Power
⚠️ Proposed — not final. Everything on this page (PSU sizing and counts, distribution, fusing, power injection, generators, and grounding) is a working proposal for planning purposes. The whole electrical design must be worked out and signed off by Alex (build) and a licensed electrician before it’s built. Numbers and approaches here are subject to change.
Generator capacity
Two ~9 kW generators (~18,000 W combined operating capacity).
Plan (2026-07-15): one generator runs the entire installation; the second is a dedicated backup for failover — not a load-share. This is viable only under a ≤50% white brightness cap (audit below): the whole show then draws ~6,700 W AC, about 74% of one 9 kW unit, leaving ~2,300 W for sound, crew, and charging.
Generator location & AC distribution (2026-07-15): the primary generator sits next to the front-right tower — all AC power originates at the front. AC feeds to the back towers run along the canopy lines (following the top diagonal cable paths, front → back). Because all power comes from the front, the canopy LEDs are fed from the front towers too (see #95) — powering them from the back only made sense when each end had its own generator. The long front→back AC runs (conductor sizing, voltage drop, GFCI, and transfer to the backup unit) are for the licensed electrician. This also collapses grounding to a single source — the two-generator shared-ground analysis (#34) is superseded; one generator = one ground reference, with all towers/scaffold bonded to it (equipotential).
Single-generator audit (≤50% white)
Full-white DC LED load is ~11,810 W across all zones (see the zone tables below). Addressable-LED power scales ~linearly with the brightness cap; AC = DC ÷ 0.91 PSU efficiency + ~162 W electronics overhead.
| Cap | LED DC | AC draw | One 9 kW gen |
|---|---|---|---|
| 100% full white | 11,810 W | ~13,100 W | ❌ needs both |
| 70% (earlier cap) | 8,270 W | ~9,250 W | ❌ ~103% — over |
| 60% | 7,090 W | ~7,950 W | ⚠️ ~88% — tight |
| 50% (mandated) | 5,905 W | ~6,700 W | ✅ ~74% |
| 30% (typical show) | 3,540 W | ~4,050 W | ✅ ~45% |
✅ Cap decided — 50% (mandated 2026-07-15, Nic H + Mark). The global port cap was revised down from 70% to 50% specifically to make single-generator operation work: at 50% the whole show draws ~6,700 W AC (~74% of one 9 kW unit), whereas 70% would draw ~9,250 W — over one unit. Enforcement is at the F48V5 port level (see the brightness-cap section below).
Backup generator: the second ~9 kW unit stands by for failover rather than carrying half the load, so a single-generator failure doesn’t go to a black show. Transfer scheme (manual vs automatic transfer switch), warm-standby vs cold-start, and switchover wiring are for the licensed electrician.
Caveats: confirm the real single-unit continuous (not peak) rating; aux loads (sound/crew/charging) are unquantified (open question #2) and eat the ~2,300 W cushion; feeding all 30 PSUs from one genset is a panel/breaker distribution job for the electrician (capacity ≠ distribution). All figures are full-white worst case within the cap — real shows sit ~20–40% (~4 kW), so there’s ample day-to-day margin.
Voltage strategy
The full installation runs 24V to minimize voltage drop over long cable runs. The rose window is the one exception — WS2815 is 12V-native, on its own dedicated 12V PSU. No converter needed: the F48V5 differential signal to each SmartReceiver is power-independent, so zones run at different voltages without issue.
| Zone | Voltage | Why |
|---|---|---|
| Arches (main, mini, quad) | 24V | Long runs, less voltage drop vs 12V |
| Spires & canopy (globe strings) | 24V | 24V RGB ordered |
| Rose window strips | 12V (dedicated PSU) | WS2815 is 12V-native; isolated from 24V |
| Controllers / SmartReceivers | Per device spec | F48V5: 12V; SmartReceivers: 12V input via buck converter from 24V rail |
Data signal is always 5V regardless of strip voltage. The WS2811 IC internally regulates to 5V for its logic regardless of whether the strip runs at 12V or 24V. Falcon SRs (powered at 12V) output a 5V data signal — no level shifter is needed between an SR and a 24V WS2811 strip. The 24V only powers the LEDs.
Rose window power — 12V system
WS2815 at 60/m draws 12W/m at full white.
| Strip length (16 petals) | ~144 m active (16 petals × ~9 m each) |
| Full white | ~1,728W |
| Typical show (20–40%) | ~350–700W |
PSU: Mean Well HLG-320H-12 — 12V / 26.7A / 320W each (IP65). Plan: 4 units, one per SRx1 (4 petals each).
| Config | Total capacity | Brightness headroom |
|---|---|---|
| 2 units (640W) | supports ~37% | Adequate for most show looks |
| 3 units (960W) | supports ~55% | Comfortable full-show buffer |
| 4 units (1,280W) | supports ~74% | Planned — one PSU per group of 4 petals |
Power injection: Both ends of every petal required — petal strips are 9m and show visible sag at midpoint without dual-end injection. Each PSU feeds both ends (hub and rim) of its 4 petals. See Wiring — Rose Window Power Injection for details.
