REVIEW 3 major objections 7 minor 6 references
The perfomance of generalized Davies-Cotton optical systems with infinitesimal mirror facets
T0 review · 3 major / 7 minor · reviewed 2026-08-12 · deepseek-v4-flash
Pith's one-line read A spherical telescope's curvature ratio $c=F/R$ can be dialed between minimal coma at $c=3/2$ and near-perfect timing at $c=1/2$.
desk verdict A useful analytic generalization of Davies-Cotton optics to arbitrary curvature ratio, internally consistent but short of an independent physical check on its light-collection model. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing object is the moment functional $I[f]=\int f(\varphi,\theta)\,|\vec g\cdot\vec n|\,R^2\cos\theta\sin\theta\,d\theta\,d\varphi$ of Equation (2), applied to the reflected-ray coordinates and propagation times. Two ingredients distinguish this analysis from earlier work: the per-facet alignment point $\vec r_A=(0,0,z_A)$ from Equation (1), which suppresses spherical aberration for any $R$, and the $|\vec g\cdot\vec n|$ projection factor, which models rays falling into the cracks between facets. The paper substitutes the ray coordinates $x_{fp}$, $y_{fp}$ and the path time $t_{fp}$ into this functional, expands in $\delta$ and $1/(4f^2)$, and identifies the ratio $c=F/R$ as the single parameter controlling coma, plate scale, and timing dispersion.
What would settle it
Ray-trace a $D=10$ m, $F=12$ m tessellated telescope with finite hexagonal or square facets at $c=1/2,1,3/2$, and compare the effective area, image RMS, and arrival-time RMS to Equations (3)--(7) as the facet size shrinks; a systematic disagreement that grows with facet tilt would falsify the $|\vec g\cdot\vec n|$ light-loss model.
Extended reading notes
Core claim
The central claim is that for a spherical reflector with infinitesimal facets laid out on a regular grid, all first and second moments of the focal-plane image and of the arrival-time distribution can be written as Taylor series in the off-axis angle $\delta$ and in $1/(4f^2)$, where $f=F/D$. The derivation generalizes the classic Davies-Cotton constraint $R=F$ to arbitrary $c=F/R$, and adds a projected-area weight $|\vec g\cdot\vec n|\,dS$ that accounts for rays lost in the gaps between tilted facets. The resulting expressions, Equations (3)--(7), show that spherical aberration is suppressed for every $c$, that the leading tangential coma term is minimized at $c=3/2$, where it equals the sagittal coma to two leading orders, and that the on-axis arrival-time variance vanishes to leading orders at $c=1/2$. At $c=1$ the new tangential-width formula differs from the earlier VFB expression by a small term of order $0.3\%/f^2$, a direct consequence of the facet-edge light loss.
Load-bearing premise
The derivation assumes each infinitesimal facet collects or loses light exactly in proportion to $|\vec g\cdot\vec n|\,dS$, with no obscuration, diffraction, or finite-facet effects; if real gaps between tilted mirrors follow a different light-loss law, the predicted optima at $R=2F/3$ and $R=2F$ would shift.
Editorial extensions
If this is right
- A designer can choose any $c$ between $1/2$ and $3/2$ to trade off-axis image quality against arrival-time spread; for a $D=10$ m, $F=12$ m telescope the maximum sag changes by roughly a metre across this range, so the structural cost is modest.
- Even an ideal infinite-facet telescope loses light to facet tilt: the effective area is $\pi(D/2)^2\cos\delta\,(1-1/(64f^2)-\cdots)$, about $1.6\%/f^2$ below the canonical aperture area.
- At $c=3/2$ the leading plate-scale correction vanishes and the tangential coma term drops to two-thirds of its classic Davies-Cotton value, while tangential and sagittal coma widths match to two leading orders.
- At $c=1/2$ a spherical dish approximates a parabola for timing, with residual on-axis dispersion suppressed to order $1/f^6$.
- Equations (3)--(7) give closed-form targets for validating custom ray-tracing simulation codes without finite-facet noise.
Reading between the lines
- If real facet layouts collect or block light differently than the $|\vec g\cdot\vec n|\,dS$ model, the two optima would shift; a natural extension is finite-facet ray tracing across facet shapes and sizes.
- The same parameter likely applies to non-imaging concentrators such as solar collectors, where timing is irrelevant but the effective-area and imaging expressions transfer directly.
- The flat $c=0$ limit, which the paper likens to a reflective Fresnel lens, suggests a testable wide-field timing reflector; its coma would be poor but its timing spread stays within twice the Davies-Cotton value.
