Solar Array

This page describes the solar array under development for the 2027 season. The design is preliminary and details may change as supplier discussions and testing progress.

Overview

For the 2027 season the team is replacing its current solar deck, a set of commercial flexible panels with 272 back-contact cells, with a custom-built array. The new deck is designed around three goals:

• Maximum power per unit area, using roughly 6 m² of cell area to make full use of the area allowed by the race regulations.
• Low mass without giving up the handling robustness needed for panels that are regularly taken on and off the boat.
• Marine durability: salt spray, standing water, UV and thermal cycling.

The array consists of 11 panels on the deck plus one spare of each type, each panel with its own MPPT.

Key figures (estimated)

Item	Value
Cell area on deck	~5.95 m²
Panel area on deck	~6.9 m²
Panels on deck	8 × 3×12, 3 × 2×10
Spares	1 × 3×12, 1 × 2×10
Laminate thickness	~1.33 mm
Laminate mass	~1.85 kg/m² of panel, ~12.5 kg for the deck
MPPTs	One per panel
Bypass diodes	None

Cells

The array uses n-type back-contact silicon cells in the 182 × 188 mm rectangular wafer format. The cell supplier is still being confirmed.

Back-contact cells were chosen because:

• All metallisation is on the rear, so no light is lost to front busbars or fingers.
• They reach the highest efficiencies currently available in mass-produced silicon.
• They go into soft reverse breakdown at a low reverse voltage, typically 2–5 V against 10–20 V for front-contact cells. A shaded cell effectively bypasses itself, which keeps hot spots under control and removes the need for bypass diodes (see Electrical design).

Cutting

A full cell in this format produces well over 10 A, which is above the input limit of our MPPTs. Short-circuit current scales with cell area, not with cell technology, so the cells are cut down to bring the current per string into range:

Format	Dimensions	Approx. current	Used in
Half cell	182.2 × 93.875 mm	~7 A	3×12 panels
Quarter cell	182.2 × 46.94 mm	~3.5 A	2×10 panels

Cutting does not change power per unit area, apart from a small loss from recombination at the new cut edges: roughly 1% for half cells and 3% for quarter cells. Thermal laser separation (TLS) is preferred over laser scribe-and-cleave, because it leaves a much cleaner edge with fewer microcracks. This matters in a thin laminate that flexes.

On back-contact cells, the cut line must suit the rear contact pattern, so that every piece keeps both polarities. Cut positions therefore follow the cell manufacturer's specification.

Each half cell has one cut edge, recognisable by its square corners. The opposite long edge is the original wafer edge, with small rounded corners about 1.2 mm across. Inner quarter cells have cut edges on both long sides.

Panel layout

Panel	Cells	Outline (mm)	Cell area	Mass (est.)	On deck	Spare
3×12	36 half cells (3 columns × 12)	590.6 × 1188.5	0.616 m²	~1.29 kg	8	1
2×10	40 quarter cells (2 columns × 20)	406.4 × 1016.8	0.342 m²	~0.74 kg	3	1

Common layout rules:

• Cell gap: 2 mm.
• Margins: 20 mm on all sides. The end margins hold the bus ribbons and cable exits; the side margins could be reduced in a future revision.
• Corner radius: 5 mm on the panel outline.
• Orientation: the long edge of each cell (182.2 mm) runs across the width of the panel, so bending along the panel's length bends the cells across their short dimension. With this orientation the bending stress acts across the cut edge, which is why cut quality matters (see Cutting).

Stringing

All cells in a panel form a single series string, laid out as a serpentine along the columns:

• The 3×12 has an odd number of columns, so the string ends at the opposite end of the panel from where it starts. The positive and negative exits are at opposite ends.
• The 2×10 has an even number of columns, so the string returns to its starting end, and both exits sit at the same end.

Interconnection is on the rear of the cells, with stress relief at every cell gap. Bus ribbons run in the margins only: nothing passes behind a cell, since the rear of a back-contact cell is live with both polarities.

Using quarter cells on the 2×10 doubles its number of cells in series, so its voltage is close to that of the 3×12. That lets every MPPT channel on the boat run in a similar voltage window.

Laminate stack-up

Numbered from the sun side

Layer	Material	Specification	Supplied thickness
1	ETFE film	Textured outer face; inner face etched or corona-treated	100 µm
2	POE encapsulant	PV grade, crosslinking	350 µm
3	Solar cells	Back-contact, half or quarter cells	~130 µm
4	POE encapsulant	As layer 2	350 µm
5	Glass/epoxy (FR4) sheet	Unclad, fully cured, Tg ≥ 170 °C	400 µm

Finished thickness is about 1.33 mm. Layer thicknesses are nominal as supplied; the encapsulant flows during lamination, so individual layers cannot be measured afterwards. The whole stack is laminated in a single cycle. The rear surface visible through the cell gaps and margins is specified as white, either with white FR4 or a white rear encapsulant.

Why these materials

ETFE front sheet. Light, UV-stable, abrasion resistant and strongly hydrophobic. Its refractive index (about 1.40) is lower than glass, so front reflection is already low without an anti-reflection coating. The textured finish reduces glare and helps at low sun angles, and costs nothing extra. 100 µm was chosen over the more common 50 µm for extra abrasion and puncture resistance during handling.

POE encapsulant. Chosen over EVA because ETFE lets water vapour through across the whole front face. When moisture reaches EVA it produces acetic acid, which corrodes cell metallisation. POE doesn't, and its high volume resistivity also helps against potential-induced degradation. The same thickness is used front and rear to keep material handling simple. The rear layer has to fully bury the interconnect, since all of it sits on the back of the cells.

