PHOTOMETRIC FILES - DOWNLIGHTS
fossLED Ltd - 30.09.2026

CONTENTS
  100 products across five families, each in two formats:
      .ldt   EULUMDAT   (DIALux, Relux, most European software)
      .ies   IESNA LM-63-2002
  One folder per family, and in "_ZIPS TO UPLOAD" one archive per family ready
  to attach to the matching product page. _INDEX.csv lists code, series, finish,
  wattage, colour temperature, CRI, beam angle, lumens and basis for every file.

      FL330   4 products    FORMA, adjustable tilt
      FL430  42 products    IRIS, adjustable tilt modular   MEASURED
      FL420  42 products    IRIS, fixed modular             derived from FL430
      FL530   6 products    ARC, adjustable tilt            MEASURED
      FL520   6 products    ARC, fixed modular IP65         derived from FL530

ROTATIONAL SYMMETRY
  Every file is symmetrised about the vertical axis. The goniophotometer runs
  left the beam measurably off-centre: across the 56 source files the half-peak
  beam width swings by 4 to 37 degrees depending which C-plane you cut through,
  and the intensity straight down the axis varies by up to 7.8% between planes.
  DIALux drew that as a lopsided polar plot, which is what specifiers were
  seeing. The supplier confirms competitor files are corrected the same way.

  The correction is the arithmetic mean over the 24 measured C-planes. With
  equally spaced planes the flux integral is

      PHI = INT INT I(C,g) sin g dC dg  =  INT [ 2*pi * mean_C I(g) ] sin g dg

  so a single rotationally symmetric curve equal to that mean carries EXACTLY the
  same luminous flux. Nothing is scaled, renormalised or fudged. Every file was
  re-integrated from disk afterwards and matches its source to 1 part in 10^6.

  The files are now Isym 1 with one stored C-plane, which is the standard
  EULUMDAT encoding for a round downlight, and the .ies carries a single
  horizontal angle of 0, which is the LM-63 equivalent.

  WHAT MOVED: centre-beam intensity, measured against the old C0 plane, by
  -4.2% to +2.1% (mean -1.4%). Total flux, beam angle from the averaged curve,
  LORL, DFF and wattage are all unchanged.

WHAT WAS CORRECTED
  The photometry in the source library was sound - all 56 files integrate to
  1000 cd/1000 lm against their declared LORL. The identification fields were not:

    - Company line blank on 34 files, "FOSS LED LTD" on 21.
    - Colour temperature written "3000K" on 33 and left blank on 13. A viewer
      that cannot parse the K falls back to a 2700 default. Now a plain integer
      throughout, taken from the code letter (G 2700 / C 3000 / N 4000 /
      W dim to warm, declared at the 3000 top of range).
    - CRI declared 0 on all 13 FL330 and FL530 files. Now 95, per the published
      specification - see OPEN POINTS below.
    - Article number and luminaire name swapped on 42 files; the filename field
      matched the actual filename on only 16.
    - No .ies existed anywhere in the downlight range. Every product now has one,
      generated from the same arrays, in absolute candela.

  C and gamma angles, LORL, DFF and wattage are untouched, and that is asserted
  file by file after every write.

BEAM ANGLES
  The order codes are unchanged - customers order by them. But the files do not
  measure what their codes say, so each luminaire name now carries both:

      code says   measures (50% of peak)
         15deg        24 - 25deg
         38deg        41 - 44deg
         60deg        65deg

  Measured from the goniophotometer data over all 24 C-planes. The discrepancy
  is with the supplier, not in these files. Note the 15deg code already delivers
  a 24deg beam, which is the beam the catalogue lists as missing.

DERIVED FAMILIES
  FL420 and FL520 have no measurements of their own. On Edward's instruction
  (30.09.2026) they are issued from FL430 and FL530 respectively - same engine,
  same reflector, differing only in whether the module tilts. The measurement
  report field of every derived file names the file it came from, and _INDEX.csv
  marks it in the "basis" column.

  FL520 carries a caveat: it is IP65 where FL530 is IP20. If the IP65 version
  puts a sealed lens or glass in the light path, both the distribution and the
  flux change and these files will need replacing with a real measurement. The
  published lumen figures for FL520 and FL530 are identical, which is the only
  evidence available here that the optic is the same.

OPEN POINTS
  1. CRI is declared 95 throughout, on Edward's instruction (30.09.2026).
     Worth noting that FL430-00-3N38R is then 1000 lm at 10W - 100 lm/W at
     CRI 95, where a CRI 95 COB normally costs 20-25% against CRI 80. The FL530
     and FL330 files sit at 79-85 lm/W, which is what CRI 90-95 usually looks
     like. If the supplier ever revises either figure, only the flux and CRI
     fields move and the set regenerates in a minute.
  2. Four files are quarantined in NEEDS RESOLUTION and are NOT in this release.
  3. The website publishes one lumen figure per wattage and applies it to all
     three colour temperatures - and that figure is the 4000K value. Warm output
     is 12-20% lower than the site implies. These files carry the real per-CCT
     values, so they will not agree with the website until the website is fixed.

VALUES
  In the .ldt, intensities are cd per 1000 lm (the EULUMDAT convention). In the
  .ies they are absolute candela. Flux is the value for the complete luminaire.

VERIFICATION
  All 200 files in this folder were checked independently of the build: EULUMDAT
  record count, spherical integration to the declared LORL, ASCII with CRLF
  endings, fossLED on the company line, numeric colour temperature, numeric CRI,
  article number, filename field, SKU leading the luminaire name, Isym 1 with
  exactly one stored plane, and for the .ies, candela integrating back to the
  declared flux. Each symmetrised curve was re-integrated from the written file
  and compared against its source measurement. Every derived file was compared
  against its source beyond the identity fields. 200/200 pass.