- The equality of tangential and sagittal coma at $c=3/2$ suggests that a dish near $R=2F/3$ could give rounder off-axis point-spread functions than the elliptical-compromise designs discussed in the wider literature.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. This manuscript extends the analytic treatment of Davies-Cotton (DC) telescopes by Vassiliev et al. (2007) to arbitrary values of the ratio c = F/R between focal length and mirror radius of curvature, in the idealized limit of infinitesimal mirror facets. The principal new ingredient is Eq. (2), which weights each aperture element by the projected-area factor |g·n| in order to model rays lost through the gaps between tilted facets. The paper presents Taylor expansions for the effective area (Eq. 3), image centroid (Eq. 4), tangential and sagittal image widths (Eqs. 5 and 6), and arrival-time variance (Eq. 7), and identifies two privileged configurations: c = 3/2, which minimizes tangential coma, and c = 1/2, which largely eliminates timing dispersion. The expansions are compared with numerical quadrature of Eq. (2) for D = 10 m, F = 12 m and field angles up to 10 degrees, with agreement better than about 1% except at the two optima, where the errors are larger and are acknowledged in the text.
Significance. If the advertised results are correct, they provide a useful parameter-free design tool for segmented imaging atmospheric Cherenkov telescopes, and they quantify a physical effect—loss of rays in the cracks between tilted facets—that is absent from the earlier VFB and Bretz-Ribordy expressions. The analytic derivation is self-contained, contains no fitted parameters, and is checked against direct numerical integration of the same model. The paper is also honest about its scope: §4 explicitly states that finite-facet-size effects are not studied and that ray-tracing simulations are required for the detailed response. The inclusion of a Python notebook implementing both the quadrature and the expansions, together with a MATLAB script for generating further terms, is a reproducibility strength.
major comments (3)
- [§2, Eq. (2)] The entire generalization beyond VFB rests on the weight |g·n| dS with dS = R² cosθ sinθ dθdφ, and the two central optima at c = 3/2 and c = 1/2 are derived from integrals of this weight. The text calls dS a surface area, but this is actually the aperture-projected area element, not the physical spherical surface area of the mirror. More importantly, the manuscript does not derive this weight from a limiting procedure on a regular grid of finite planar facets, and it does not test the weight against any independent ray-tracing simulation: Fig. 1 and the notebook only compare the Taylor expansions with numerical quadrature of Eq. (2) itself, so they cannot validate the physical content of the loss law. Because §4's disclaimer about finite-facet effects does not by itself guarantee that Eq. (2) is the correct continuum limit of a faceted telescope, the paper should add either a derivation of Eq. (2) from the facet-layout geometry (including the no-obscuration assumption) or a small-facet ray-tracing check in the limit of many facets for the three values of c used in Fig. 1.
- [§3, Fig. 1 and Eq. (7)] The accuracy of the expansions is not uniform over the parameter values that are the paper's main conclusions. At c = 1/2, the error in σt exceeds 50% at δ = 10°, and at c = 3/2, the errors in σx and σy reach about 1% at δ = 10°. These are exactly the two 'optima' highlighted in the abstract and in §4. The text discloses these errors, but the abstract and §4 present c = 1/2 and c = 3/2 as design-relevant results without this caveat. The paper should give an explicit domain of validity for the timing and imaging claims (for example, restrict the δ range, or state the number of additional Taylor orders needed at the optima), so that the central claims are not stronger than the demonstrated accuracy.
- [§4, Discussion of design trade-off] The proposition that c can be used to trade imaging resolution against timing dispersion, with c = 3/2 as the minimum-coma configuration, is made for a model that explicitly assumes no obscuration just before Eq. (2). For the example D = 10 m, F = 12 m, the c = 3/2 case has a maximum sag of about 1.76 m, i.e. a rather deep dish, and mutual shadowing of facets for off-axis rays grows with dish depth and field angle. This effect can change the effective area and the moments even for small facets, and the manuscript offers no estimate of its magnitude. The paper should either give a first-order estimate of obscuration for the deep-dish configurations or state more forcefully that the optimum is a prediction of the no-obscuration infinitesimal model that requires ray-tracing verification before being applied to a real telescope.
minor comments (7)
- [Title] The title contains a typo: 'perfomance' should be 'performance'.
- [Abstract] The sentence 'Configurations that range between two "optima" are, one of which minimises tangential comatic aberration and the other that minimises timing dispersion' is ungrammatical; it should be rephrased as two complete clauses, for example: 'Configurations range between two optima: one minimises tangential comatic aberration, and the other minimises timing dispersion.'
- [§2, before Eq. (2)] The phrase 'Calculation the moments' should read 'Calculating the moments'.
- [§2, Eq. (2) and surrounding text] Calling R² cosθ sinθ dθdφ the 'surface area' of the infinitesimal mirrors is misleading; this is the area element projected onto the aperture plane. Please rename it consistently as the grid-cell area or aperture-projected area element.
- [§2, Eq. (8)] The refractive index n in the timing expressions is not defined in the text; state that n ≈ 1 for air and clearly state the units assumed in the prefactor 1.2036 ns of Eq. (8).
- [§3, Fig. 1] The panel labels use 'F/R=3/2', 'F/R=1', and 'F/R=1/2', whereas the paper defines c = F/R; using c values consistently in the figure would avoid confusion.
