Ethan Sheehan
← All work IGNITER — 2024-25

BSEP Torch Igniter

Codename Tuppence

Hydrogen/oxygen torch igniter: eight firings, retired after melting through 5 mm of stainless. Succeeded by Pascal.

RoleIgniter designer, HyPower Bristol
Firings
8
Inlet pressure
3→12 bar
Wall
5 mm melted through
01

Tuppence

HyPower Bristol's first liquid torch igniter, and the first liquid ignition system I designed. The brief was to make it as small as possible, which forced an extremely short and stubby layout: a single impinging jet for fuel and oxidiser, a spark-plug port, and two palm-sized steel discs bolted together. Short enough that none of us were certain it would light at all. Designed for NOS/IPA, adapted to run on Astron's hydrogen/oxygen rig at Westcott. Built for Race 2 Space 2025 with Jacob Cox, Jose Luis Partida and Jake Marsh.

The assembled steel igniter, spark plug and brass fittings included, held in one black-gloved hand.
The complete igniter in one gloved hand. The size constraint, in one image.
The assembled igniter standing on a bench, spark plug pointing up from the steel body with two brass inlet fittings.
The assembled igniter: two inlet fittings and a spark plug. 9 Jul 2025.
Top-down view of the assembled igniter showing the spark plug and the two brass inlet fittings on the hexagonal steel body.
Assembled, from above. 9 Jul 2025.
02

Workshop

The body was machined from steel by Jack Marsh at the UoB Machine Shop and came out on 25 and 26 June 2025. Two halves: an injector disc carrying the impinging-jet ports and a chamber cap with the spark-plug port. Assembly on 9 July was brass inlet fittings, a copper gasket between the halves, and a standard automotive spark plug.

A hand holding the freshly machined injector half, two small impinging-jet ports visible in the face.
The injector half in hand, day one out of the machine shop. Impinging-jet ports visible.
Machined steel chamber cap on a bench, its concave internal face and central port catching the light.
Machined internal face of the chamber cap.
Hexagonal steel flange with six bolt holes and a central port lying on a workshop bench.
Chamber cap outer face and dowel on the workshop bench.
Both machined steel halves of the igniter lying face-up side by side on a workshop bench.
Both halves face-up. The whole engine is two palm-sized discs.
Low bench-level view of the machined igniter parts lined up on a workshop bench.
The machined steel parts on the workshop bench.
The two igniter halves bolted together into a short hexagonal steel body, fresh from machining.
Body bolted together, fresh machined.
Hands threading a brass inlet fitting into the steel igniter body, spark plug already installed.
Threading the brass inlet fittings into the body.
Close view of the assembled igniter with its two brass inlet fittings and spark plug.
The assembled igniter, fittings side. 9 Jul 2025.
03

Test day, Westcott

One test day, 10 July 2025. The DAQ was up at 09:11, ambient 19.5 °C; sequences ran from 11:29 to 12:38. We armed 13 ignition sequences and got 8 successful firings, stepping the inlet pressures up from around 3 bar to around 12 bar across the day and stretching the armed duration from a few seconds to over twelve.

The body thermocouple tells the day as a staircase: every firing leaves its heat behind, starts climbing from 21 °C to 72 °C, peaks around 110 °C. The torch itself is nearly invisible in daylight on hydrogen and oxygen.

The igniter bolted to a steel bracket on a perforated test plate, thermocouple wires taped to the body.
Pre-fire: mounted on the test plate, thermocouples wired. 10 Jul 2025, 11:04.
Close view of the mounted igniter with a green thermocouple connector, yellow tape and copper gasket visible.
Pre-fire close-up, thermocouples wired. 10 Jul 2025, 11:07.
Side view of the mounted igniter between firings, scorch marks and orange discolouration spreading around the copper gasket.
Between firings: scorch building up around the copper gasket. 11:52.
Top view of the mounted igniter between firings, the joint darkened with scorch.
Between firings, from above. 11:53.
A routine firing (no. 7 of 8), close angle. The torch flame is nearly invisible in daylight.
Line chart of body thermocouple temperature against time of day across 13 armed sequences: a staircase of heat spikes rising from about 20 to 110 degrees, then a final spike to about 159 degrees marked as the melt-through.
Body thermocouple across all 13 armed sequences, 11:20 to 12:59. Every firing left its heat behind; the final burn spikes off the staircase to about 159 °C.
04

The melt-through

The last firing was a deliberate longer burn at the highest inlet pressures, to test the limits of the design. At about t+4.5 s the wall let go: a hole melted through 5 mm of stainless steel and flame torched out of the side of the body, then a shower of sparks. The hot-side thermocouple jumps about 100 °C at the breach; the body soaks to around 159 °C a minute after shutdown.

The inlet pressures are labelled A and B: the DAQ channel map was inferred from behaviour, and which inlet carried hydrogen and which oxygen was not recorded. If we can melt through 5 mm of stainless steel we should hopefully be able to ignite a liquid rocket engine.

The final burn, close angle. Deliberate long burn, breach and spark shower at about 5 s.
The final burn, wide angle. Same 14 s, from the other camera.
GoPro frame of the igniter engulfed in a spray of bright orange sparks on the test plate.
Two frames later: the full spark shower.
GoPro frame of the igniter on the test plate with a jet of flame torching sideways out of the body wall.
Frame from the final burn: flame torching out of the side of the body. Westcott, 10 Jul 2025.
Two-panel chart of the final burn: hot-side and body thermocouple traces, the hot-side jumping about 100 degrees at t plus 4.5 seconds, above two flat inlet pressure traces near 12 and 9 bar.
The melt-through sequence: both thermocouples and both inlet pressures, breach at about t+4.5 s. Which inlet carried which gas was not recorded.
05

Teardown

Off the stand it was sooted but in one piece from the outside, the charred gasket squeezed out of the joint. Split in the vice, the internal dome of the chamber was coated in re-solidified melt and the spark-plug electrode had slumped into a blob of metal. Tuppence was retired here. Its successor, Pascal, carries the lessons into the EuRoC 2026 engine.

The igniter still mounted after the final burn, a charred and squeezed-out gasket protruding from the joint.
Straight after the final burn, still mounted: charred gasket squeezed out of the joint. 12:56.
The igniter off the stand on a bench, heavily sooted but intact, spark plug still fitted.
Off the stand, sooted but in one piece. 13:06.
Chamber half clamped in a blue vice, its internal dome coated in re-solidified grey metal.
Chamber half in the vice: the internal dome coated in re-solidified melt. 13:13.
Both igniter halves split open in a vice, the spark-plug electrode reduced to a slumped blob of metal.
Both halves split: the spark-plug electrode reduced to a slumped blob of metal. 14:10.