Pith. sign in

REVIEW 3 major objections 5 minor 17 references

The Measurement of Position Resolution of RD53A Pixel Modules

T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Pith's one-line read A 13-degree tilt improves RD53A pixel resolution to 10.9 and 6.8 micrometers in a test beam.

desk verdict A workmanlike RD53A testbeam measurement whose headline numbers are probably right, but the 3.85 vs 3.35 µm Mimosa26 ambiguity needs to be propagated into the final error budget. read the letter →

arxiv 1908.10973 v3 pith:PGS6GXZY submitted 2019-08-28 physics.ins-det hep-exphysics.data-an

classification physics.ins-dethep-exphysics.data-an PACS 29.40.Gx
keywords positionresolutionRD53ApixeldetectortestbeamtiltanglebinaryreadouttrackHL-LHC
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper measures the intrinsic position resolution of RD53A pixel modules, the prototype detectors for the upgraded LHC tracking systems, using an 11 GeV electron test beam. It reports that tilting the modules by 13 degrees improves the short-pitch direction: the 50x50 module reaches 10.9 micrometers in the tilted 50-micrometer direction, and the 25x100 module reaches 6.8 micrometers in the tilted 25-micrometer direction, gains of roughly 26% and 14% over the non-tilted values. The resolutions are derived by subtracting a simulated track resolution from measured unbiased residuals, using a fitted intrinsic resolution for the telescope's reference sensors. If the result is correct, the RD53A geometry will meet the position-resolution needs of the HL-LHC innermost pixel layers.

What carries the argument

The load-bearing identity is the unbiased-residual relation $\sigma_{\rm unbiased}^2 = \sigma_{\rm intrinsic}^2 + \sigma_{\rm track}^2$, which converts the measured residual width into the desired intrinsic resolution. The track resolution on each telescope plane is supplied by a simulation of the same geometry and material; the unknown intrinsic resolution of the six reference sensors is fixed by scanning a range of values and minimizing a $\chi^2$ against the biased residuals on those planes, yielding 3.85 micrometers. The 13-degree tilt works geometrically, reducing the effective pixel pitch as seen by the incoming particle, and this shrinking projected pitch is what drives the reported improvement.

What would settle it

Redo the analysis with a reference telescope whose track resolution is independently known to be much smaller than the expected DUT resolution, so the subtraction in Eq. (2) becomes negligible; if the directly measured intrinsic resolutions differ from 10.9 and 6.8 micrometers by more than the stated uncertainties, the material model or the fitted Mimosa26 resolution is wrong. Alternatively, recompute the DUT values using the 3.35 micrometer intrinsic Mimosa26 resolution obtained from unbiased residuals; a shift exceeding the quoted systematic uncertainties would indicate that the single-common-resolution assumption is inadequate.

Watch

Extended reading notes

Core claim

The central measurement is that non-tilted RD53A modules with binary-readout clusters perform close to the ideal pitch/√12 limit, and that a 13-degree tilt effectively shrinks the projected pixel pitch, improving the intrinsic resolution of the 50-micrometer direction from 14.51 to 10.86 micrometers and of the 25-micrometer direction from 7.92 to 6.81 micrometers. The quoted resolutions include systematic uncertainties; the ratios of tilted to non-tilted resolution are 0.74±0.09 and 0.86±0.30 for the two directions. The authors establish these values by measuring unbiased residuals on the device under test and using the identity σ²_unbiased = σ²_intrinsic + σ²_track, with the track resolution obtained from a simulation tuned to the telescope geometry and a fitted common intrinsic resolution of 3.85 micrometers for the six reference sensor planes.

Load-bearing premise

The final numbers rest on subtracting a simulated track resolution from the measured residual width; that simulation assumes the test-beam material and beam energy are known to about 10% and that all six reference sensors share one fitted intrinsic resolution of 3.85 micrometers. If those assumptions are wrong, the subtraction changes and the quoted 10.9 and 6.8 micrometer values shift.

