The home ultrawide had been running at 60 Hz for weeks, and every reconnect made it worse-looking, not better. Cause is in the desk profile, not the monitor:
dotfiles/zz-blankpad/niri/.config/niri/cfg/desk-1-home.kdl asked the Sceptre M34 for mode "3440x1440@239.999".
The panel's EDID has no 240 Hz mode. niri msg outputs, read live, tops out at 3440x1440@180.000.
niri does not fail on an unavailable mode. It logs configured mode 3440x1440@239.999 could not be found, falling back to "preferred" and drives the EDID preferred mode — 60 Hz. Twenty-eight fallbacks in the current boot, one per connector event.
The wrong number was not invented by the config: omp-docs/tools/niri/displays.md documented 239.999 as the rig spec, in three places, including the copy-paste pattern block. The page was the vector.
Fix
desk-1-home.kdl: Sceptre identity block 239.999 -> 180.000, plus the two legacy connector-keyed blocks (DP-1) aligned to 180 so a connector-named reconnect cannot land on 120; four comment lines recording that the EDID stops at 180 and what a phantom mode costs.
niri validate against the live config: config is valid.
Applied without touching the session: niri msg output "Sceptre Tech Inc Sceptre M34 Unknown" mode "3440x1440@180.000" (the temporary per-output interface — no config reload, no restart). Verified Current mode: 3440x1440 @ 180.000 Hz, compositor PID unchanged, zero new Page flip / DMCUB errors after the change.
Docs
tools/niri/displays.md: every 239.999 -> 180.000; rule 4 rewritten with the real failure mode (silent fallback to preferred, not a dark output); new triage section separating config / DP-link / amdgpu-driver faults by log signature with the exact probes, plus the safe live remedy; wlr-randr corrected to not installed.
Tonight's connector drops (20 since Oct 14) and the DMCUB error after the 23:46 s2idle resume are amdgpu driver state, not config: they need a reboot Frank schedules, and a cable if they survive it. The page now says how to tell the difference. tools/noctalia-shell.md also lists wlr-randr as a noctalia dependency while nothing matching wlr is installed — left alone here because that is a dependency claim, not an inventory claim, and deserves its own check.
The home ultrawide had been running at 60 Hz for weeks, and every reconnect made it worse-looking, not better. Cause is in the desk profile, not the monitor:
- `dotfiles/zz-blankpad/niri/.config/niri/cfg/desk-1-home.kdl` asked the Sceptre M34 for `mode "3440x1440@239.999"`.
- The panel's EDID has no 240 Hz mode. `niri msg outputs`, read live, tops out at `3440x1440@180.000`.
- niri does not fail on an unavailable mode. It logs `configured mode 3440x1440@239.999 could not be found, falling back to "preferred"` and drives the EDID preferred mode — 60 Hz. Twenty-eight fallbacks in the current boot, one per connector event.
The wrong number was not invented by the config: `omp-docs/tools/niri/displays.md` documented 239.999 as the rig spec, in three places, including the copy-paste pattern block. The page was the vector.
## Fix
- `desk-1-home.kdl`: Sceptre identity block 239.999 -> 180.000, plus the two legacy connector-keyed blocks (`DP-1`) aligned to 180 so a connector-named reconnect cannot land on 120; four comment lines recording that the EDID stops at 180 and what a phantom mode costs.
- `niri validate` against the live config: `config is valid`.
- Applied without touching the session: `niri msg output "Sceptre Tech Inc Sceptre M34 Unknown" mode "3440x1440@180.000"` (the temporary per-output interface — no config reload, no restart). Verified `Current mode: 3440x1440 @ 180.000 Hz`, compositor PID unchanged, zero new `Page flip` / `DMCUB` errors after the change.
## Docs
- `tools/niri/displays.md`: every 239.999 -> 180.000; rule 4 rewritten with the real failure mode (silent fallback to preferred, not a dark output); new triage section separating config / DP-link / amdgpu-driver faults by log signature with the exact probes, plus the safe live remedy; `wlr-randr` corrected to *not installed*.
- `tools/applications.md`, `tools/cachyos.md`: dropped `wlr-randr` rows — `pacman -Qs wlr` matches nothing on blankpad.
- `log.md`: dated entry with the laws.
## Not in scope
Tonight's connector drops (20 since Oct 14) and the `DMCUB` error after the 23:46 s2idle resume are amdgpu driver state, not config: they need a reboot Frank schedules, and a cable if they survive it. The page now says how to tell the difference. `tools/noctalia-shell.md` also lists `wlr-randr` as a noctalia dependency while nothing matching `wlr` is installed — left alone here because that is a dependency claim, not an inventory claim, and deserves its own check.
---
cwd: /home/blank/.omp/wt/wt-20261019-125509-7c8a99d · session: 01a155bb-6474-7366-be9d-629f74af0e05
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
The home ultrawide had been running at 60 Hz for weeks, and every reconnect made it worse-looking, not better. Cause is in the desk profile, not the monitor:
dotfiles/zz-blankpad/niri/.config/niri/cfg/desk-1-home.kdlasked the Sceptre M34 formode "3440x1440@239.999".niri msg outputs, read live, tops out at3440x1440@180.000.configured mode 3440x1440@239.999 could not be found, falling back to "preferred"and drives the EDID preferred mode — 60 Hz. Twenty-eight fallbacks in the current boot, one per connector event.The wrong number was not invented by the config:
omp-docs/tools/niri/displays.mddocumented 239.999 as the rig spec, in three places, including the copy-paste pattern block. The page was the vector.Fix
desk-1-home.kdl: Sceptre identity block 239.999 -> 180.000, plus the two legacy connector-keyed blocks (DP-1) aligned to 180 so a connector-named reconnect cannot land on 120; four comment lines recording that the EDID stops at 180 and what a phantom mode costs.niri validateagainst the live config:config is valid.niri msg output "Sceptre Tech Inc Sceptre M34 Unknown" mode "3440x1440@180.000"(the temporary per-output interface — no config reload, no restart). VerifiedCurrent mode: 3440x1440 @ 180.000 Hz, compositor PID unchanged, zero newPage flip/DMCUBerrors after the change.Docs
tools/niri/displays.md: every 239.999 -> 180.000; rule 4 rewritten with the real failure mode (silent fallback to preferred, not a dark output); new triage section separating config / DP-link / amdgpu-driver faults by log signature with the exact probes, plus the safe live remedy;wlr-randrcorrected to not installed.tools/applications.md,tools/cachyos.md: droppedwlr-randrrows —pacman -Qs wlrmatches nothing on blankpad.log.md: dated entry with the laws.Not in scope
Tonight's connector drops (20 since Oct 14) and the
DMCUBerror after the 23:46 s2idle resume are amdgpu driver state, not config: they need a reboot Frank schedules, and a cable if they survive it. The page now says how to tell the difference.tools/noctalia-shell.mdalso listswlr-randras a noctalia dependency while nothing matchingwlris installed — left alone here because that is a dependency claim, not an inventory claim, and deserves its own check.cwd: /home/blank/.omp/wt/wt-20261019-125509-7c8a99d · session: 01a155bb-6474-7366-be9d-629f74af0e05