Brightness cap — 50% on every port (set at the F48V5 port level)
Global cap (mandated 2026-07-15, Nic H + Mark; revised down from 70%): two drivers — 100% is too bright in the dark playa night, and the whole show must fit on one generator (the second is a backup), which requires ≤50% white. So all 48 F48V5 ports are set to 50% brightness — a uniform hardware ceiling for the whole show that also bounds the worst-case draw to ~6,700 W AC (see Generator capacity). 50% is well under the rose window’s PSU-safety limit (~74% ceiling; ports 33–48, issue #67), so the rose PSUs gain headroom too.
The cap is enforced at the F48V5 port level — each port set to 50% brightness in the controller config (see F48V5-config-runbook.md).
Why the port level and not the xLights model: the port cap is a true hardware backstop — it scales the physical output regardless of what any sequence contains, so it protects the PSUs against every show file, community submission, and raw test pattern. It cannot be bypassed by a sequence.
⚠️ Do NOT also set a model-level Max Brightness in xLights. A model dimming curve is baked into the rendered output and would compound with the port cap (0.50 × 0.50 = 0.25), under-driving the window to ~25%. Pick one enforcement point — we use the port. (This supersedes earlier guidance in this doc that said to set Max Brightness on the Rose Window A/B models.)
The PSUs have built-in overcurrent protection (OCP) as a further hardware backstop — they won’t fail dangerously — but OCP means the zone goes dark rather than limiting gracefully, so the 50% port cap is the right first line of defense.
Power implication of the global 50% cap: the real worst-case draw is ~50% of the full-white figures used throughout this doc (e.g. system full-white ~13,000 W → ~6,700 W AC actual ceiling; each quad’s ~20 A → ~10 A). This is the basis of the single-generator plan (see Generator capacity) and adds headroom on PSUs, fuses, and wire — but keep all hardware sized to the full-white numbers anyway (the cap is a config setting, and sizing to worst case is the safe practice).
Rose window PSU assignment
The rose window uses 4× SRx1 (one petal per dedicated port, ports 33–48) — see Rose window controller architecture. 4 PSUs, one per SRx1 (4 petals each). Each PSU serves 4 petals × both ends = 8 injection points.
| PSU | SRx1 (ports) | Petals | Max brightness |
|---|---|---|---|
| PSU-A | SRx1 #1 (33–36) | 1–4 | ~74% |
| PSU-B | SRx1 #2 (37–40) | 5–8 | ~74% |
| PSU-C | SRx1 #3 (41–44) | 9–12 | ~74% |
| PSU-D | SRx1 #4 (45–48) | 13–16 | ~74% |
Because each SRx1 is a separate board powered independently from its own 12V PSU, there is no shared-rail wiring hack — the earlier single-SRx4 design needed isolated V+ rails inside one enclosure; with four SRx1 boards each PSU simply feeds its own board. Much simpler.
| Spec | Value |
|---|---|
| Output per unit | 12V / 26.7A / 320W |
| Two units combined | 640W |
| IP rating | IP65 |
| Input | 90–305V AC |
Main arch PSU — per-foot injection
Each main arch leg gets its own PSU at the arch foot. The PSU powers a 6m straight leg plus a 5–6m curve half, with current meeting at the crown (V+ cut at crown). 10 PSUs total (2 per arch × 5 arches), all at ground level — no AC power on scaffold.
PSUs: E–N (E/F = arch 1 front, G/H = arch 2 … M/N = arch 5 back; odd letter = right leg, even = left leg).
Mean Well HLG-320H-24 — 24V / 13.3A / 320W, IP67 sealed. Provides headroom above the 7.2A peak per foot and is dust/weather-safe for playa use.
Minor arch PSU — 2 per side
Each side (left and right, 5 arches each) is served by 2 PSUs, with three injection points per arch: right foot, peak T-connector, and left foot. Single PSU per arch — no V+ cut.
All 5 arches on one side draw ~33A at full white, requiring more than one HLG-320H-24 (13.3A). Two HLG-600H-24B (25A each) cover full white with headroom.
