# S-7.5 Prospector v2 - Critique, round 1

Authority: `Assets/allinone.png` (sheet px, y down). Coordinates: s = m aft of bow, x = lateral, z = height above
nose-lens centre. Sheet-to-metre conversions use the Designer's scales unless stated.
Evidence crops (blueprint next to / above the model render, same sheet-px grid) are in
`models/renders_v2/critique_r1_*.png`. Nothing in the .blend, the build script or the GLB was changed.

Tags: **[measured]** = pixel-measured on the sheet and/or read from the scene; **[judged]** = visual judgement.

---------------------------------------------------------------------------------------------------------------

## 1. Verdict per requirement

| # | Requirement | Verdict | Why |
|---|---|---|---|
| 1 | Matches the sheet in every view | **FAIL** | Side: silhouette good, but pods, canopy frame, hinge, seams, scoop wrong. Top/bottom: outline forward of s=2.5 and outboard gear good; hull in the tank bay and tail is too wide, pods wrong shape. Front and rear: do not read as the drawing (hull width, buried pods, no hoops, clamp lugs, heights). |
| 2 | 7.5 m long, 2.6 m beam | **PASS** | [measured] scene bounds: y -3.750..+3.750 (Nose_Lens .. Propulsor), x -1.300..+1.300 (lower-tank clamp bolt heads; ducts +-1.298). |
| 3 | Moving parts separate, origin on axis | **PARTIAL** | Rotors PASS (own objects, origin on duct axis at s=6.60, x=+-0.829, z=0.13, spin about local Y, parented to the propulsor; same in the GLB). Hatch_Lid PASS (origin on aft rim, s=4.065, z=1.22, hinge about local X). Canopy_Glass is separate with origin on a transverse axis (s=2.37, z=1.26) but **cannot open**: the inner frame, the hinge bar and the hinge blocks are all in the static `Canopy_Frame`, the centre hinge bar (z 1.20..1.31) already intersects the glass roof (z 1.228), and the origin is not on the modelled hinge pins (s=2.02). See item 5. |

GLB sanity: `S75_Prospector_v2.glb` exists (959 000 B, glTF 2), 22 meshes, 12 materials, one root, both rotors are
children of their propulsor, glass material is BLEND. OK. Note: the open .blend reports unsaved changes
(`is_dirty`), it was like that before the critique touched it.

Per-view score (how much of the drawn content is right): side ~70 %, top ~65 %, bottom ~70 %, front ~25 %, rear ~35 %.

---------------------------------------------------------------------------------------------------------------

## 2. Conflict rulings

**R1 beam - UPHELD, with one big correction.** Keep the separate lateral scale in the plan views (117.5 / 112.7
px/m). Three independent checks agree with it [measured]: the stretched plan view puts the upper-pod outer face at
1.055 m (front view: 1.05), the lower-tank clamp face at 1.29 (front 1.29, rear 1.29) and the hatch collar at
r = 0.52 laterally (side 0.53, front 0.54, rear 0.51). Cost: round items look 10 % long in the plan overlays
(hatch collar 138 px long vs 122 drawn) - accepted.
But the claim "all widths come from the top view" is not true for the hull aft of the canopy. Top and bottom views
draw the hull edge in the tank bay at +-77.5 px = **+-0.66 m** (top: lines at y=543 and y=698, the line is separate
from the pod cone edge at x=660; bottom: hull visible between the tanks 876..1025 px), the front view draws the
hull at +-0.70..0.79 and the rear view at +-0.62. The model keeps 0.93 from s=2.3 to 4.6. Only the canopy collar
(s 1.5..2.6) is 0.93 wide. The "pods half-buried in the hull flank" problem is this misread, not a sheet conflict.
See item 1. Residual cost after the fix: the canopy collar is still 0.93 in the model against ~0.79 in the front view.

