Status: research-only and inconclusive. Nothing was implemented or applied on the lab machine. No driver, service, task, policy, registry value, firmware option, fan behavior, power setting, or Windows security/update control was changed.

The evidence question

Can supported, already available Windows and HP surfaces provide a trustworthy, low-overhead CPU-package temperature, ambient-temperature, and fan-state gate for unattended EXP-001 startup measurements?

Not yet. The project will not relabel an unidentified ACPI thermal zone as a CPU temperature, calculate battery wear without design capacity, or invent a readiness band. A future instrumentation experiment must identify the sensor, units, access requirements, sampling behavior, overhead, and relationship to throttling.

What the ZBook reported

  • Seven signed Microsoft ACPI Thermal Zone device instances were present, using machine.inf driver version 10.0.26100.1150.
  • The non-elevated MSAcpi_ThermalZoneTemperature query returned access denied. This does not prove that the firmware exposes no thermal zones.
  • Win32_TemperatureProbe, Win32_Fan, and BatteryTemperature returned no instances or readings.
  • PwrTest was not installed. HP PC Hardware Diagnostics Windows, HP Support Assistant, and an HP_TOOLS volume were not detected by the inspected Windows inventory surfaces.
  • Signed Intel Dynamic Tuning Technology Manager, two Generic Participants, and a Processor Participant were present at version 8.7.10802.26924.
  • The battery reported AC online, full/remaining capacity 41407, 101 cycles, and voltage 12560. Design capacity was unavailable, so no health percentage was inferred.

These are single read-only observations on the validated HP ZBook Firefly 14 inch G8, Windows build 26200, BIOS T76 01.24.02. They are not repeated benchmark runs.

Documented limits and safety decision

Microsoft describes ACPI thermal zones as firmware-defined abstractions that can contain several heat-generating or managed devices. A zone reading is not automatically a CPU-package reading. Microsoft also says current WMI implementations do not populate Win32_TemperatureProbe.CurrentReading.

HP documents a Fan Speed Test and a roughly 320-second Fan Thermal Test. Those are interactive hardware diagnostics, not low-overhead unattended benchmark sensors. Running a stress-oriented diagnostic could also heat the computer and contaminate the next performance run.

Intel documents DTT as an OEM-configured system coordinating performance, power, acoustics, and thermals, and advises against disabling or uninstalling it. This layer therefore rejects disabling DTT or forcing fan behavior.

Evidence ledger

Documented facts

  • Windows thermal management depends on firmware, Windows, and driver cooperation.
  • ACPI thermal zones are abstract firmware descriptions, not guaranteed CPU sensors.
  • HP's fan checks are explicit diagnostic runs.
  • Intel advises against disabling DTT.

Lab measurements

One non-elevated, observation-only inventory was captured. No repeated thermal samples, ambient reading, fan stress test, startup benchmark, apply, rollback, or reboot was performed.

Hypotheses

An elevated, identified ACPI or trusted vendor sensor may provide useful data, but its identity, overhead, and relationship to throttling remain unproven.

Unresolved questions

  • Which sensor can identify CPU-package temperature and throttling on this exact ZBook?
  • What elevation and sampling overhead does it require?
  • How should ambient temperature, fan state, settling time, and the pre-run band be recorded?

Primary sources

All sources were retrieved July 27, 2026.

YouTube briefing

Presenter: The first hourly layer asked whether the ZBook can prove a consistent thermal starting point before every startup test.

Windows found ACPI thermal-zone devices and Intel Dynamic Tuning components, but the non-elevated inbox interfaces did not return an identified CPU-package temperature or fan state. An unknown ACPI zone is not being renamed as a CPU sensor.

HP's documented fan diagnostics are useful hardware checks, but they are not low-overhead unattended sensors. Intel also advises against disabling Dynamic Tuning Technology, so no DTT component, fan control, driver, firmware option, service, policy, or power setting was changed.

This layer is preserved as inconclusive. A later instrumentation experiment must prove sensor identity, units, access, overhead, ambient and fan evidence, and a repeatable pre-run band. The cursor now moves to layer 2: hardware resources and bottlenecks.

Read the complete source script

Source and next layer

The next hourly cycle position is layer 2, hardware resources and bottlenecks.

Back to all development updates