PSUs: O–R
| PSU | Side | Arches | Full-white draw |
|---|---|---|---|
| PSU-O | Left (FL) | Arches 1 + 2 (front two) | ~15.2A |
| PSU-P | Left (FL) | Arches 3 + 4 + 5 (back three) | ~17.8A |
| PSU-Q | Right (FR) | Arches 1 + 2 (front two) | ~15.2A |
| PSU-R | Right (FR) | Arches 3 + 4 + 5 (back three) | ~17.8A |
4 HLG-600H-24B total (2 per side) for all 10 minor arches.
Recommended: Mean Well HLG-600H-24B — 24V / 25A / 600W, IP67, ~$110–130 each. Consistent with the general 24V system PSU.
Tower PSU — per tower
Each tower runs two rails: a 24 V rail (tower-mounted HLG-600H-24B) for the spire, quad faces, and — on back towers — canopy; and a 12 V rail (HLG-185H-12) for the spirelet + wash floods, which moved to 12 V on 2026-07-15. The 12 V flood rail is separate from the SR-board buck (which stays isolated for SR logic).
24 V rail
Front towers (FL, FR) — 1 PSU each (PSU-S, PSU-T):
| Component | Detail | Full white |
|---|---|---|
| Spire | 104 globe nodes | 3.1A |
| QUAD (4 faces) | 34m strip | 20.4A |
| Total | 23.5A |
One HLG-600H-24B (25A) per front tower — 94% at full white, ~47% under the 50% cap. 2 PSUs.
Back towers (BL, BR) — 2 PSUs each (lower quad on its own; the rest share):
| Component | PSU | Detail | Full white |
|---|---|---|---|
| Spire + QUADTOP + canopy | top (U / W) | 104 nodes + 17.6m + 126 nodes | 17.5A |
| QUADBOT (4 faces) | bottom (V / X) | 33.2m strip | 19.9A |
Two HLG-600H-24B per back tower. 4 PSUs. Under the 50% cap the bottom-quad PSU runs ~40%, the top PSU ~35% — comfortable. 6 HLG-600H-24B total for the towers.
12 V flood rail — spirelets + wash (new, 2026-07-15)
Wash and spirelets are the same fixture — the 12 V (10 W) variant of the spirelet fixture — powered by dedicated HLG-185H-12 (12 V / 15 A) PSUs at the tower enclosures, kept off the SR-board buck.
| Tower | Floods | Load @ 10 W | PSU |
|---|---|---|---|
| FR / FL (front) | 7 wash + 6 spirelets = 13 | ~11 A | 1× HLG-185H-12 each |
| BR / BL (back) | 4 spirelets | ~3 A | 1× HLG-185H-12 each |
6 HLG-185H-12 total (4 in use + 2 spares).
Full PSU summary
| PSU | Zone | Model | Count | Notes |
|---|---|---|---|---|
| A–D | Rose window | HLG-320H-12 | 4 | 12V; 1 per SRx1 (4 petals each). All 4 in hand. |
| E–N | Main arches | HLG-320H-24 | 10 | 1 per leg × 5 arches × 2 legs; E/F = arch 1 … M/N = arch 5 |
| O–R | Minor arches | HLG-600H-24B | 4 | O = FL arches 1–2, P = FL arches 3–5, Q = FR arches 1–2, R = FR arches 3–5 |
| S | FR tower (QUAD + spire) | HLG-600H-24B | 1 | 23.5A full white |
| T | FL tower (QUAD + spire) | HLG-600H-24B | 1 | 23.5A full white |
| U | BR tower top (QUADTOP + spire + canopy) | HLG-600H-24B | 1 | 17.5A full white |
| V | BR tower bottom (QUADBOT) | HLG-600H-24B | 1 | 19.9A full white |
| W | BL tower top (QUADTOP + spire + canopy) | HLG-600H-24B | 1 | 17.5A full white |
| X | BL tower bottom (QUADBOT) | HLG-600H-24B | 1 | 19.9A full white |
| — | Spirelet + wash floods (12 V rail) | HLG-185H-12 | 6 | 12V; 1 per tower (4) + 2 spares. New 2026-07-15. |
| Total | 30 | 20× 24V + 4× 12V rose + 6× 12V flood |
Arch strips power — 24V system
All arch strips: WS2811 24V 60/m. At full white: ~14.4 W/m.
| Zone | Strip length | Full white |
|---|---|---|
| Main arches (5) | ~114 m | ~1,642W |
| Mini arches (10) | ~265 m | ~3,816W |
| Quad arches (24 faces) | ~264 m | ~3,802W |
| Arch total | ~643 m | ~9,260W |
The 60/m spec roughly doubles arch power vs. an earlier 30/m design. At 50% show brightness, arch draw drops to ~4,600W. Warm glow (20%) is ~1,850W.
