BSEP Engine 4 — Ethan Sheehan

01

Elsa: Bristol's first LOx engine

Engine 4 keeps the 5 kN, 20 bar architecture of Engines 1 to 3 and swaps nitrous oxide for liquid oxygen. That one change drove the rest. The chamber goes back to Inconel 718 from CP1 aluminium, because an ox-rich shutdown with LOx has to be survivable. The injector becomes an ox-centred coaxial swirler, which let us delete film cooling. O/F 1.6, IPA fuel, 8 s design burn. Built for Race 2 Space 2026, with Engine 1 declared as the backup: chambers and injectors stay interchangeable across every BSEP engine. The long goal is the 9 km liquid category at EuRoC 2027.

CAD render of Engine 4: the isogrid-latticed chamber and nozzle, three-quarter view from the injector end — Elsa: isogrid chamber and nozzle, three-quarter view.

CAD render of the full Engine 4 assembly: bolted injector top plate, isogrid chamber and nozzle — Full assembly: injector stack, isogrid chamber, nozzle.

CAD render of the Engine 4 chamber on its L-bracket test mount with feed plumbing attached — On the L-bracket test mount with feed plumbing.

02

Design point

5.0 kN thrust at 20 bar chamber pressure, 30 bar delivery on both sides. O/F 1.6: 0.828 kg/s of IPA against 1.325 kg/s of LOx. The chamber is 281.0 mm long, 117.8 mm across, with a 48.10 mm throat, expansion ratio 3.81 and contraction ratio 6.00. Regen cooling runs through 48 spiral channels at a constant 35 degrees behind a 0.8 mm hot wall, inside an isogrid shell to save mass. Five identical coax swirl elements, sized by the Bazarov method, ox inside and fuel fed to the lower manifold straight from the regen exits. PTFE cryo face seals rated to minus 200 degrees C, a silicone piston seal, no inter-propellant seals. Ignition by the airborne torch igniter.

03

The injector

The ox-centred layout is the headline. LOx drops straight through the top plate into the inner swirlers; fuel never has to cross the ox manifold, so the seal count falls. Ox in through a 3/4 inch BSPP port at the centre top, fuel in through 1/2 inch BSPP at the regen inlet. The machined parts arrived in August 2026. Read the strip as the stack, top plate down to the body.

Machined top plate exterior with a central hex boss holding the threaded oxidiser inlet and two smaller instrumentation ports — Top plate, ox inlet side.

Flat machined plate face with a bolt pattern and a ring of small drilled ports near the rim — Plate face, bolt pattern and drilled ring.

Machined plate with a central threaded port and a fine spiral finish across its face — Plate with central threaded port.

Plate with five inner oxidiser swirler posts protruding from its face — Oxidiser posts, five elements.

Oxidiser posts, profile.

Injector body face-on with five element bores standing in a counterbore and a ring of small drilled ports on the shoulder — Body, five element bores.

Injector body at an angle showing the depth of the element bores and the drilled ring — Body, angled.

Reverse side of the injector body with five machined element seats inside the bore and a lobed flange with threaded holes — Body, reverse: element seats.

04

Printed

The Inconel 718 chamber came off the LPBF printer in June 2026, brace webs still on. Isogrid lattice, E-4 and the Elsa snowflake embossed on the shell, the same three-quarter framing as the render. Going back to Inconel from CP1 aluminium is what lets a LOx engine survive an off-nominal shutdown, and the chamber stays interchangeable with every other BSEP injector.

05

Sizing and analysis

The CDR carries the numbers behind the design point. The five swirl elements are sized by the Bazarov method, iterated until the injector parameters converge; the plot below is one element, inner and outer spray cones on a tip-mixing scheme. Cold-flow testing of tip against external mixing settled the scheme: at 8 to 15 bar of water pressure drop, tip mixing mists less.

The isogrid is the structural story. A plain 1 mm outer wall under combined thermal and 49 bar pressure loading sits at about 1 GPa of von Mises away from the throat, which against a 1100 MPa yield at 473 K is a safety factor of 1.1 — effectively no margin. The interlocking-helix lattice drops the global average to about 450 MPa and the safety factor to 2.4, close to the smeared NASA-handbook optimiser’s target of 2. These figures are the team’s, from the CDR.

Coaxial element from the Bazarov implementation: inner and outer spray cones, R_o 5.44 mm, R_v 7.10 mm. Tip mixing, non-impinging fan angles.BSEP, Race 2 Space 2026 CDR

Two-dimensional development of the optimised interlocking-helix isogrid: rib radius against axial distance along the chamber, the two helix families crossing at every node — The optimised interlocking-helix isogrid, developed flat: the two rib families and their intersection nodes along the chamber.BSEP, Race 2 Space 2026 CDR

Dimensioned cross-section of the isogrid T-flange rib: rib width 1.0 mm, depth 2.0 mm, flange width 2.0 mm, flange thickness 0.5 mm on a 1.0 mm skin, sized for safety factor 2 — Rib section at the SF = 2 design point: b 1.0, d 2.0, w_f 2.0, t_f 0.5 mm on a 1.0 mm skin.BSEP, Race 2 Space 2026 CDR

Von Mises stress field on the plain 1 mm outer wall under combined thermal and 49 bar pressure loading, peaking red at the throat and uniform along the barrel — Plain 1 mm outer wall, thermal plus 49 bar: about 1 GPa away from the throat, peaking near 2.2 GPa at it. Safety factor 1.1 — no margin, which is what the isogrid is for.BSEP, Race 2 Space 2026 CDR

Von Mises stress field on the isogrid-reinforced outer wall under the same thermal and pressure loading, the lattice ribs carrying load in parallel with the skin and the field dropping to blue — The same load case with the isogrid: global average about 450 MPa, a 2.5x reduction. Safety factor 2.4 against the smeared optimiser's 2.0.BSEP, Race 2 Space 2026 CDR

Finite element analysis of the aluminium L-bracket engine mount under 5 kN of nominal thrust, meshed and colour-mapped, deformation scaled 85 times — L-bracket at 5 kN nominal thrust: 0.51 mm peak displacement, 52.5 MPa von Mises, safety factor 5. Deformation scaled 85x.BSEP, Race 2 Space 2026 CDR

06

Documents

The Preliminary Design Report from January 2026 and the Critical Design Report from May 2026. The CDR is the canonical document: parameter tables, injector scheme, seals, test plan, sizing calculations, the isogrid appendix and a STAR-CCM exhaust simulation. The test plan is two full-duration burns, then depth of throttle, then increased chamber pressure. Hot-fire status is pending.

Race 2 Space 2026 Preliminary Design ReportPreliminary Design Report, 24 pp. January 2026.24 pp

Race 2 Space 2026 Critical Design ReportCritical Design Report, 30 pp. May 2026.30 pp