**R2 heights - UPHELD, no geometric compromise exists.** [measured] Front and rear agree with each other and
both disagree with the side: hull depth 195 px = 1.33 m (front), 210 px = 1.32 m (rear), against 1.86 m (hatch
section) to 2.09 m (canopy section) in the side view. Both are the side-view heights squeezed to 0.64x (front)
and 0.71x (rear); after that squeeze the lower tank centre lands on z = -0.47 in both (side: -0.485). Matching
them would mean flattening the hull by ~30 %, which wrecks the largest view and the only one tied to the 7.5 m
arrow. Cost: front and rear never overlay isotropically.
What to do instead: keep the isotropic overlays, and add a second "normalised" overlay per view, the same trick
already used for the plan views: front ppy = 93.3 px/m with z=0 at y=1270.3; rear ppy = 113.2 px/m with z=0 at
y=1274.6 (ppx unchanged). With those, lateral placement and vertical ordering can be checked and must match.

**R5 no roof frame band - UPHELD.** The side view's double line along the roof is 4 px (0.03 m) against a 16-18 px
sill band, and the top view shows bare glass over the centreline. Read it as the glass silhouette. Cost: one line
instead of two along the roof in the side view.

**R6 hinge blocks - OVERTURNED.** Remove the centreline hinge bar (it is in no plan view). The two lateral hinges
of the top view, built taller, give the side-view silhouette by themselves; the front view shows their lugs at
x = +-0.52..0.55 under the hatch collar. See item 5.

**R7 hatch posts fore/aft - UPHELD** (side + top against front + rear). Cost: front and rear show one post on the
centreline instead of two at +-0.42.

**R8 duct diameter 0.93 - UPHELD.** [measured] side 1.09 m (140 px) and only 1.09 m long; top 0.945 (111 px) and
bottom 1.02 (115 px), both 1.385 m long (179 px); rear 0.79 (r = 63 px). 0.93 is the middle and matches two
views. Cost: side view duct 9 px short of the drawn outline top and bottom and its front lip 0.30 m further
forward than drawn (s 6.10 vs 6.41); rear view ring 11 px too big in radius and its centre at 0.829 vs 0.90 drawn.

**R10 guard hoops - OVERTURNED, build them.** They are the outermost item of the front view together with the
clamp lugs (hoop outer face at x = 108.5 / 489 px = +-1.30). Put them at the two clamp stations so the other views
are barely affected: in the top view they sit inside the drawn clamp blocks, in the side view the forward one
lies on the drawn seam line at x=447 px. See item 3. Cost: two vertical tubes in the side view that the side
view does not draw; the aft one crosses the pod near its rear band.

R3 (canopy rear from the side view), R4 (shelf), R9 (flattened pod), R11, R12, R13: upheld. R3 gets a cheap
improvement (item 11). R12 costs 0.26 m in the bottom view (dome drawn at s 0.09..0.74 there).

---------------------------------------------------------------------------------------------------------------

## 3. Corrections - MUST FIX (ordered by visual impact)

### 1. Hull half-width aft of the canopy collar  [measured]  views: top, bottom, front, rear
Blueprint (top view, half-width px from y=621 -> m at 117.5): shoulder curve 108 px at x=385 -> 78 px at x=462,
then a straight edge at 77.5 px to x=690, then the tail.
Model: `W_T` = 0.93 from s=2.3 to 4.6, 0.74 at 4.95, 0.64 at 5.25, 0.55 at 5.55, 0.39 at 6.0, 0.22 at 6.45.
Change `W_T` aft of s=2.3 to:

| s | 2.30 | 2.54 | 2.73 | 2.885 | 3.04 | 3.13 | 4.90 | 4.97 | 5.28 | 5.59 | 5.90 | 6.21 | 6.48 | 6.83 | 6.88 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| W | 0.93 | 0.92 | 0.834 | 0.757 | 0.69 | 0.664 | 0.66 | 0.647 | 0.545 | 0.477 | 0.35 | 0.22 | 0.20 | 0.20 | 0.13 |

That is 0.27 m narrower per side in the bay and 0.06..0.09 narrower along the tail. Tail stub: 0.20 half-width
from s=6.48 to 6.83 (drawn 895..940 px, +-23.5 px), then the nozzle r 0.12 to s=6.91.
Follow-on: the lower tanks (inner face 0.715) then stand 0.06..0.17 m off the flank. Bridge that with saddle
brackets at the two clamp stations; the front view shows the inner clamp lugs bearing on the hull at x 0.70..0.80.