Editorial extensions

If this is right

  • Non-tilted RD53A modules already operate near the ideal pitch/√12 limit, so binary readout does not degrade the position measurement.
  • A 13-degree tilt improves the short-pitch direction resolution to 10.9 micrometers (50-micrometer direction) and 6.8 micrometers (25-micrometer direction), reductions of 26% and 14%.
  • RD53A modules can satisfy the position-resolution requirements for the HL-LHC innermost pixel layers, informing the geometry of the phase-II upgrades.
  • The methodology of unbiased residuals with simulated track-resolution subtraction transfers directly to future pixel-module test-beam campaigns.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Because the measured gain from tilting is close to the geometric foreshortening of the projected pitch, larger tilt angles would plausibly improve resolution further until charge sharing and cluster fragmentation degrade the binary centroid estimate; the paper does not test this limit.
  • The difference between the fitted common intrinsic resolution of 3.85 micrometers (from biased residuals) and the 3.35 micrometers obtained from unbiased residuals suggests that a per-plane variation may be hidden; if propagated, this could widen the quoted DUT uncertainties.
  • The same subtraction method could serve as a general characterization pipeline for other pixel geometries, allowing different sensor pitches and thicknesses to be compared for future collider detectors.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper reports a testbeam measurement of the intrinsic position resolution of two RD53A pixel modules, one with 50×50 µm² pitch and one with 100×25 µm² pitch, using an 11 GeV electron beam at SLAC. Tracks are reconstructed with the GBL fitter in EUTelescope using the CALADIUM Mimosa26 telescope. The DUT resolution is extracted from the unbiased residual width after subtracting a simulated track resolution, and the Mimosa26 intrinsic resolution is calibrated from the biased residuals of the telescope planes. The authors report non-tilted resolutions close to pitch/√12 and improved resolutions for a 13-degree tilt: 10.9 µm for the 50 µm direction and 6.8 µm for the 25 µm direction.

Significance. The measurements are relevant for the HL-LHC pixel detector development, and the non-tilted results confirm the expected binary-readout resolution. The paper includes a useful consistency check in Fig. 8 and a derivation of the biased/unbiased residual relations in Appendix A. However, the central numbers rely on a simulated track-resolution subtraction whose input is calibrated on the same data, and the uncertainties from that calibration are not propagated into the final results. The claimed 14% improvement in the 25 µm direction is not statistically significant. With additional analysis and clearer reporting, the result could become a solid reference for RD53A characterization.

major comments (3)
  1. [3 (Eq. (4) and Fig. 7)] The determination of the Mimosa26 intrinsic resolution is not propagated into the DUT resolutions. The parabola fit in Fig. 7 gives the minimum at 3.849 µm with a statistical uncertainty of ±3.368 µm (as printed on the figure), yet the text takes 3.85 µm as a fixed value. The systematics bullet also states that the unbiased residuals give 3.35 µm. Because Eq. (2) subtracts the simulated track resolution, which scales with this input, the DUT resolutions in Table 1 (especially the 6.81 µm tilted 25 µm value, whose error bar already exceeds its separation from the non-tilted value) are missing a potentially large source of uncertainty. Please repeat the extraction using 3.35 µm and using the ±3.4 µm range, and add the resulting variation to the systematic uncertainties in Table 1.
  2. [3 and Appendix A (Eqs. (1), (2))] The residual relations used to calibrate and subtract the track resolution assume that the track fit uses the true intrinsic resolution of each plane as its hit error (see Appendix A). In the actual analysis, EUTelescope uses a fixed 4.5 µm hit error for the Mimosa26 planes, while the fitted intrinsic resolution is 3.85 µm. The track-resolution simulator is run with the scanned intrinsic resolution, not with the 4.5 µm value used in the fit. The paper does not demonstrate that Eqs. (1) and (2) remain correct when the fit errors differ from the true resolutions. Please either verify with a Monte Carlo or simulation that this mismatch has negligible impact, or modify the simulation to use the same hit-error assumptions as the data fit.
  3. [4 and Table 1] The claimed improvement in the 25 µm direction is not statistically significant. Table 1 gives the ratio of tilted to non-tilted resolution as 0.86 ± 0.30 for the 25 µm direction, which is consistent with 1.0 within one standard deviation. The abstract and conclusion state that a 6.8 µm resolution 'can be achieved' and that the tilted modules improve the 25 µm direction by 14%, without quoting this uncertainty. Please qualify these statements and report the uncertainty on the improvement.
minor comments (5)
  1. [2] The word 'teatbeam' appears in 'was used in SLAC teatbeam'; it should read 'testbeam'.
  2. [1] The phrase 'for the the HL-LHC' contains a duplicated definite article.
  3. [3 (Eq. (3))] State explicitly whether the box width d in Eq. (3) is fixed to the known pitch or is a free parameter, and report the fitted values of d and Σ for the DUTs.
  4. [Table 1] Indicate whether the quoted uncertainties are statistical only or include the systematic effects listed in the text, and show the breakdown of statistical and systematic components.
  5. [References] Reference [15] for the GBL Track Resolution Calculator should include a full citation with the URL or DOI for the Zenodo record.