FR4 backing. The glass/epoxy sheet provides stiffness for handling. Its advantages over other backing options:

• It is an electrical insulator, so no separate dielectric layer is needed.
• Its thermal expansion (~14 ppm/K) is close to that of silicon (~2.6 ppm/K) compared with polymers. After cooldown the cells sit in mild compression, which is the benign direction for silicon.
• It can be laminated in the same cycle as the rest of the stack, so there is no separate bonding step and no adhesive bondline that could fail later.
• The sheet must be fully cured before lamination; outline and cable exits are machined beforehand. Every cut edge is resin-sealed to stop water wicking along the glass fibres.

Stiffness and flatness

Bending stiffness comes almost entirely from material far from the neutral axis. In this stack the silicon and the FR4 act as a thin sandwich, with the rear POE between them. The shear-lag length through the POE is about 15 mm, much shorter than the cells, so the composite action develops across almost every cell.

Relative stiffness (modelled, per unit width)

Build	Thickness	Mass	Stiffness
Bare laminate, no backing	~1.1 mm	~1.3 kg/m²	1×
Current commercial panels (estimate)	~1.9 mm	2.2 kg/m²	~9–10×
New stack, 0.3 mm FR4	1.23 mm	1.66 kg/m²	~10×
New stack, 0.4 mm FR4	1.33 mm	1.85 kg/m²	~15×
New stack, 0.5 mm FR4	1.43 mm	2.04 kg/m²	~21×

The chosen build is about 1.5 times as stiff as the current panels while being lighter per square metre. Thermal-bow modelling predicts less warping than an all-polymer panel, but these results are only reliable in relative terms. A full-size trial panel will be measured for bow and compared against the current panels in a cantilever deflection test.

Edges

No butyl edge seal is used. In glass-glass modules an edge seal matters because the edge is the only route for moisture, but with an ETFE front the moisture path through the face is thousands of times larger, so sealing the edge alone achieves very little. The real edge risk is mechanical: delamination starting from a knock or a peel, after which liquid saltwater can wick along the interface. This is addressed with:

• 20 mm margins of well-bonded encapsulant.
• A flexible edge bead (neutral-cure silicone or PU) applied after lamination, which is easy to renew.
• Resin-sealed FR4 edges.

Electrical design

• One MPPT per panel. Nothing else is in series with a panel, so nothing can drive it into deep reverse bias.
• No bypass diodes. With one MPPT per panel, a panel-level diode would never conduct. Within the panel, the cells' soft breakdown limits the reverse voltage on shaded cells to a few volts each. A fully shaded half cell dissipates in the order of 20 W, which is hot but survivable, and far below the level at which a standard cell would be damaged.
• MPPT tracking. Cells moving in and out of breakdown under changing shade create steps in the power curve. The MPPTs should perform periodic global sweeps rather than purely local tracking.
• Current below 10 A on every panel, set by the cell cutting.
• Voltage. Both panel types run in the region of 25–30 V open-circuit. The MPPT input window must cover the cold-morning open-circuit voltage, which rises by a few percent below 25 °C.

Cable exits

Each exit is on the rear of the panel, within the end margin and never behind a cell. The flat bus ribbon passes through a slot in the FR4 and connects to a 2.5 mm² fine-stranded, marine-suitable cable inside a small termination on the rear face.

Requirements:

• 1000 mm pigtail, polarity marked red and black.
• Termination at most 15 mm high on the rear face. The deck needs a recess or pass-through under each exit.
• Cable anchorage able to take a 50 N pull without loading the bus ribbon or laminate.
• Watertight sealing for marine exposure. No acetoxy-cure silicone, since it releases acetic acid.
• An expansion loop in the bus ribbon inside the laminate, before it reaches the slot.

Handling

The panels are much stiffer than a bare laminate, but silicon still cracks under sharp bending. The worst event for a removable panel is being lifted by a single corner.

• Lift with two people, supporting the panel along its length.
• Never lift by a cable or a corner.
• Store and transport flat on a rigid carrier with the front face protected.
• Check suspect panels under electroluminescence (EL) imaging. On the FR4 panels, polarised sunglasses also reveal stress patterns in the polymer layers, which is a useful quick check after lamination.

Quality checks

• EL imaging of each string before lamination and of each finished panel.
• I-V flash test of each finished panel.
• A first article approved before batch production.
• Trial laminate, possibly with standard cells, to check FR4 adhesion and flatness before production cells are used.

Alternatives considered

Option	Outcome	Reason
Thicker encapsulant for stiffness	Rejected	Encapsulant has a modulus around 20 MPa and contributes almost no stiffness. Heavier than FR4 for a fraction of the benefit.
Carbon/foam sandwich bonded after lamination	Not chosen for now	Stiffest option per unit mass, but needs a second process step, a separate dielectric and an adhesive bondline. Remains an option for panels that are permanently mounted.
ETFE-only rear sheet	Not chosen	About 4–5 kg lighter across the deck, but roughly a tenth of the stiffness of the current panels. Too fragile for frequent removal.
Light-redirecting (prismatic) film	Rejected	Its gain comes from recovering light off front-side busbars, which back-contact cells don't have.
Anti-reflection coating on the ETFE	Rejected	ETFE's low refractive index already does most of the job, and coatings tend to be the first layer to abrade.
Butyl edge seal	Rejected	See Edges.
Bypass diodes	Rejected	See Electrical design.
Pairing two half-cell 2×10 panels in series	Fallback	Would give a similar voltage to the quarter-cell design if quarter cutting turns out not to be possible.