### 2. Upper pods (Tank_Upper_*)  [measured]  views: all five
Blueprint: top view body y 497..543 px = 0.39 wide, centre 101 px = **x 0.86**, outer face 1.055; front view outer
face 1.05, inner face 0.81, hull 0.70, i.e. the pod stands outboard of the hull with a bracket gap. Side view:
body x 500..647 px (s 3.41..4.55), top y=206, bottom y=311 = **0.816 tall, centre z = +0.354**; cones to flat caps
at s 3.11 and 4.87, caps 52 px = 0.40 tall; double band rings at s 3.36..3.43 and 4.52..4.58; two seam lines at
mid height (y=258 and 268 -> z +0.36 and +0.28) and faint ones at y=217 and 300. Top view: both ends taper to a
point in plan (tips at x=410 and 683 px), rear cap ring 14 px = 0.12 wide.
Model: centre x 0.80 (0.602..0.998), 0.79 tall at z 0.365, constant 0.38 width with round ends; because the
hull is 0.93 there the cones are inside the hull, so the side render shows a bare rectangle x 494..645 px and the
plan render shows a box.
Change: `POD_X` = 0.855 (inner 0.665, outer 1.045); half-height 0.408, `POD_Z` = 0.354; cones must taper in
width too, 0.19 -> 0.07 half-width at the caps (cap = 0.14 wide x 0.40 tall); seam strips at z +0.36 and +0.28
on the outboard face (replace the three at -38, -52, +55 deg); two standoff brackets per pod to the hull.
The front-view "0.38 wide" in ruling R9 is wrong: the front view draws the pod 34.5 px = 0.24 wide
(x 143.5..178 px); 0.38-0.39 comes from the top view. Keep 0.38.

### 3. Guard hoops (R10)  [measured]  view: front
Blueprint: D-shaped tube hoop each side, tube 11 px = 0.075 m, outer extreme x = +-1.30, from the pod's upper
outboard corner round to the lower tank's upper clamp lug; a small cap block on top of the pod at x 0.86..1.00.
Model: absent.
Change: add a new static object per side (or part of the tank object) with two hoops at s = 2.96 and 4.475.
Tube diameter 0.08. Centreline path in (x, z): (0.97, +0.74) -> (1.20, +0.62) -> (1.26, +0.25) -> (1.20, -0.05)
-> (1.05, -0.20). Outer face 1.30. Upper end on a bracket to the pod / hull shoulder, lower end on the upper
outboard clamp lug.

### 4. Canopy frame does not read as a frame  [measured + judged]  views: side, top, front, beauty
Blueprint: a bold band all round the glass. Side: glass edge y=297, outer frame line y=315 along the sill
(16-18 px = 0.13 m), at the rear glass edge x=378, outer line x=398..400 (0.17 m). Top: band 11 px = 0.094 m in
plan all round (apex x=117 outer, 128 glass).
Model: band geometry is the right size (0.165 at 0.6 flare) but its outer edge blends into the hull, so the line
renders show the glass edge plus broken strokes and the beauty shot shows a thin black gasket only.
Change: give the band a hard outer step, at least 0.02 m proud of the hull with a sharp edge, continuous round
the whole perimeter, so both edges draw as lines in side and top. Also:
- The rear arch sticks up as a tilted fin to z=1.37 at s 2.45..2.55 (side render x 378..398 px, y down to 127);
  the drawn roofline there is flat at y=137 = z 1.30. Nothing of the frame may exceed z=1.30.
- Remove the shading lumps and stray crease lines at the apex (top render x 112..150 px, side render
  x 115..125 px, y 258..270) and at the rear lower corner.

