Expeditions

New to MARGIN?

It shows how close NASA’s fire tests in space come to the cabin air future crews will breathe. One minute to learn the screen.

Read the guide

Reference atmospheres

Findings checked against primary sources

Scenario

Landmark

Question

O₂%

Pressure

Material

Assumed flow, cm/s

MARGIN, created by Alexander de Gentile-Murnieks for NASA Space Apps 2026 · Apache-2.0 · NASA PSI data; no NASA endorsement · AI-assisted development, disclosed under Sources

Evidence field · Microgravity

Evidence proximity, not a safety rating. Not a hazard map.

Source: NASA PSI run tables (snapshot 2026-11-14).

9 comparable runs have no recorded pressure: not drawn here, still counted for the verdict as partial matches.

PSI snapshot2026-11-14 · 1,870 runs

Inspector

Microgravity · Material flammability · any material · PSI only

EvidenceDIRECT

D0.00computed

Comparable tests close on every compared axis^[1]

partial match Compared on oxygen + pressure only; flow not recorded. A missing axis can only make a test look closer. On all three axes: NEAR (D 2.02).

Evidence proximity, not a safety rating.

DIRECT (D ≤ 1), NEAR (≤ 2.5) and DISTANT are classes defined by MARGIN’s method, not established scientific categories. How D is computed

At 20.9% O₂ and 101.4 kPa (14.7 psia), with 10 cm/s ventilation, the nearest comparable test is (3/8" clear rod)^[2], 0.00 units away on oxygen and pressure only; flow not recorded, matching on every compared axis. Differences are test minus scenario; one unit = 2 pts O₂, 25% pressure or 5 cm/s flow. Best fully-compared test: ^[3], D 2.02 (NEAR). Its pressure is nominal (stated), not measured.

Gaps to the nearest test (partial match)

Axis	Gap in scale units	Physical difference
O₂1 unit = 2 pts		±0.00 (±0.0 pts)
Pressure1 unit = ×1.25		−0.00 (−0 %)
Flow1 unit = 5 cm/s	not recorded · not compared	—

D = √(Σ gap²) = 0.00 on 2 of 3 axes. Signed as test minus scenario.

Gaps to the best fully-compared test

Axis	Gap in scale units	Physical difference
O₂1 unit = 2 pts		+0.30 (+0.6 pts)
Pressure1 unit = ×1.25		−0.00 (−0 %)
Flow1 unit = 5 cm/s		+2.00 (+10.0 cm/s)

D = √(Σ gap²) = 2.02 on 3 of 3 axes. Signed as test minus scenario.

Published evidence is off: only PSI run-table rows are counted.

15% of the charted window lies within D 2.5 of comparable evidence (computed on a 120 × 90 grid; uniform in log pressure).

Findings checked against primary sources

MEASURED checked against primary source · AI-assisted, page and quote recorded

On the ISS, PMMA rods in low-speed forced flow kept burning at a lower O₂ than the 1-g upward-spread limit reported for the same material.

13.5% vs 15.4% O₂ · BASS-II, cast PMMA rods in concurrent flow, minimum of the flammability boundary at stretch rate 8–12 s⁻¹; the 1-g value is upward spread with an added 4 cm/s forced flow (Hirsch et al. 2003, as cited in the report)

Scope Cast PMMA rods (radius 0.318–0.635 cm) in the ISS Microgravity Science Glovebox

Not valid for Other materials or geometries; partial gravity; rating any material

“13.5 percent O2 at a of 8 to 12 s⁻¹ is significantly lower than the observed upward flame spread limit of 15.4 percent O2”

Olson et al., BASS-II Summary Report, NASA/TM-20210011385, §6.1.4.4, p. 76 (PDF p. 84) (opens in a new tab)

MEASURED checked against primary source · AI-assisted, page and quote recorded

Near the low-O₂ limit, PMMA rods on the ISS took far longer to ignite than the igniter-on time used in the NASA standard test, and ignition limits sat above flammability limits.

up to ≈10× the standard igniter-on time · BASS-II, cast PMMA rods, hot-wire igniter, ISS Microgravity Science Glovebox; ignition attempts of up to 500 s failed at very low O₂

Scope PMMA rods with one igniter design

Not valid for Other igniters, materials or geometries

“The ignition delay time can be up to almost 10 times longer than the NASA STD–6001 standard igniter on time.”

Olson et al., BASS-II Summary Report, NASA/TM-20210011385, §6.1.6, p. 89 (PDF p. 97); 500 s attempts §6.1.3, p. 67 (PDF p. 75) (opens in a new tab)

MEASURED checked against primary source · AI-assisted, page and quote recorded

Without buoyancy to carry it away, fuel vapour from heated samples accumulated and ignited as a sudden deflagration in two Saffire tests.

2 of 5 Saffire IV–V tests · Saffire IV-2 (2-sided PMMA, 100 kPa, 22% O₂, after flow was cut to 5 cm/s) and V-4 (structured PMMA, 72.6 kPa, 25.7% O₂, at ignition in 20 cm/s flow), conditions as in ICES-2021-266 Table 1

Scope Two events in large-scale tests; the authors suggest curled ash upstream may have helped V-4

Not valid for Predicting when or where a deflagration will occur

“Two Saffire samples (IV-2 and V-4) exhibited intense ignition processes of deflagration because of a stratified mixture”

Urban et al., ICES-2021-266 (NTRS 20210011521), §III.B, p. 5; V-4 at ignition p. 7 (opens in a new tab)

MEASURED checked against primary source · AI-assisted, page and quote recorded

The gravity effect on limiting O₂ depends on the material: Ultem's 0-g limit sat slightly above its 1-g limit, and Nomex's Martian-g limit sat below its lunar-g limit.

Ultem ΔO₂ (0 g) +0.25 · Nomex ULOI* Mars 19.9 vs Moon 21.0 % O₂ · Same drop-tower study: Ultem 1000 at 10.2 psia, Nomex HT90-40 at 14.7 psia; 5.18 s tests

Scope Two charring materials, one thickness each

Not valid for Ranking gravity levels for untested materials

“For these charring materials, the oxygen margin of safety in 0g is very close to zero”

Olson & Ferkul, ICES-2012-3492 (NTRS 20130010991), Table I, p. 6; charring-material discussion p. 7 (opens in a new tab)

Key: what the marks mean

DIRECT · D ≤ 1 to a comparable test
NEAR · D ≤ 2.5
no comparable evidence for this question and flow (D > 2.5)
NEAR boundary if the O₂ and pressure scales were halved or doubled
PSI run, comparable for this question
PSI run, other fuel class
published (not in PSI)
reported limit: a threshold, not a tested condition
planned: never counted as evidence
reference cabin atmosphere
your scenario (the probe)
Marks: ● solid · ◆ liquid droplet · + gas burner · ■ smoke source. Bands are computed at the scenario’s flow, question and material, in this gravity regime only. Lightness encodes closeness of tested conditions, not hazard.