Globe strings power — 24V system
~0.72W per globe node (estimate — verify against product listing).
| Zone | Nodes | Full white |
|---|---|---|
| Spires (4 × 104 px) | 416 | ~300W |
| Canopy (252 px) | 252 | ~182W |
| Globe total | 668 | ~482W |
Flood fixtures power — 12V system
Spirelets + wash — the 12 V spirelet fixture, 10 W each (own 12 V flood rail).
| Zone | Fixtures | Full white |
|---|---|---|
| Spirelets | 20 | ~200W |
| Wash | 14 | ~140W |
| Flood total | 34 | ~340W |
Total AC draw (what the generators see)
Includes electronics overhead (~162W) and PSU efficiency (~91%).
Electronics overhead (~162W breakdown):
| Component | Count | Est. draw | Total |
|---|---|---|---|
| Falcon F48V5 controller | 1 | ~20W | ~20W |
| SRx4 SmartReceivers | 4 | ~5W each | ~20W |
| SRx2 SmartReceivers | 3 | ~5W each | ~15W |
| SRx1 SmartReceivers | 8 | ~3W each | ~24W |
| Buck converters (24V→12V, one per 24V SR enclosure; rose-window SRx1s run 12V, no buck) | 11 | ~5W each | ~55W |
| WiFi access point (dedicated, playa AP) | 1 | ~15W | ~15W |
| Misc (cable losses, sensor nodes, misc) | — | — | ~35W |
| Total | ~162W |
| Scenario | AC watts needed |
|---|---|
| Full white, all zones | ~12,000W |
| 50% brightness | ~6,200W |
| 30% brightness | ~3,800W |
| Warm glow (20%) | ~2,600W |
Per the single-generator plan (see Generator capacity): one ~9 kW generator runs the whole show under a ≤50% white cap (~6,700 W AC), the second stands by as backup. The combined 18,000 W covers full white if both are ever run together, but the operating plan is one-primary-plus-failover at the 50% cap.
PSUs — 24V system
Recommended: Mean Well HLG-600H-24B
| Spec | Value |
|---|---|
| Output | 24V / 25A / 600W |
| Input | 90–305V AC (generator-friendly wide range) |
| IP rating | IP67 (sealed, playa-safe) |
| Approx. cost | ~$110–130 each |
| # Units | Total capacity | Notes |
|---|---|---|
| 8 | 4,800W | Warm glow through moderate shows |
| 12 | 7,200W | 50% brightness all arches |
| 16 | 9,600W | Full arch white with margin |
PSU count proposed 2026-07-15 (#44) — see the Full PSU summary above: 10× HLG-320H-24 (main arches) + 10× HLG-600H-24B (4 minor arches + 6 towers), + spares. (The capacity table above is illustrative; the summary is authoritative.)
Power injection — voltage drop
Long strip runs need power injection at multiple points. At 14.4W/m / 0.6A/m (24V, 60/m strips):
| Run length | Injection |
|---|---|
| Under 5 m | One end only |
| 5–10 m | Both ends |
| Over 10 m | Both ends + midpoint(s) |
| Zone | Face / half length | Injection |
|---|---|---|
| Main arch leg | 6 m | Both ends |
| Main arch curve | 6 m | Both ends |
| Minor arch 1 half | 6.5 m | Both ends |
| Front QUAD face | ~8.5 m | Both ends (60/m density doubles effective current vs. earlier 30/m estimate) |
| Back QUADTOP face | ~4.4 m | One end sufficient |
| Back QUADBOT face | ~8.3 m | Both ends |
| Canopy diagonal | 22 m | Single-end from back tower — test on-site; add front-tower tap if far globes dim |
| Rose window petal | 9 m | Both ends |
Injection-tap fusing
Proposed approach (2026-07-13): keep the installed 5 A SR output fuses; power all quads from raw taps at both ends. The SR output fuse is 5 A (manual, #52), and even half a quad exceeds it — and upsizing the fuse risks the SR board trace. So every quad’s V+ comes from the 24 V PSU bus at both feet-pair corners, each fused: the data-in corner via the SR pigtail’s V+ conductor (landed on the bus inside the enclosure, not the SR board — don’t pull the strip-side pin), and the far corner via a separate power-only tap (V+/GND). No quad current passes through the SR. GND is common (bus GND ↔ PSU GND ↔ SR GND). (Gauge note: the made-up pigtail V+ conductor at the data-in end must be sized for its share of the ~20 A quad load — see #53.) Rose petals keep the SR V+ at the near end (rose SRx1 is 12 V, so ~2.2 A @ 50% legitimately runs through it, within the 5 A fuse) and raw-tap only the far end.
Sizing: fuse ≥ ~1.3× the tap’s current, next standard size, ≤ the tap wire’s ampacity. Each raw tap is a 2-conductor V+/GND run from the PSU bus to the injection corner.
| Zone | V | Strip full-white | