidth and less CPU overhead for rendering
- Touch: GT911 capacitive I2C touchscreen (uses backup address 0x14)
- WiFi: ESP32-C6 co-processor via SDIO (ESP-Hosted), WiFi 6 capable
- Weather Station: AAG CloudWatcher Solo (HTTP polling, no cloud dependency)
- Dome Controller: AstroShell Arduino (HTTP GET, plain text status response)
- Imaging Software: N.I.N.A. with Advanced API plugin (ninaAPI by Christian Palm)
The ESP32-P4 is a significant step up from earlier ESP32 variants for display-heavy applications:
- 360 MHz dual-core RISC-V vs. 240 MHz Xtensa on ESP32-S3 — 50% higher clock speed with a more efficient ISA
- Native MIPI-DSI for the display interface — dedicated peripheral, zero CPU involvement for pixel transfer
- 32 MB PSRAM — enough to hold multiple full-resolution decoded images in memory without swapping or streaming
- Hardware JPEG decoder available (though this project uses the software TJPGD decoder for flexibility with non-standard JPEG streams)
- Sustained 30 fps LVGL rendering on a 720x720 display while simultaneously running HTTP clients, JSON parsing, and JPEG decoding on background tasks — no frame drops
main/
main.c — Entry point, FreeRTOS task creation, polling loops
display_driver.c/h — Manual DSI + GT911 init with graceful I2C fallback
wifi_manager.c/h — ESP-Hosted WiFi + NTP time sync
cloudwatcher_client.c/h — HTTP fetch + key=value parser for lastData.pl / graphData.pl
nina_client.c/h — N.I.N.A. image fetch, JPEG decode, dome status query
ui_main.c/h — Screen management, swipe + button navigation
ui_home.c/h — Home screen: status + chart + clock + dome banner
ui_nina.c/h — N.I.N.A. image screen with metadata overlay
ui_dashboard.c/h — Sensor card grid + safe/unsafe banner
ui_charts.c/h — 6 tabbed 24h charts with today/yesterday overlay
Kconfig.projbuild — All configurable parameters (WiFi, IPs, ports, intervals)
You need ESP-IDF v5.4+ (the ESP32-P4 is only supported from v5.4 onwards).
# Install system dependencies via Homebrew
brew install cmake ninja dfu-util python3
# Clone ESP-IDF v5.4 (or newer) into ~/esp
mkdir -p ~/esp && cd ~/esp
git clone -b v5.4 --recursive https://github.com/espressif/esp-idf.git
# Run the ESP-IDF install script (downloads toolchains for ESP32-P4)
cd ~/esp/esp-idf
./install.sh esp32p4This takes a few minutes on first run. It installs the RISC-V toolchain, Python packages, and build tools into ~/.espressif/.
cd ~/Desktop # or wherever you keep your projects
git clone https://github.com/joergsflow/cloudwatcher-waveshare.git
cd cloudwatcher-waveshareYou need to source this once per terminal session before building:
source ~/esp/esp-idf/export.shTip: Add an alias to your
~/.zshrcfor convenience:alias get_idf='source ~/esp/esp-idf/export.sh'Then just type
get_idfin any new terminal.
idf.py menuconfigNavigate to CloudWatcher Configuration and set:
| Setting | Default | Description |
|---|---|---|
| WiFi SSID | MyNetwork | Your WiFi network name |
| WiFi Password | MyPassword | Your WiFi password |
| CloudWatcher IP | 192.168.1.151 | CloudWatcher Solo IP address |
| CloudWatcher Port | 80 | CloudWatcher HTTP port |
| Poll Interval | 60s | Current data refresh interval |
| Graph Poll Interval | 240s | Graph data refresh interval |
Navigate to NINA Configuration and set:
| Setting | Default | Description |
|---|---|---|
| NINA PC IP | 192.168.1.114 | IP of the imaging PC running N.I.N.A. |
| NINA API Port | 1889 | ninaAPI Advanced API plugin port |
| NINA Dome Port | 1888 | NINA dome control API port |
| Dome Controller IP | 192.168.1.177 | AstroShell Arduino dome controller IP |
| Dome Controller Port | 80 | AstroShell dome controller HTTP port |
| NINA Poll Interval | 120s | Image + dome refresh interval |
Press S to save, then Q to quit. The settings are stored in sdkconfig.
idf.py buildFirst build takes a few minutes (compiles LVGL, ESP-Hosted, etc.). Subsequent builds are incremental and much faster.
