Starting DirectConvert engine

Preparing…

DirectConvert Research Preview

MODEL, NOT MACHINE. Public outputs are research simulations of how a charged-particle spectrum might convert to electricity. No fabrication dimensions, insulation stacks, high-voltage operating procedures, vacuum-system guidance, or component sourcing. All physics runs in your browser with the real directconvert_core kernel — no science is reimplemented in JS.
Starting…

What each parameter means

Every input carries a parameter-source tag so you can see what is a user assumption vs a literature/dataset value. Assumed loss factors are labelled unreviewed and are never engineering predictions. Change any input and the result is marked OUTDATED until you re-run.

Spectrum shape Model default — illustrative, not evaluated data.

Run a scenario to see the spectrum support, mean energy and content hash.

Charge +2e; a stage at V collects a particle iff q·V ≤ E.

DC-L1 loss factors. Each shows a parameter-source tag; all are assumed and unreviewed. Provide a standard deviation to include it in the uncertainty study.

OUTDATED — inputs changed since this result was computed. Re-run before exporting (export disabled).

Set a scenario and run. The energy ledger, efficiency, warnings and provenance appear here.

Morris elementary-effects ranking of how strongly each assumed loss factor moves η.

Uses the per-channel standard deviations from the Loss models tab (Normal distribution).

Run the study to see the median efficiency, 5th–95th percentile interval, exceedance probability, histogram and convergence.

Equations

  • EQ-DC-STAGE-BOUND — recovered ≤ q·V, with q·V ≤ E ▸ record
  • EQ-DC-ETA — collection-efficiency integral ▸ record
  • EQ-DC-BALANCE — exact energy accounting ▸ record

Sources

  • Barr & Moir — Direct Energy Conversion review Curator-reviewed
  • Rax et al. (2025) — Adiabatic direct energy conversion Curator-reviewed

Reviewer state: unreviewed. See limitations.

Export levels

Quick export

Input + output JSON and CSV tables for the current run.

Reproducibility package

Complete manifest, Python reproduction script, environment lock, checksums, equations, sources, warnings and RO-Crate.

See example package →

Research release

Everything above plus DataCite metadata, review decisions, a licence, a release manifest and an immutable identifier.

Releases →

Export is disabled while a result is OUTDATED. Same inputs always produce the same canonical run hash.

DC-L2 trajectory — validation matrix

DC-L2 (electrostatic particle trajectories) is institutional-only and not exposed in this public tool. Status of its validation:

TestStatusMeaning
Energy conservationPassedSolver preserves declared energy accounting
Time-step convergencePassedOutput stabilizes as the time step decreases
L2-to-L0 ideal limitPassedL2 respects the ideal upper-bound model (q·V ≤ E)
Analytic field testRequiredCompare trajectory against an analytic solution
Independent solver comparisonRequiredCompare with a separate trajectory code
Reviewed field mapRequiredConfirm imported field data and units
Experimental comparisonFutureCompare with institutional measurements
External scientific reviewRequiredIndependent specialist approval

DC-L2 stays restricted

The public interface allows
  • Educational geometry abstractions
  • Normalized particle distributions
  • Safe parameter exploration
  • Numerical convergence demonstration
It does not provide
  • Hardware fabrication dimensions
  • Insulation construction instructions
  • High-voltage operating procedures
  • Vacuum-system construction guidance
  • Component sourcing for dangerous systems