Circularity Check

1 steps flagged · score 4.0 of 10

Fig. 8's predicted biased resolution is the in-sample fit of Eq. (4); Table 1 is not tautological but inherits the unpropagated 3.85/3.35 μm telescope calibration.

  1. fitted input called prediction [Section 3, Eq. (4) and Figure 8]
    "The value leading to the smallest χ2, 3.85 µm, is taken as the intrinsic resolution of all Mimosa26. Figure 8 shows the good agreement between measured and predicted biased resolution on six Mimosa26 planes based on 3.85 µm intrinsic resolution."

    3.85 µm minimizes χ2 in Eq. (4), which compares measured biased Mimosa26 residuals to sqrt(σ_intrinsic^2 - σ_track,i^2) with σ_track,i simulated using that same σ_intrinsic. Thus Fig. 8's 'predicted biased resolution' is the in-sample best fit, so the agreement is forced by construction, not an independent validation. The same fitted value is then used to simulate the track resolution subtracted in Eq. (2) for the DUTs, while the paper's alternative estimate of 3.35 µm from unbiased residuals is listed as a systematic but not propagated into Table 1.

full rationale

The DUT intrinsic resolutions in Table 1 are not defined to equal the fitted Mimosa26 resolution: they come from measured unbiased residuals on the DUT combined with Eq. (2), so the reported values would shift if the telescope calibration changed. However, the telescope calibration itself is derived from the same biased-residual data via Eq. (4) rather than from an external anchor, and the only validation shown—Fig. 8—compares the data to the very fit that produced the parameter, making the agreement in-sample. The paper explicitly acknowledges an alternative 3.35 µm intrinsic resolution from unbiased residuals but does not propagate the resulting 0.5 µm shift through Eq. (2) into Table 1, so the headline 10.86 and 6.81 µm values carry an unquantified calibration sensitivity. No self-citation or uniqueness-theorem argument appears, and the pitch/sqrt(12) comparison provides an external sanity check, so the central derivation is not purely definitional. The circularity is therefore partial and concentrated in the calibration-and-validation step, warranting a score of 4.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

The central claim rests on one fitted calibration, the Mimosa26 intrinsic resolution, and four unvalidated modeling assumptions: the residual-variance subtraction formulas, the uniformity of telescope resolution, the accuracy of the track-resolution simulator, and the box-Gaussian residual model for large-pitch and tilted DUTs. No new physical entities are introduced.

free parameters (1)
  • Intrinsic resolution of Mimosa26 telescope planes = 3.85 um (scan minimum in the 3 to 5 um range, step 0.01 um)
    Determined by minimizing Eq. (4) against biased residuals from the same SLAC testbeam data, then used as input to the track resolution simulator for the DUT correction. This is a calibration of the telescope, not an independently measured quantity.
assumptions (4)
  • domain assumption The residual-variance relations sigma^2_unbiased = sigma^2_intrinsic + sigma^2_track and sigma^2_biased = sigma^2_intrinsic - sigma^2_track hold for the real GBL fit with multiple scattering and non-perpendicular incidence.
    Eqs. (1) and (2) are used to correct DUT residuals, but the proof in Appendix A covers only a simplified three-plane, constant-track model with no slope and no scattering.
  • domain assumption All six Mimosa26 planes have the same intrinsic resolution representable by a single number.
    The chi2 scan in Eq. (4) fits one sigma_intrinsic for every plane; plane-to-plane variations are not modeled.
  • domain assumption The GBL Track Resolution Calculator with the configured geometry and material describes the actual track resolution at each plane.
    The simulator is used to subtract tracking uncertainty and to predict biased residuals in Fig. 8, but no independent validation of the simulator against the data is shown.
  • domain assumption For large-pitch or tilted binary-readout DUTs, the residual is a convolution of a box of width d (about the pitch) and a Gaussian, with resolution sigma^2 = d^2/12 + Sigma^2.
    This model is used for the 100 um direction and for the tilted cases; the effective widths under tilt are estimated geometrically from Fig. 5 rather than derived or fitted.