Connect the board via USB-C. The Waveshare ESP32-P4 board uses a CH340 USB-serial chip. On macOS the port name looks like /dev/cu.usbmodemXXXX (the exact number changes on reconnect).
# List available serial ports
ls /dev/cu.usbmodem*If nothing shows up, make sure you're using the correct USB-C port on the board (the one labeled USB or UART, not the one labeled USB-OTG). You may also need to install the CH340 driver if macOS doesn't recognize it.
# Flash firmware and open serial monitor (replace port if needed)
idf.py -p /dev/cu.usbmodem* flash monitorThe serial monitor shows boot logs and WiFi connection status. Press Ctrl+] to exit the monitor.
# Full workflow in one go
source ~/esp/esp-idf/export.sh
idf.py build
idf.py -p /dev/cu.usbmodem* flash monitorFreeRTOS Tasks:
LVGL task — display rendering (~30 fps, managed by BSP)
cw_poll_task — CloudWatcher current data + graph fetching
nina_poll_task — N.I.N.A. image fetch + dome status query
Screens: Home <-> NINA <-> Details <-> Charts (swipe or tap nav bar)
Network stack:
ESP32-P4 (main CPU) --SDIO--> ESP32-C6 (WiFi co-processor) --WiFi--> LAN
The ESP32-P4 has no native WiFi — it uses an onboard ESP32-C6 co-processor connected via SDIO (ESP-Hosted). All HTTP clients run on the P4 and transparently route through the C6.
| Source | Endpoint | Format |
|---|---|---|
| CloudWatcher current | http://<cw_ip>/cgi-bin/lastData.pl |
Key=value pairs (~300 bytes) |
| CloudWatcher graphs | http://<cw_ip>/cgi-bin/graphData.pl |
JavaScript arrays (~160 KB, ~750 pts/series) |
| N.I.N.A. image | http://<nina_ip>:<port>/v2/api/equipment/camera/image |
JPEG binary |
| N.I.N.A. metadata | http://<nina_ip>:<port>/v2/api/history/latest |
JSON |
| Dome status | http://<dome_ip>/?$S |
Plain text: "OPEN" or "CLOSED" |
Based on the AAG CloudWatcher Solo configuration for observatory Wietesch (52.17N, 7.25E):
| Sensor | Safe | Warning | Unsafe |
|---|---|---|---|
| Sky Temp (clouds) | < -6 C (Clear) | < 5 C (Cloudy) | >= 5 C (Overcast) |
| Rain | > 2500 (Dry) | > 2000 (Wet) | <= 2000 (Rain) |
| Sky Quality | > 17 mpsas (Dark) | > 13 mpsas (Light) | <= 13 mpsas (Bright) |
| Humidity | < 50% (Dry) | < 80% (Normal) | >= 80% (Humid) |
- NTP time sync (pool.ntp.org, CET/CEST timezone) with large clock overlay on home screen chart
- Direct AstroShell dome controller query (HTTP
/?$S) — independent of N.I.N.A. - Dome status banner with colored box: CLOSED (green), OPEN (red), Domestatus? (yellow)
- NINA/dome poll interval reduced from 5 minutes to 2 minutes
- Clock overlay on charts screen
- N.I.N.A. live image display with software JPEG decoder (TJPGD)
- Dome status display on home screen
- Manual image refresh button
- N.I.N.A. image screen with hardware JPEG decoder
- Fixed chart Y-axis scaling with per-series ranges
- 4-screen navigation (Home, NINA, Details, Charts)
- 3-screen layout: Home, Dashboard, Charts
- Swipe navigation + bottom nav bar
- 24h historical charts with today/yesterday overlay
- Manual Y-axis labels (LVGL 9 workaround)
MIT