Sensors & configuration
seglcd live and seglcdd read the sensors with the same code (src/SegLcd.Linux). The LB360-S only shows CPU and
GPU temperature (integer °C, max. 99); the other values are sent along like the original software does.
Sensors are read every sample_interval_ms (1000); the last values are sent every update_interval_ms (250, max. 400 –
otherwise the display turns off).
Sources
| Field | Source | Default (first match) |
|---|---|---|
| CPU temperature | hwmon | k10temp:Tctl, zenpower:Tdie, zenpower:Tctl, coretemp:Package id 0 |
| CPU usage | /proc/stat | fixed |
| CPU power | RAPL or hwmon | rapl, zenpower:SVI2_P_Core+zenpower:SVI2_P_SoC |
| CPU clock | cpufreq scaling_cur_freq of all cores | average (cpu_clock_mode) |
| CPU voltage | hwmon | zenpower:SVI2_Core |
| GPU temperature | hwmon or NVML | amdgpu:edge, nvml:temperature, xe:pkg |
| GPU usage | DRM or NVML | drm:busy, nvml:usage |
| GPU power | hwmon, NVML or RAPL | amdgpu:PPT, nvml:power, rapl:uncore |
| GPU clock | hwmon, NVML or DRM | amdgpu:sclk, nvml:clock, drm:freq |
| fan, pump | hwmon | none – depends on the mainboard header |
| RAM usage | /proc/meminfo | fixed |
Missing or unreadable values are sent as 0.
Notes:
- CPU power: RAPL is the exact package power (as shown by btop), but readable by root only
(PLATYPUS). The optional udev rule opens
it for the group
rapl. Without RAPL, zenpower’s Core + SoC is an approximation; k10temp has no power sensor. - Intel:
coretempfor the temperature, RAPL for the power; integrated GPUs: clock viadrm:freq, power viarapl:uncore, no temperature. - NVIDIA: via NVML (
libnvidia-ml.so.1, loaded at runtime); fan only in percent. - Hybrid graphics: configure the GPU explicitly (
nvml1:…,drm0:freq).
Sensor specs
The same syntax works for CLI options, config.toml and D-Bus SetSensor:
| Spec | Example | Meaning |
|---|---|---|
chip:label | k10temp:Tctl, it8686_31050607:fan5 | hwmon sensor – stable, prefer this |
hwmonN/name | hwmon5/fan1 | hwmon path (the number may change on reboot) |
a+b | zenpower:SVI2_P_Core+zenpower:SVI2_P_SoC | sum |
rapl[:domain] | rapl, rapl:uncore | RAPL power: package (default), core, uncore, dram, psys |
nvml[N]:metric | nvml:temperature | NVIDIA GPU N: temperature, power, clock, memclock, usage, fan |
drm[N]:metric | drm:busy | DRM card N: busy, freq |
| empty | "" | automatic (defaults above) |
seglcd sensors lists every available source with its spec and current value; “(permissions)” marks an unreadable
RAPL counter.
Fan and pump: the mainboard manual or BIOS names the headers (CPU_FAN, AIO_PUMP, …). The LB360-S pump runs
constantly at about 3000 RPM; radiator fans follow the CPU load.
Configuration
| Field | seglcd live | config.toml / D-Bus SetSensor |
|---|---|---|
| CPU temperature / power / voltage | --cpu-temp-sensor, --cpu-power-sensor, --cpu-voltage-sensor | cpu_temperature, cpu_power, cpu_voltage |
| GPU temperature / usage / power / clock | --gpu-temp-sensor, --gpu-usage-sensor, --gpu-power-sensor, --gpu-clock-sensor | gpu_temperature, gpu_usage, gpu_power, gpu_clock |
| fan / pump | --fan-sensor, --pump-sensor | fan, pump |
| CPU clock mode | --cpu-clock-mode | cpu_clock_mode (TOML only) |
Recommended: find the specs with seglcd sensors, try them with seglcd live --show --fan-sensor … --pump-sensor …,
then put them into config.toml and run systemctl --user reload seglcdd.
config.toml
User service: ~/.config/seglcd/config.toml, system service: /etc/seglcd/config.toml (also seglcdd --config);
values saved via D-Bus are in /var/lib/seglcd/config.toml and take precedence – delete it to use /etc only.
A missing file or key means the default; invalid values are logged and replaced by defaults. Complete default file:
# seglcdd – configuration
# Location: $XDG_CONFIG_HOME/seglcd/config.toml (usually ~/.config/seglcd/config.toml), or: seglcdd --config <path>
# Format: TOML 1.1. This file contains ALL options with their default values – a missing file or
# missing keys behave the same.
#
# Sensor specs (section [sensors]):
# "" automatic selection (default order in the comment)
# "chip:label" hwmon sensor, e.g. "k10temp:Tctl", "it8686_31050607:fan5" (stable, preferred)
# "hwmonN/name" hwmon path, e.g. "hwmon5/fan1" (the number may change)
# "a+b" sum, e.g. "zenpower:SVI2_P_Core+zenpower:SVI2_P_SoC"
# "rapl[:domain]" RAPL power: package (default), core, uncore, dram, psys (permissions: udev/72-rapl-energy-read.rules)
# "nvml[N]:metric" NVIDIA GPU N: temperature, power, clock, memclock, usage, fan (%)
# "drm[N]:metric" DRM card N: busy (%), freq (MHz, Intel)
# List the available sources: seglcd sensors
# Feed the display. false = send the off packet and release the device (e.g. for experiments with seglcd).
enabled = true
# Time between two packets in ms (50–400). The display turns off after about 400–500 ms without a packet.
update_interval_ms = 250
# Time between two sensor reads in ms (100–10000). In between, the last values are repeated.
sample_interval_ms = 1000
# Device path, only needed with several displays, e.g. "/dev/hidraw5". Empty = first segment LCD found.
device = ""
[sensors]
# CPU temperature. Automatic: k10temp:Tctl, zenpower:Tdie, zenpower:Tctl, coretemp:Package id 0
cpu_temperature = ""
# CPU power (W). Automatic: rapl, zenpower:SVI2_P_Core+zenpower:SVI2_P_SoC
cpu_power = ""
# CPU core voltage (V). Automatic: zenpower:SVI2_Core
cpu_voltage = ""
# CPU clock from the values of all cores: "avg" (average), "max", "min", "first" (core 0).
cpu_clock_mode = "avg"
# GPU temperature. Automatic: amdgpu:edge, nvml:temperature, xe:pkg
gpu_temperature = ""
# GPU usage (%). Automatic: drm:busy, nvml:usage
gpu_usage = ""
# GPU power (W). Automatic: amdgpu:PPT, nvml:power, rapl:uncore
gpu_power = ""
# GPU clock (MHz). Automatic: amdgpu:sclk, nvml:clock, drm:freq
gpu_clock = ""
# Fan and pump speed (RPM). No automatic selection – the mainboard header depends on the system, e.g.:
# fan = "it8686_31050607:fan1"
# pump = "it8686_31050607:fan5"
fan = ""
pump = ""