- [§2, Eqs. (4)–(7)] The derivation of the Taylor expansions is not shown. The numerical notebook is helpful, but one representative derivation or a short appendix describing the integration and expansion procedure would make the algebra verifiable by readers and referees.
Circularity Check
No significant circularity: the analytic expansions are self-contained geometry, with only a model-level assumption (the |g·n| crack-loss weight) that is openly stated and not fitted to the target predictions.
full rationale
The paper derives Taylor expansions for the moments of the image and arrival-time distribution of a generalized Davies-Cotton telescope directly from the geometric setup and the stated projected-area weighting in Eq. 2. No parameter is fitted to data and then renamed a prediction. The only author-overlapping citation is to VFB, but it is used for notation and methodology, not as a load-bearing premise; the present results generalize and correct VFB by adding the |g·n| crack-loss term, and the final expressions are obtained by explicit integration in the paper. The numerical quadrature in Figure 1 integrates the same Eq. 2, so it verifies the algebra of the Taylor expansions rather than independently validating the physical light-loss model. This is not circular: the paper does not claim the quadrature validates the model, and it explicitly states that finite-facet effects are not studied and that ray-tracing simulations are required for the detailed response. The optima at c = 3/2 and c = 1/2 are mathematical consequences of the assumed projected-area weighting; whether that weighting correctly describes real segmented reflectors is an open modeling question, not a circularity in the derivation. Accordingly, no circular step is identified and the score is 0.
Assumptions & free parameters
assumptions (4)
- domain assumption The VFB reflection geometry and moment-integral method are correct and can be extended to arbitrary c = F/R.
- domain assumption The collecting area of an infinitesimal facet is |g·n| dS with dS = R² cosθ sinθ dθdφ, which models rays lost through facet cracks.
- ad hoc to paper The Taylor expansions truncated at the stated orders in δ and 1/(4f²) are accurate enough for design use, without a proven general error bound.
- domain assumption There is no obscuration by the camera or support structure, and no diffraction or finite-facet-size effects.
Cite this review
Pith. "Pith review of The perfomance of generalized Davies-Cotton optical systems with infinitesimal mirror facets." pith.science (2026). https://pith.science/paper/5QMSVY4H
@misc{pith2026241116434,
author = {Pith},
title = {Pith review of: The perfomance of generalized Davies-Cotton optical systems with infinitesimal mirror facets},
year = {2026},
howpublished = {\url{https://pith.science/paper/5QMSVY4H}},
note = {Machine review of arXiv:2411.16434}
}
read the original abstract
This paper presents Taylor expansions for the imaging and timing characteristics of spherical optical systems with infinitesimal mirror facets, sometimes referred to as ''modified Davies-Cotton'' telescopes. Such a system comprises a discontinuous spherical mirror surface whose curvature radius is different from its focal length, and whose mirrors are aligned to suppress spherical aberration. Configurations that range between two ''optima'' are, one of which minimises tangential comatic aberration and the other that minimises timing dispersion.
Figures
Reference graph
Works this paper leans on
-
[1]
author Bretz , T. , author Ribordy , M. , year 2013 . title Design constraints on Cherenkov telescopes with Davies-Cotton reflectors . journal volume 45 , pages 44--55 . :10.1016/j.astropartphys.2013.03.004
-
[2]
author Davies , J. , author Cotton , E. , year 1957 . title Design of the quartermaster solar furnace . journal Journal of Solar Energy, Science and Engineering volume 1 , pages 16--22
work page 1957
-
[3]
author Fazio , G.G. , author Helmken , H.F. , author Rieke , G.H. , author Weekes , T.C. , year 1968 . title An experiment to search for discrete sources of cosmic gamma rays in the 10^ 11 to 10^ 12 eV region. journal Canadian Journal of Physics volume 46 , pages S451--S455
work page 1968
-
[4]
author Garczarczyk , M. , et al., year 2015 . title Status of the Medium-Sized Telescope for the Cherenkov Telescope Array , in: booktitle Proc. 34th International Cosmic Ray Conference (ICRC2015)
work page 2015
-
[5]
Wide-field prime-focus Imaging Atmospheric Cherenkov Telescopes: A systematic study
author Schliesser , A. , author Mirzoyan , R. , year 2005 . title Wide-field prime-focus imaging atmospheric Cherenkov telescopes: A systematic study . journal volume 24 , pages 382--390 . :10.1016/j.astropartphys.2005.08.003, http://arxiv.org/abs/astro-ph/0507617 arXiv:astro-ph/0507617
work page Pith review arXiv 2005
-
[6]
author Vassiliev , V. , author Fegan , S. , author Brousseau , P. , year 2007 . title Wide field aplanatic two-mirror telescopes for ground-based gamma-ray astronomy . journal volume 28 , pages 10--27 . :10.1016/j.astropartphys.2007.04.002
Reviewed August 12, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.