Skip to content

About

True hardware-level bottom display off, wake repair, and live CPU/display telemetry for the AYN Thor.

Topics

Resources

Contributing

Security policy

Stars

6 stars

Watchers

2 watching

Forks

Latest commit

 

History

129 Commits

Folders and files

Repository files navigation

Jesty Thor Fix

Fix three AYN Thor issues in one lightweight app.
Truly power off the bottom screen in TOP mode, fix the AYN Dashboard CPU clock bug,
and help the Thor return to sleep after unexpected closed-lid wake-ups.

Download v1.7.1 · Features · Results · What's next · ☕ Support

Release v1.7.1 AYN Thor Android 13 No Magisk or manual rooting Lightweight background operation GPL-3.0

Signed releases · Tested on a physical Thor · Works with the app closed

AYN NATIVE TOP · FIXES OFF TRUE OFF + CPU FIX ON
v1.7.1 app screenshot: stock AYN black-screen mode, CPU clocks pinned high, battery-draw readout around 1.72 W v1.7.1 app screenshot: TRUE OFF, CPU Fix active, normal CPU clocks, battery-draw readout around 0.95 W

Current app screenshots on the same Thor. With both fixes enabled, CPU clocks are no longer pinned high and the bottom screen is powered off. The app shows about 1.72 W (fixes OFF) versus 0.95 W (ON) in these TOP-mode screenshots.

Lightweight and safety-focused. In physical idle tests, the daemon used around 0.09% of one CPU core in BOTH and 0.32% in TOP (tested v1.7.1 candidate). It does not change CPU frequencies, voltages, thermal limits or firmware, and verifies the display state after applying a fix. Performance evidence · Safety and validation.


What it fixes

Feature Why it matters
True Bottom Screen Off When you use just the top screen, actually powers down the lower display, not just the image on it. Restores the correct state after sleep and wake.
AYN Dashboard CPU Fix Stops the Dashboard from keeping CPU speeds unnecessarily high while the Thor is idle. No manual CPU tuning needed.
Closed-Lid Wake Guard Helps the Thor return to sleep after an unexpected wake while the lid is closed.

You can close the app after setting it up; the fixes keep working.


Measured on a real Thor

The fix does more than display a black image. On physical AYN Thor hardware, the lower display pipeline was verified as powered off in TOP mode, while the CPU fix addressed unusually high idle clock speeds.

Controlled A/B/A benchmark

The September 2026 test used repeated, logged 1-second samples under matched conditions. The results below are system-power proxy measurements, not a direct battery-life test.

Controlled TOP-mode comparison: native mode showed about 2.03 W of estimated system-power proxy; True Bottom Screen Off showed about 1.24 W, roughly 39 percent lower in this test.

About 39% lower estimated system-power draw in this short test. This is a power proxy from the Thor's hardware readings, not a measured battery-life improvement. Brightness, workload and charging conditions affect real-world results.

High-performance CPU cluster: 2.707 GHz average in native TOP mode versus 1.654 GHz with True Bottom Screen Off; near-maximum CPU readings changed from 75 of 75 samples to 0 of 45.

From stuck-near-maximum to normal clock behaviour: all 75/75 native TOP samples were near the performance cluster's maximum speed, compared with 0/45 after the screen was really powered off.

v1.7.1: the difference at a glance

A simpler look at the battery-draw readings visible in nine TOP-mode photographs. The bars show the average of the on-screen values in each configuration, with the observed ranges beneath.

Easy-to-read comparison of photographed app values: native TOP with fixes OFF averages 1.86 W across five snapshots (range 1.43 to 2.27 W); True Off plus CPU Fix ON averages 0.97 W across four snapshots (range 0.95 to 0.98 W).

These are snapshots of different fix configurations, not a controlled battery-life test. The separately documented A/B/A benchmark above provides the controlled comparison.

Full A/B/A benchmark and raw measurements · How display power-off was validated

Lightweight in the background

The app's own CPU usage stays tiny, too. In separate physical idle tests, daemon CPU use with both screens active fell from about 0.29% to 0.09% of one CPU core — around 70% less than the earlier RC2 candidate. In TOP mode it remained around 0.32% of a core.

View the full hardware validation dossier · CPU and sleep measurements


Current investigations

We are still looking into the following issues. These are research areas, not advertised fixes in v1.7.1.

  • 🖥️ 120 Hz / screen tearing — Understanding mixed-refresh behaviour when the two Thor displays run at different rates. Research #37.
  • 🔌 Device-service reliability and wake black screens — Investigating rare cases where Android temporarily loses access to a built-in device service, and a separately tracked TOP wake black-screen incident (#84). No cause or fix is established for the blackout.
  • 🎮 TOP-mode game focus — Checking reports of games losing focus when the bottom screen is off.
  • ⚡ Further efficiency and reliability — More hardware edge cases, safer recovery and precise profiling. The engineering hardening PR #67 is merged; stable v1.7.2 is not yet published.

Interested in future device support, including Retroid Pocket Duo? That's research only and requires separate volunteer hardware testing — it is not supported today. Full roadmap.


How to use

  1. Download Jesty Thor Fix v1.7.1 (official signed APK)
  2. Install and open Jesty Thor Fix
  3. Enable the controls you want
  4. Close the app (or swipe it away). The fixes continue in the background.

The app uses the Thor’s own privileged bridge. No Magisk or root required.

CPU Fix note: Turning this fix on or off restarts the Android interface once. Open apps may close, and the screen can briefly go black. This is not a full device reboot.


Safety-first design

Built for stock AYN Thor hardware — no overclocking, firmware flashing, Magisk or manual rooting required.

  • No CPU or thermal tuning: does not write CPU frequencies, voltages or thermal limits.
  • Verified changes: checks the hardware display state after applying the fix.
  • Guarded recovery: restores normal behaviour when validation detects a problem.
  • Optional protection: Closed-Lid Wake Guard is off by default and can be enabled only if wanted.

The safety mechanisms, physical tests and known limitations are documented in the validation evidence and architecture.


Compatibility

Device Status
AYN Thor Supported (tested on Android 13 firmware TKQ1.231222.001)
Retroid Pocket Duo Research / testers wanted — not supported yet
Other devices Unsupported

Support the project

Jesty Thor Fix is free and open source. No features are locked behind donations.

Buy Me a Coffee

  • ⭐ Star the repository so other Thor owners can find it
  • 🧪 Share results from another firmware
  • 🐛 Report reproducible issues
  • 💬 Join the original Reddit discussion

Documentation

Physical validation & performance evidence · Public evidence policy · Architecture · Diagnostics · Diagnostics RC validation · Compatibility · Benchmarks · What's new in v1.7.1 · Build from source · Release integrity

Technical implementation

Jesty Thor Fix follows the physical TOP/BOTH switch. It waits for Android and both displays to settle, then verifies the hardware state (CRTC).

The CPU Fix applies the tested Thor-side setting and performs its required one-time Android display-component recovery.

Closed-Lid Wake Guard performs a guarded sleep request only after verifying the current lid and device state.

See Architecture for the full service, daemon, verification and boot behaviour.

License, provenance & independence

Independent community project. Not affiliated with or endorsed by AYN.
Hardware claims are based on physical device evidence.

About

True hardware-level bottom display off, wake repair, and live CPU/display telemetry for the AYN Thor.

Topics

Resources

Contributing

Security policy

Stars

6 stars

Watchers

2 watching

Forks

Releases

Sponsor this project

Packages

Used by

Contributors

Languages