Why EEPISAT Mengangkasa scored 568 of 688 on the CDR

Separation-system optimization study: the PDR trade-off is narrative-only and the CDR "optimization" section is really just the specs of the already-chosen CO₂ system — no comparison of alternatives at the CDR stage.

Shear-pin optimization is otherwise solid, but Table V-14 states 268 N/pin while the same pin/material is derived as 172.9 N/pin everywhere else in the section — an unreconciled contradiction that undercuts the stated safety factor.

Separation ground test: titled "Test Plan" in future tense, then the closing paragraph claims it "has been carried out." No confirmed ≤30s video, no reported measured result, and the CO₂ force used here is double the value used elsewhere for the same joint.

Mass budget totals 4.023 kg — only 23 g above the 4.0 kg floor, a thin margin, printed with a units typo ("4.023 g / 4 kg"). The optimization study compares two designs narratively, no quantitative trade matrix.

Buzzer specified as ">100 dB" but there's no attenuation calculation actually proving 100 m audibility — the spec sheet rating is asserted, not demonstrated.

Anti-jamming is genuinely tested (physical PVC ejection test, compared scenarios). Anti-slide-toward-motor retention — half of what this rubric line asks for — has no comparative study or load calculation at all.

PDR→CDR differences are well quantified but never close with the explicit "this design complies with the specification" confirmation the item literally asks for.

Highest risk in this section: the described surface process (epoxy primer, 600→1500 grit, compound polish) reads as a near-mirror finish, yet the report claims it yields the 60 µm "Regular Paint" value — with no Ra measurement or coupon evidence either way. This is the one item in the whole CDR closest to the rubric's explicit disqualification clause; get real measurement data before your next submission.

The intro narrative claims fin thickness changed 8mm→3mm between PDR and CDR, while the fin table and flutter section both say fin geometry is unchanged — the section contradicts its own headline claim.

The fin bending-load calculation uses Cr=150/Ct=200/b=280mm, a different geometry than the fin specified everywhere else (220/80/120mm) — casts doubt on whether the reported safety factor was actually computed against the real fin.

Assembly-strategy CAD and a 4-phase procedure are shown, but there is no optimization study at all — no alternative sequences, no constraints, no comparison. This is the cleanest single miss in the section: full narration, zero trade study.

Motor integration has a strong photographed step-by-step, but — unlike the separation-integration item right next to it, which names two responsible people — no named personnel or stated competency is given here.

Indigenous-FCC optimization has real scenarios (A/B/C) but the "weighted trade-off" table lists priorities with no actual weights or scores, and it optimizes architecture choices rather than the I-FCC design itself.

Claims 4 sensors "used" by the algorithm, but the text itself says GPS and the current sensor are excluded from the separation decision — only 2 of the 4 actually drive the trigger logic.

Antenna trade-off tables are solid but skip the explicit "PDR-phase summary vs. CDR analysis" structure the rubric requires — the adjacent link-budget item (§24) does this correctly, this one doesn't.

Indigenous FCC algorithm test: a bare video link with no in-report description of the setup, the simulated inclination/altitude inputs, or the result — the rubric explicitly wants this "presented in detail" in the text itself.

Communication test: same pattern — link only, no distance/data/setup written up, and it's unclear whether this reuses the Suramadu Bridge field test rather than a dedicated ≥100 m prototype test as specified.

Sensor & filtering test: one link stands in for what should be a separate test per sensor — no evidence of per-sensor testing or filtered-vs-raw data, which is the actual point of this test.

Present and page-referenced correctly, but all ~114 requirement rows are marked "Fully Met" with zero exceptions — the exact self-assessment pattern that reads as unreliable rather than reassuring to a grader who is told to independently verify.

Filename convention is correct, but nothing in the report body visibly cross-checks the spreadsheet's totals against the OpenRocket data it's supposed to match.

Filename convention correct; "all options selected" on export isn't independently confirmable from the report itself.

All 10 mandatory sentences and the time/altitude/velocity table are filled with real numbers — strong overall. Small blemish: the table says primary-chute deployment at t0+25.60s, the narrative says t0+25.65s, and one sentence is truncated mid-number.

Real 27-entry reference list with correctly matched in-text citations — but the template's own 7 sample references (Anderson, Cesaroni, Fleeman...) were never deleted and still sit in the section.