how reviews work

0 comments
Cite this review

Pith. "Pith review of The Measurement of Position Resolution of RD53A Pixel Modules." pith.science (2026). https://pith.science/paper/PGS6GXZY

@misc{pith2026190810973,
  author       = {Pith},
  title        = {Pith review of: The Measurement of Position Resolution of RD53A Pixel Modules},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PGS6GXZY}},
  note         = {Machine review of arXiv:1908.10973}
}
abstract

Position resolution is a key property of the innermost layer of the upgraded ATLAS and CMS pixel detectors for determining track reconstruction and flavor tagging performance. The 11 GeV electron beam at the SLAC End Station A was used to measure the position resolution of RD53A modules with a $50\times50$ and a $25\times100\ \mu$m$^2$ pitch. Tracks are reconstructed from hits on telescope planes using the EUTelescope package. The position resolution is extracted by comparing the extrapolated track and the hit position on the RD53A modules, correcting for the tracking resolution. 10.9 and 6.8 $\mu$m resolution can be achieved for the 50 and 25 $\mu$m directions, respectively, with a 13 degree tilt.

Figures

Figures reproduced from arXiv: 1908.10973 by the authors.

Figure 1
Figure 1. Picture of SLAC testbeam set-up. 3 Testbeam reconstruction and analysis The tool used to reconstruct the testbeam is EUTelescope v2.0.0 [7][8][9], which is a collection of processors in Modular Analysis and Reconstruction for the Linear Collider (Marlin). The track reconstruction algorithm used is the General Broken Lines (GBL) fitter [10][11]. The procedures to process the testbeam data are similar to other testbea… view at source ↗
Figure 2
Figure 2. Hits position correlations between two planes. (a) correlation in X direction between [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Residual distribution of Mimosa26. The blue line is fitted gaussian function. (a) residual [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: (a) residual in X direction on 100 × 25 µm2 RD53A module. (b) residual in Y direction on 100 × 25 µm2 RD53A module. If the DUTs are tilted by 13◦ , the equivalent pitch is reduced, which means a better position resolution, as depicted in [PITH_FULL_IMAGE:figures/full_…
Figure 5
Figure 5. Figure 5: Plot to demonstrate the effect from the tilt angle. The example pixel is 100 [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 6
Figure 6. Figure 6: (a) residual in X direction on 50 × 50 µm2 RD53A module. (b) residual in X direction on tilted 50 × 50 µm2 RD53A module. In the testbeam measurement, only unbiased position resolutions are measured for DUTs, how￾ever, intrinsic resolutions are desired. Intrinsic resolu…
Figure 7
Figure 7. Figure 7: χ 2 distribution. Red dots are χ 2 for each scanned intrinsic resolution of Mimosa26. Blue curve is the fitted parabola. The configuration of material in EUtelescope varies by 10% as the systematic uncertainty. • The beam energy is also varied by 10%. • The position of…
Figure 8
Figure 8. Figure 8: Comparison between measured and predicted biased resolution on six Mimosa26 planes [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

17 extracted references · 12 canonical work pages

  1. [1]

    Apollinari, O

    G. Apollinari, O. Bruening, T. Nakamoto and L. Rossi, High Luminosity Large Hadron Collider HL-LHC, CERN Yellow Report CERN-2015-005, pp.1-19 (2017), arXiv:1705.08830 [physics.acc-ph]

  2. [2]

    https://cds.cern.ch/record/2287593

    RD53 Collaboration, The RD53A Integrated Circuit, CERN-RD53-PUB-17-001. https://cds.cern.ch/record/2287593

  3. [3]

    Phys.: Conf

    Timon Heim, YARR - A PCIe based Readout Concept for Current and Future ATLAS Pixel Modules, J. Phys.: Conf. Ser. 898 032053 (2017)

  4. [4]

    Ahlburg et al., EUDAQ - A Data Acquisition Software Framework for Common Beam Telescopes, arXiv:1909.13725 [physics.ins-det]

    P. Ahlburg et al., EUDAQ - A Data Acquisition Software Framework for Common Beam Telescopes, arXiv:1909.13725 [physics.ins-det]

  5. [5]

    Jansen, et al., Performance of the EUDET-type beam telescopes, Eur

    H. Jansen, et al., Performance of the EUDET-type beam telescopes, Eur. Phys. J. Tech. Instr. 3 (1) (2016) 7. doi:10.1140/epjti/s40485-016-0033-2

  6. [6]

    Baudot et al., First test results Of MIMOSA-26, a fast CMOS sensor with integrated zero suppression and digitized output, IEEE Nucl

    J. Baudot et al., First test results Of MIMOSA-26, a fast CMOS sensor with integrated zero suppression and digitized output, IEEE Nucl. Sci. Symp. Conf. Rec. (NSS/MIC) 1169 (2009)

  7. [7]

    Rubinskiy, EUTelescope

    I. Rubinskiy, EUTelescope. Offline track reconstruction and DUT analysis software, EUDET- Memo-2010-12 (2010)

  8. [8]

    Bisanz, A

    T. Bisanz, A. Morton, I. Rubinskiy, EUTelescope 1.0: Reconstruction Software for the AIDA Testbeam Telescope, AIDA-NOTE-2015-009 (2015)

Show all 17 references
  1. [9]

    https://github.com/eutelescope/eutelescope/tree/v2.0.0

  2. [10]

    Claus Kleinwort, General Broken Lines as advanced track fitting method, NIM A, 673 (2012), 107-110, doi:10.1016/j.nima.2012.01.024

  3. [11]

    https://www.terascale.de/wiki/generalbrokenlines/

  4. [12]

    Bisanz, Test-beam activities and results for the ATLAS ITk pixel detector, JINST 12 C12053 (2017)

    T. Bisanz, Test-beam activities and results for the ATLAS ITk pixel detector, JINST 12 C12053 (2017)

  5. [13]

    Frank Gaede, Ties Behnke, Norman Graf, Tony Johnson, LCIO - A persistency framework for linear collider simulation studies, arXiv:physics/0306114 [physics.data-an]. 9

  6. [14]

    Blobel et al., Fast alignment of a complex tracking detector using advanced track models, Comput

    V. Blobel et al., Fast alignment of a complex tracking detector using advanced track models, Comput. Phys. Commun. 183 1760 (2011)

  7. [15]

    Spannagel and H

    S. Spannagel and H. Jansen, GBL Track Resolution Calculator v2.0, 2016. doi:10.5281/zenodo.48795 10 A The relationship between (un)biased residuals, intrinsic reso- lution and track resolution Suppose N pixel planes where on each plane we measure a position ⃗ ytransverse to th...

  8. [16]

    (A15) σ2 biased track =⟨(θbiased−ytrue 1 )2⟩ (A16) = 1 1 σ2 0 + 1 σ2 1 + 1 σ2 2 (A17) 12 Next, let’s compute the residuals: runbiased,i =y1− y0 +y2 2 (A18) rbiased,i =y1− y0 σ2 0 + y1 σ2 1 + y2 σ2 2 1 σ2 0 + 1 σ2 1 + 1 σ2 2 (A19) Note that with our assumptions, the average res...

  9. [17]

    (A24) =σ2 intrinsic,i +σ2 unbiased track (A25) σ2 biased residual,i =⟨(rbiased,i−⟨rbiased,i⟩)2⟩ (A26) = ⟨( y1− y0 σ2 0 + y1 σ2 1 + y2 σ2 2 1 σ2 0 + 1 σ2 1 + 1 σ2 2 )2⟩ (A27) =σ2 1 + 1 σ2 1 ( 1 σ2 0 + 1 σ2 1 + 1 σ2 2 )2− 2 1 σ2 0 + 1 σ2 1 + 1 σ2 2 + 1 σ2 0 + 1 σ2 2 ( 1 σ2 0 + 1...

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.