Return delay and acquisition order.

Question and rationale

Does the small return residual after brief source activation follow elapsed return time when acquisition order and schedule position are balanced? The preceding experiment found median control-adjusted return differences of +1, 0 and 0 counts at added return waits of 0, 4 and 32 microseconds in both of its blocks. This experiment holds departure wait fixed, samples the early return interval more closely and gives each delay equal representation in both pair orders and every schedule position.

The wider objective is sovereignty over all available silicon components and their interconnections: SEIGRASM discovery, direct control and observation, replacement of inherited firmware dependencies, and Hyphos/SeigrOS use of discovered electrical capabilities through native Seigsit relations. This experiment tests one component's dependence on preceding actions. Its source-enable settings and delays are not assigned Seigsit states.

Method

Twelve blocks each contain six adjacent pairs. Every pair contains a disabled–enabled–disabled path and a disabled–disabled–disabled control. Return waits are 0, 1, 2, 4, 8 and 32 microseconds; departure wait is always zero. Rotating the delay order and reversing pair order gives 72 pairs, 144 sequences and 432 commanded conversions. Each wait occupies every position once in each pair order. Allocation is deterministic.

Each sequence establishes the disabled reference, waits at least 1,000 microseconds and admits the stopped converter on channel 4 with zero divider, disabled empty FIFO and disabled ADC interrupts. That baseline wait does not establish complete settling. Sixteen native accesses record a reference conversion, a departure source write and conversion, then a disabled return write and conversion. Both source writes occur even in controls. Each conversion records its command, readiness, RESULT and CS. Every access retains values, counts, status and counter intervals.

No SD transaction occurs inside a sequence. Recording and admission occur between sequences, including between pair members. Each member therefore has its own reference. The paired return contrast is (changed return − changed reference) − (control return − control reference); the away contrast uses the corresponding away readings. Current and sticky conversion errors are retained without automatic retry. Guarded restoration follows completed execution.

The instruction streams, ADC channel, clock settings and mapping are unchanged from the preceding experiment. Descriptor scheduling and the evidence reader were extended. The frozen protocol contains the complete schedule, access sequence and dependencies. Acquisition calls no firmware service; entry firmware, processor execution, SD transport and retained bridge configuration remain apparatus dependencies.

Assessment fixed before return

Check identity, completeness, source readbacks, readiness, errors, surrounding controls and restoration. Retain all individual results. Compare each wait's twelve paired contrasts by order, block and position, with away contrasts and reference drift alongside. Compare each early wait with the 32-microsecond contrast within its own block, using the recorded timing bounds.

A return-time relation recurring in both orders would motivate separation of source history from converter history. Order or position dependence would motivate an acquisition-history control. If the previous small residual did not recur consistently, the fine-delay comparison would close as unresolved at the available resolution. No response sign or magnitude was required, and no further sweep was prescribed merely to obtain a preferred outcome.

Preparation and acquisition

Preparation 284BD3F0F5FD froze 31 source inputs and passed 56 image-bound supplied-response checks before installation. All 4,096 image chunks and two full image readbacks were verified. The prescribed allowance was four minutes. The returned coordinator interval is 186.996197 seconds, excluding entry loading and the terminal record's write/readback; powered duration is not independently measured by the capture.

Results

All 4,768 records, 817 operations, 144 sequences and 432 conversions completed. All 288 source-write readbacks matched. No current or sticky conversion errors or admission failures were reported. ADC and clock restoration passed.

Changed paths departed by +116 to +119 counts and returned within 0 to +3 of their references. Controls had away and return differences of −4 to +2. After subtracting each adjacent control, away contrasts span +114 to +121 and return contrasts −1 to +7. All 71 return contrasts after the first pair lie between −1 and +2; the first pair remains included in every summary.

The first pair

The largest contrast is +7 counts at zero added wait in block 1. Its control records 748 → 744 → 744, while the changed path records 741 → 858 → 744. Both return readings are 744. Their local return differences are −4 and +3; subtracting them gives +7. The large paired contrast therefore includes a difference between the two local references, rather than a seven-count separation between the return readings.

All associated records, commands, source readbacks and status flags pass their checks. The first reference is also the highest disabled-source reference in this run. This is a specific observation to investigate, not grounds to discard the pair or identify a startup mechanism.

Order, block and schedule position

At zero added wait the return-contrast median is +1 when the control runs first and 0 when the changed path runs first. At 32 microseconds the corresponding medians are −1 and 0. The 0-versus-32 contrast within each block is positive in all six control-first blocks, but has mixed signs or ties in changed-first blocks. The same lack of agreement between orders appears at other early waits. This does not establish a delay-only response that is independent of acquisition order.

Pooling waits within each schedule position gives a median of 0 at all six positions. That pooled statistic does not exclude interactions with delay or order. Pair order is assigned by block, so order, preceding actions, inter-sequence recording and gradual drift are not independently isolated by this run.

Timing and reference drift

The 0, 1 and 2-microsecond conditions have overlapping access bounds. They cannot be treated as independently resolved internal conversion times. Departure-to-return write bounds span 15.037–21.852 microseconds despite the fixed zero added departure wait. All accesses report a 54 MHz counter.

Disabled-source reference medians fall from 740.5 in block 1 to 723 in block 12. There are twelve references per block. This drift and the separate references within each pair matter at the one-count scale; their physical cause is unmeasured.

Individual sequences and pairs

Case, pair, block and schedule-position numbers below are one-based. Every block includes its twelve sequences and six paired contrasts. Values are raw counts; no observations are omitted. C means unchanged control; X means source activation followed by deactivation. Departure wait is zero throughout.

Interpretation and next question

Direct source interventions again produce a large departure followed by a return close to the local reference. The finer delay schedule does not establish a consistent residual relation across both pair orders. The comparison therefore closes with the small residual unresolved at this measurement resolution, as specified before return. Selecting a longer delay from this result would not identify the cause.

The next discriminating investigation should separate source history from conversion history: vary whether the source was activated and whether an intervening conversion was commanded, while matching the elapsed intervals and retaining unchanged-source controls. Repeated disabled-source observations before any intervention should expose the behavior of the first reference rather than silently replace it. Compact recording should retain every action and observation, so inter-sequence storage effects can be distinguished from actions inside the controlled interval. The subsequent source and conversion history experiment implements this comparison and reports its own methods and results.

This component-level knowledge must support the wider cluster-sovereignty work: determine which electrical relations SEIGRASM can establish and maintain, which interconnected components affect them, and which inherited control dependencies must be replaced. The present source controls do not establish six distinct maintained electrical states, native physical Seigsit creation or transport, independently fresh analog conversions, replacement firmware or applicability to another cluster.

Integrity, restoration and provenance

Two read-only SD captures agree. The independent reconstruction verifies all senary record fields, integrity checks, commands, readbacks, observations, pair identities and interval bounds. Unused journal reservations match preparation. Restored ADC control and divider are zero, interrupts disabled and FIFO empty; clock control is zero with divider/select retained at 1.

Partition table, geometry and both boot files match preparation; FAT copies agree. Seven sectors outside the journal differ, including System Volume Information. Their writer and timing are not established.

Capture SHA-256: db54b6df727b5daff9da1923c0d3f7a10061f2f012a59b6961ddd7439b17db6b. Image SHA-256: 0b1bbda097a463ed4ac38188e833bb08b194d93e77e73706b47dc2b2a75239ec.