### 5. Canopy hinges, and what moves with the canopy (R6, requirement 3)  [measured]  views: side, top, front
Blueprint: top view two hinge blocks at x 333..362 px, lateral centre 64.5 px = +-0.55, 29 x 19 px (0.22 x 0.16 m),
straddling the glass rear edge; nothing on the centreline. Side view one two-tier block: upper bar x 310..366 px
(s 1.94..2.37), y 137..150 (z 1.20..1.30), rounded front end with a pin at (317,143) = s 1.99, z 1.25; lower
bracket x 318..362, y 150..163 (z 1.10..1.20).
Model: centre bar 0.16 x 0.40 x 0.11 at z 1.255 plus a base block, plus lateral blocks at z ~1.05..1.13 that sit
wholly forward of the glass edge and 0.03 too far outboard.
Change: delete the centre bar and its base. Build the two lateral hinge arms at x = +-0.55 as two-tier arms:
upper bar s 1.94..2.37, z 1.20..1.30, 0.16 wide, pin diameter 0.05 at s 1.99, z 1.25 with axis along X;
lower bracket s 2.00..2.34 from the glass up to z 1.20.
Requirement 3: the glass, its inner frame band and the hinge arms must be one moving assembly (one object, or
frame and arms parented to `Canopy_Glass`); only the hull coaming and hinge bases stay static. Put the origin
on the real pivot (the hinge base on the hoop, x=0, s 2.37, z 1.25, axis X) and check that rotating it 60 deg
clears everything.

### 6. Lower-tank clamps  [measured]  views: front, rear, top, bottom, side
Blueprint: each clamp has four lugs on the 45 deg diagonals. Front view lug bolts at +-27 px = +-0.19 m in x and z
from the tank axis; bottom view two bolts per clamp at +-22 px = +-0.195 laterally, the clamp spans the whole
tank width (y 808..878 px); side view bolts at y 338 and 396 = +-0.225 from the axis; top view clamp block reaches
1.28..1.29 m.
Model: two boxes per clamp (top and bottom, 0.38 x 0.15 x 0.17) pushed outboard to x 1.286, bolts at +-0.285.
In the bottom render they cover only the outer half of the tank.
Change: strap all round plus four lugs per clamp at x = TNK_X +-0.19, z = TNK_Z +-0.20, each ~0.13 x 0.15 x 0.13,
one bolt on its vertical face (seen from the side) and one on its horizontal face (seen in plan). Outboard lug
faces at x 1.29, bolt heads to 1.30.

### 7. Nose  [measured + judged]  views: side, top, front
- Cap seam: drawn C-shaped, s 0.30 at the top (x=100 px), 0.24 at mid height (x=92), 0.40 at the bottom
  (x=113). Model: straight ring at s 0.27. Make it follow those three points.
- Two crease lines from the top of the cap seam to the canopy apex, drawn in the side view from (100,279) to
  (140,258) = s 0.30, z +0.19 to s 0.61, z +0.36, and in the top view as the edges of the two cheek panels
  (x 100..155 px, y 555..580 and mirrored). In the front view they are the two lines fanning up from the lens.
  Model: absent.
- Lens: drawn 20 px tall = 0.155 m and standing 0.08 m proud as a rounded button (x 61..71 px). Model: 0.20
  diameter, flat. Use r 0.078, 0.07 proud, and let the nose taper down to it.

### 8. Hull seams in the side view  [measured]
- Missing: horizontal seam at z = +0.70 (y=213..216) from s 3.00 to 4.97, visible forward of and behind the pod,
  continuing into the existing tail line at z 0.68.
- Missing: horizontal seam at z = +0.124 (y=288) from s 2.52 to 3.00 (x 385..447 px).
- Missing: spine line at z = +1.04 (y=170) from s 3.10 to 4.90.
- Extra: the top-deck seam `long_x(3.01, 4.95, 0.63)` draws at y=183 in the side view and curls down to y=210
  at x 660..695; nothing like it is drawn. Delete it (the plan-view line at +-0.66 is the hull edge after item 1).
- Extra: the s=0.80 ring seam runs under the belly; drawn only from the frame down to the chine (y 287..337).
  Stop it at z = -0.27.
- Chine seam: drawn in the bottom view from s 0.60 to 2.90 (x 135..430 px), in the side view from (162,337) to
  (280,363). Model: s 0.95..1.70 only. Extend to s 0.60..2.90 at z -0.26 (s 0.78), -0.37 (1.47), -0.46 (1.70+).
- Ring seam at s 3.0 curves forward at the top in the side view (x=447 below y=230, 437 at y=180, 425 at y=160).
  Model: straight. [nice to have]

