Drawing set HR-001 · Rev A
Annular multi-pass photovoltaic collector. A thin glass guide, photovoltaic segments on the inner wall, and a launch that can hold rays by total internal reflection.
Not a device claim. Under full-beam illumination this ring still collects less than an equal-area flat absorber facing the beam. The drawings below keep that result in the ledger.
HR-001 / A · live sample, not the Monte Carlo
Seven rays across the facet. Impact parameter 6.96–8.19 mm. Outer TIR needs b > 6.67. Inner wall needs b < 8.5. Index 1.5.
Gold rays reach a copper segment. Dim rays leave through a facet or the outer wall. This sample illustrates the routing condition. Sheet 03 holds the 16,000-ray and full-beam ledgers. Turning on the coat, fibers, or shell changes the drawing only — those layers were not in this live trace.
Why a smooth ring fails, and what the thin guide fixes
A ray that refracts in through the outer wall keeps an incidence angle too small for later total internal reflection on that same wall. Specular hits on a circle preserve the impact parameter. Most of that light leaves after one or two passes.
b > Rout / nouter-wall trapping
b < Rinreaches the inner absorber
Both can be true only if the inner radius is larger than the outer radius divided by the index. At n = 1.5 and R = 10 mm, the inner wall must sit outside 6.67 mm. A 6 mm bore has no overlap. 8.5 mm does.
A strictly tangential launch has impact parameter equal to its entry radius, so it misses the inner wall. The working path is an inward tilt: 35° inside the glass, which is 59.36° outside the facet. Every ray from that facet then sits in the open window.
At normal incidence on the same facet the window closes and patch absorption falls to zero. The mechanism is angular, not omnidirectional.
Idealized 2D · lossless glass · perfect patches
Fraction of power that hits the 1.5 mm injection facet. Not whole-device collection. 16,000 rays.
| Destination | Share |
|---|---|
| Ideal inner patches | 85.17% |
| Entrance reflection | 8.61% |
| Escape, both open ends | 6.21% |
| Outer-wall escape | 0% |
Patch absorption as a share of power on the facet. Samples, not a fitted curve.
Every ray across the projected width is counted. The flat reference has the same absorber length as the patches and faces this beam. The ring does not beat it.
| Configuration | Both sides | Glass side only |
|---|---|---|
| Equal-area flat, facing the beam | 50.07% | — |
| Ring, inner mirrors only | 43.45% | 15.76% |
| Add a rear mirror, ideal | 45.75% | 18.06% |
| Rear mirror at 98% | 45.43% | 17.76% |
| Mirror the entire outer wall | 34.40% | 6.73% |
About two thirds of the two-sided ring absorption is light that first hits the exposed inner arc from the open center, not light injected through the facet. Covering the whole outer wall blocks more entrance than it saves. A weak gray absorber (10% per encounter) is the case where recycling helps: about 6.9% in the guide versus 5.0% on the equal flat — still not a cell with its own antireflection.
On the drawing · not yet in the power ledger
A
A fluorescent layer absorbs selected sunlight and re-emits inside the glass. Emission born in the guide can enter trapped paths that refraction through the outer wall cannot. This is a luminescent concentrator, not a self-running glow. Quantum yield, Stokes shift, and escape still have to be counted. Visible brightness is light leaving toward the eye.
B
Non-collecting stretches of the inner arc are already ideal mirrors in the model. Additional mirrors return a miss for another pass. They do not pull photons in. A selective rear shell raised full-beam absorption from 43.4% to about 45.4%. A sealed outer shell lowered it, because the entrance closed.
C
Clear fiber only accepts a designed end. Fluorescent fiber absorbs along its length and guides a fraction of the new emission to collectors at both ends. Arcs inside the ring would give a defined route without requiring a complete circuit. Bend radius, cladding, and end coupling are unspecified at this 10 mm scale.
Next drawings, not next adjectives
Illuminate the whole footprint, not only the facet, and keep the equal-area flat cell as the reference.
Replace ideal patches with a measured semiconductor interface, both faces, including the reflection the model currently swallows.
Weight an AM1.5 spectrum and a stated direct-plus-diffuse sky. One design angle is not a year.
Put the luminescent coat and fiber arcs into the same power ledger, with quantum yield and self-absorption.
Report net electrical power, not optical power entering a perfect absorber.