OFF-GRID BUILD BLUEPRINT
v7 · AUDITED FREEZE · 2026-11-06
Supersedes battery-chemistry page (v1, 2026-11-02) · chemistry winner unchanged, hardware locked
SYSTEM AT A GLANCE
The full pipeline — photons in, radio waves out
STRING ACCOUNTING — SINGLE SOURCE OF TRUTH
1 string = 2× 24V/400W panels in series (~48V nominal, ~95–100V Voc)
| Configuration | Panels | Strings | Per MPPT | Verdict |
|---|---|---|---|---|
| Starter | 12 | 6 | 1.5 (uneven) | ASYMMETRIC |
| BASELINE 6.4 kW | 16 | 8 | 2 str / 4 panels | BUILD TARGET |
| Doubling 9.6 kW | 24 | 12 | 3 str / 6 panels | COMFORTABLE |
| Hard ceiling 12.8 kW | 32 | 16 | 4 str / 8 panels | ~20% CLIPPING |
| Beyond 32 panels | — | — | — | ADD 2 MPPTs |
MPPT ceiling is 45A PV short-circuit input per unit = 4 strings (8 panels) max — the input-current limit, not just wattage. NEVER 3 panels per string: 150V Voc max in frost.
LOCKED HARDWARE
Decision tree closed 2026-11-06 · audited same day
| Stage | Hardware | Key Numbers | Status |
|---|---|---|---|
| Storage | 3× DatouBoss 48V/314Ah LiFePO4 | 16 kWh/pack · 210A BMS · 5-yr warranty | LOCKED |
| Solar charge | 4× Victron BlueSolar MPPT 150/45 | 2,600W @ 48V each · no Bluetooth (RF hygiene) | LOCKED |
| Inversion | 3× Studer XTH 8000-48 | abs 56.0V · float 54.0V · no CAN comms | EXISTING |
| Transmission | Combiner @ array → 25 mm² bundle, 30m | gPV string fuses + DC breaker + T2 SPD | LOCKED |
| EMC | CMCs + ferrites + bulkhead plate | measure-first: RTL-SDR before & after | DOCTRINE |
WHAT CHANGED SINCE THE CHEMISTRY PAGE
Chemistry conclusion survived the audit — the hardware plan evolved
- Chemistry: LiFePO4 still the winner (matches the v1 table). Only the vendor changed: Deye SE-F16 → DatouBoss 48V/314Ah at ~€119/kWh — same 16 kWh class, 210A BMS, ~€5,700 total vs €8,955 planned.
- MPPT: SmartSolar family swapped for BlueSolar (identical electricals, Bluetooth removed) — broadcast-site + home RF hygiene, biological considerations.
- Topology: shared battery bus chosen; escape hatch to 1-pack-per-Studer is rewiring only, no hardware swap.
- Transmission: combiner doctrine added — string fuses + main DC breaker + T2 SPD at the array, one 25 mm² bundle per MPPT over the 30m run, 8 conductors in the trench instead of 16.
- Audit fixes folded in: string ceiling = 32 panels (not 24), winter doubled-array yield ~19–20 kWh/day @ 24 panels, busbar trunk upsized for ~500A aggregate (2× 95 mm² or 120–150 mm²).
20-YEAR COST — UPDATED FOR DATOUBOSS
Normalized to ~43 kWh usable · replacements included
The vendor swap didn't change the chemistry conclusion — it improved it by ~37%. Flooded still ends the 20-year race at ~10× the DatouBoss cost.
EXPANSION ROADMAP
Growth without hardware regret — every stage reuses the fleet
- Stage 1 — baseline: 16 panels, 3 packs, 4 MPPTs. 43 kWh usable, clean 2/2/2/2 string loading. Winter baseline ~10 kWh/day.
- Stage 2 — doubling: +8 panels → 24 total, 3 strings per MPPT. Winter ~19–20 kWh/day, full bank recovery in bad weather. No controller change; resize combiner breakers/fuses FIRST.
- Stage 3 — ceiling: 32 panels on the same 4 MPPTs = ~20% midday clipping. Above 32 panels: +2× 150/45s (6 total, 19.2 kW). Escape hatches everywhere: spare PV ceiling (10.4→19.2 kW), parallel pack slots (up to 15), battery layout switchable by rewiring.
BUILD CHECKLIST
Live from PROJECT-CONTEXT.md v7 · tick as Avantida goes dark-grid
- Confirm daily consumption (kWh) via Studer logs/meter
- Order 3× DatouBoss 48V 314Ah (verify sale variant at checkout)
- Order 4× BlueSolar MPPT 150/45 (verify VE.Direct port; Skroutz quote)
- Combiner gear baseline: 8× 15A gPV fuses, 4× 25–32A DC breakers, 4× T2 SPDs
- EMC parts: sendust toroids, MnZn rings, clamp-on split ferrites
- Copper: 4× 25 mm² bundles (30m), 95 mm² Studer taps, 2× 95 mm² trunk
- Busbar + per-pack Class-T fusing layout
- Bulkhead entry plate (single penetration, bonded)
- Earth array frames + earth rod; mount combiner boxes
- Top-balance 3 packs → parallel onto shared bus
- Tune Studer profiles (abs 56.0V / float 54.0V)
- Pull 30m bundles → terminate at MPPTs (pigtails at terminals)
- Mount MPPTs >2m from receiver racks
- Baseline SDR sweep → chokes/filters → re-sweep verify
- Future: double PV — resize combiner protection first