### 9. Tail scoop  [measured]  views: side, rear
Blueprint: smooth tube leaving the belly at s 5.63..5.75; mouth raked, lower lip at (812,395) = s 5.84, z -0.71,
upper lip at (852,366) = s 6.15, z -0.48, drawn as an oval with a lip ring; rear view mouth 63 x 35 px = 0.40 x 0.22,
opening 0.27 x 0.10.
Model: square-cut end at s 6.15 with a forward-pointing spike at (830,371) px in the side render.
Change: rake the mouth plane between those two lip points, add a rounded lip ring, remove the spike. Width and
height are already right (0.38 x 0.21).

### 10. Propulsor ducts  [measured]  views: top, bottom, side, rear
- Lengthwise seams: plan views show a straight seam on the centre plus a bowed one at about 0.7 of the radius
  each side, i.e. 8 seams at 45 deg. Model: 4 at 90 deg. Use 8.
- Everything else (length 1.40, lip, max diameter, band seam at s 6.9, exit ring 0.77) matches the plan views.
- Rear view shows only the 8-point star and hub; the model's 9 wide rotor blades show between the vanes.
  [nice to have] 8 blades, rest angle aligned behind the vanes.

### 11. Canopy rear arch rake (R3 improvement)  [measured]  views: top, side
Blueprint side: the rear glass edge leans forward at the top: x=378 px at y 200..240, 372 at y=180, 365 at
y=165, i.e. s 2.46 below z 0.8, 2.42 at z 0.96, **2.36 at the roof**. Top view: rear edge at x=348 px on the
centreline, bowed aft. Model: flat arch at s 2.47 (straight line at x 373..376 px in the top render).
Change: rake the arch to those three stations. Costs nothing in the side view and cuts the top-view error on the
centreline from 24 px to about 14.

---------------------------------------------------------------------------------------------------------------

