RESEARCH DIRECTORATEPX–TB–027
REVISION 1 / PUBLIC EDITION
PARALLAX / Technical brief
The earlier clock
The controller had not spoken. The hydraulic plant had already moved. Three retained windows distinguish the two events.
Tomas Eide / Field Operations; PARALLAX timing review
SOURCE DATE 2026.07.13 / RELEASE DATE ABOVE
01 / The omitted channel
The independent accelerometer records a small mechanical onset 96 milliseconds before the command in the original window. A plant-pressure channel records preparation in the same interval. E2-A responds 14 milliseconds after that local onset. The control-room display used the command as time zero and omitted the earlier physical input. Its timestamps were accurate; its account of what began the event was incomplete.
02 / Timing changed
With plant preparation moved earlier and the main command held at zero, a retained comparison window places the precursor at −211 milliseconds and the panel response at −195 milliseconds. The response follows the precursor by 16 milliseconds. This intervention supports a local-input explanation for the apparent anticipation. It does not identify every physical feature to which the material responds.
03 / Preparation isolated
In a retained window with the preparation path mechanically decoupled, no precursor is detected above the witness threshold. The command occurs at zero, the main load at +14 milliseconds, and the panel response at +32 milliseconds. Relative to the first detected local load, the response follows by 18 milliseconds. “No precursor detected” is the scope of the observation; it is not proof that every possible earlier signal was absent.
04 / What the public extract contains
The evidence desk presents these three selected windows. The nominal values are rounded to the nearest millisecond; the estimated uncertainty of each reported offset relative to the command is ±2 milliseconds. A conservative uncertainty bound for a difference between two offsets is ±4 milliseconds. The extract is not a count of all trials, a detection rate, or a complete waveform. Changing the displayed reference subtracts the same origin from every event. It changes no event order and cannot turn a response into a prediction.
05 / Remaining work
The lead on the command clock is explained sufficiently to withdraw the anticipation interpretation. The commissioning team still cannot reproduce E2-A’s exact response in a newly fabricated panel. That is a different question. Its examination must not inherit the drama of the clock error or assume that a corrected display has explained the material.