MAXWELLX RESEARCH WORKBENCH — USER GUIDE
Version 1.1.0 · 21 October 2026

All tools run locally. Browser refresh clears imported raw data. Changes are temporary unless explicitly exported. There is no raw-data upload endpoint.

1. IDEAL CASIMIR
Start at 100 nm and 1 mm². Double the gap: pressure falls by a factor of 16. Negative sign denotes attraction. Vary uncertainty and inspect local spring stability. JSON exports parameters/results; CSV exports sweep rows; SVG exports the first plot. Perfect conductors, zero temperature, no patches/roughness/finite-edge corrections.

2. GEOMETRY & MATERIALS
Sphere–plane PFA requires d/R ≤ 0.01, a policy restriction rather than a quantified error bound. Planar finite-temperature mode offers ideal, Drude or plasma scenarios. Run calculation after changing inputs. Two cutoff/tolerance settings produce a convergence difference. Materials are analytic scenarios, not measured optical datasets. Arbitrary geometry uses provenance-bearing external JSON imports, not a hidden built-in solver. Download the schema and replace every example field.

3. ENERGY & MOMENTUM
Use one boundary and interval. Blank channels remain unknown. Enter zero only when justified. Stored-energy change may be negative. Default unknown heat intentionally prevents a complete balance. Canonical time-series CSV: time_s,input_W,output_W,loss_W,stored_J. All cells required; strict increasing timestamps. Sampling intervals are taken as intentional; gaps are not automatically inferred. Photon recoil is a collimated-emission benchmark, not reactionless propulsion.

4. PHOTON MODES
Change frequency and temperature to see thermal occupation. Pair flux is your assumption, not predicted yield. Pair output exceeding pump input is flagged. No coupling, cavity quality factor, detection efficiency or photon generation rate is modeled.

5. HELIOS DATA LAB
Try Load synthetic teaching pair. Confirm the same passive load for this demonstration and keep maximum gap 20 s. Analyze locally. Primary energy is approximately 0.000574 J, coverage 64.5%, with three excluded intervals. Compare with the synthetic blank. These are teaching files, not MaxwellX measurements.
For actual data: choose a CSV, explicitly map time/voltage/current/run ID/condition, set units, choose timestamp format and measurement mode, define the maximum gap, then analyze. ISO timestamps require Z or an offset. Limit: one run, 50,000 samples, 5 MB per file. Missing numeric values break integration. Non-finite values, duplicate times and multiple run IDs are rejected. Comparison uses the same mapping; review matched intervals, coverage, polarity and conditions. Signed energy is not biological attribution. Biased half-cell results are terminal energy; bias input must be measured separately. Exports record raw SHA-256 and mapping. Clear data removes application references, not forensic memory erasure. The blank template intentionally has missing measurements.

6. TANK → HOME
Default 0.1% conversion is hypothetical, not measured Helios efficiency. Generation is insufficient for the assumed auxiliaries/home. Scrub the hour to inspect stored and served energy; artwork changes day/night. Import 24 hourly irradiance/load rows or use default day. Initial battery energy counts as a resource. Sensitivity uses ±20% one parameter at a time. Hourly electrical balance includes losses, curtailment and unmet auxiliary/home demand. Eight-foot height does not determine projected collecting area or conversion.

7. NO-SIGNALLING
Move the relative Bloch-sphere angle and visibility. Bob remains 50/50. Reveal Alice's classical record to inspect conditional correlations. Exact probabilities, not collected photons. Conditioning is not a faster-than-light channel. General local Kraus-channel checks are included in the Python reference.

8. EVIDENCE
Search title/author/country/claims. Filter year and publication status. Expand each record for access depth, locator, notes and replication status. Nineteen curated sources, ten dated 2026; not a comprehensive systematic review. Missing affiliations/methods stay unknown.

EXPORTS
JSON includes parameters, version and result. CSV contains result rows or metric/value pairs; spreadsheet-formula strings are escaped. SVG saves the first available line plot. Model parameters and evidence filters can be saved/reloaded. Helios mappings travel with analyzed results and must be reconfirmed when raw files are imported. Download full source and the reference package from the workbench.

PHYSICAL LIMITS
No actual MaxwellX raw measurements were provided. No experimental replication was performed. No autonomous useful power, reactionless propulsion or faster-than-light result is established. Measured optical data, film stacks, arbitrary-geometry solvers, correlated metrology uncertainty and confirmatory biological statistics remain future integrations.