## 4. Corrections - NICE TO HAVE

12. **Lower tank ends** [measured]. Rear: drawn as a blunt dome ending at x=681 px (s 4.82) with a ribbed neck
    (plan: 30 px = 0.27 wide at x=680) and a small nipple; model has a 0.27 m long cone down to r 0.06. Keep
    r >= 0.20 to s 4.74, neck r 0.14 over s 4.78..4.86, nipple r 0.07 to 4.89. Front: the flange plate
    (R+0.022, side render x 411..419 px) is larger than anything drawn; reduce to R+0.005. Clamp centre 0.03 m
    too far aft (s 2.995 vs 2.96).
13. **Lower tank lamp** [measured]. Front view lens ring r = 25 px = 0.17, housing r 0.29. Model bezel r 0.135,
    lens r 0.10. Use bezel 0.17, lens 0.14.
14. **Tool housing** [measured]. Round fitting drawn r = 9 px = 0.07 at (333,378) px; model r 0.12 -> 0.075.
    Bottom-view cover is an oval x 277..352, y 852..908 px = 0.58 x 0.50 m centred s 1.99, x 0.64; model oval
    0.72 x 0.47 centred s 1.92 on a rectangular box outline. Use half-length 0.29, half-width 0.25, centre s 1.99.
    Bolted collar between nose dome and box (x 252..262 px) is missing. Nose dome is visibly faceted (20 segments
    is fine, but shade smooth). Extra rail at s 1.98..2.28 under the motor is not drawn.
15. **Hatch** [measured]. Posts drawn at s 3.225 and 4.064 (side) / 3.36 and 4.06 (top); model 3.19 and 4.13.
    Use 3.25 and 4.06. Lid: drawn 4 bosses on the 45 deg diagonals at r ~0.36 plus 2 small bolts on the lateral
    axis; model 6 bolts at 60 deg. Skirt ring at r 0.605 shows as arcs outside the collar in the top render; not
    drawn, reduce to r 0.555. Spine fairing (drawn x 598..656 px, 30 px wide = s 4.19..4.63, 0.255 wide, rounded
    rectangle) is in the model but draws no outline: give it a crisp edge.
16. **Rear-view round fitting** [measured]. Drawn at 80 px left of centre = 0.50 m to port, on the hull bottom
    line, r = 17 px = 0.11, two bolts at 3 and 9 o'clock. The model's cap (x 0.57, z -0.60, r 0.10, 4 bolts) is
    hidden behind the belly in the rear render. Re-check after item 1; target centre x 0.50, z -0.68, r 0.11.
17. **Chin dome** [measured]. Side centre s 0.70 (x=151 px), model 0.67; flange/visor drawn to s 0.92 with a ramp
    down to the belly. Move 0.03 aft.
18. **Side latch blocks** [judged]. Drawn only in the plan views (40 x 14 px = 0.31 x 0.12 m at s 2.05..2.36,
    outboard face 1.055). The side view draws nothing there, the model shows a 0.24 x 0.09 box at z -0.13.
    Keep, but make the block thinner in z (0.05).
19. **Bottom-view belly seams** [measured]. Bottom view draws the two forward belly seams at s 1.92 (x=305 px,
    straight) and 2.34 (curved, bowed aft); model has them at 1.70 and 2.48 from the side view. Side wins; no
    change unless the side view can keep its lines.
20. **Ring seams are drawn bowed** aft by ~8 px in every view (s 3.0, 4.97, 5.6). Not achievable for one seam in
    all views; leave straight.
21. **Render script** [measured]. Bottom view centreline: the drawing's own outline is centred on y = 948..950
    (x=449: 806..1094; x=940: 801..1094), not on the dash-dot line at 955. Use oy = 949, otherwise the whole bottom
    overlay is 5-6 px off sideways. Add the normalised front/rear overlays from ruling R2.

---------------------------------------------------------------------------------------------------------------

## 5. What already matches (do not touch)

Side silhouette of the hull, roof and belly (ZT/ZB tables agree with my scan within 1-2 px); canopy glass
outline in the side view; hull plan outline forward of s=2.5; hatch collar and lid in the side view; tool housing
box, cover plate, feet; belly panel and its four bolts (s 3.72..4.52, +-0.62); pylon box, neck, feed strut and
aft strut in the plan views; duct plan outline; tail lines at z +0.68 / -0.12; rotor hub and 8 stator vanes.

---------------------------------------------------------------------------------------------------------------

## 6. Designer's report - statements that did not check out

1. "All widths come from the top view at that scale" (R1): false for s 2.5..6.5. The top and bottom views draw
   the hull at +-0.66 in the bay; the model has 0.93.
2. "Pods half-buried in the hull flank" listed as a consequence of view conflict: it is a consequence of item 1.
   Plan, front and rear views all agree the pods stand outboard of the hull.
3. R9 "front view = chamfered rectangle 0.38 wide": the front view pod is 0.24 wide; 0.38-0.39 is the top view.
   R9 "side view = 0.79 m tall": 105 px = 0.816.
4. R10 "guard hoops appear in no other view": true, but they define the 2.6 m beam in the front view together
   with the clamp lugs, so leaving them out removes the outermost item of the dimensioned view.
5. R2 "everything below the lens is compressed": the squeeze is over the whole height. The model's roof is also
   32 px above the drawn roof in the front view (y 1117 vs 1149), and the offsets below the lens range from
   36 px (chin dome) to 78 px (hull bottom), not a uniform 55-60.
6. Bottom view "centreline y=955": that is the dash-dot line; the drawn geometry is centred on ~949.
7. R6: the centreline hinge bar is not needed to explain the side view and is contradicted by the top view.
8. Not reported: `Canopy_Frame` protrudes to z 1.37 above the roofline; the canopy cannot open (requirement 3);
   the deck seam shows as a stray curved line in the side view; the pod cones are invisible in every render.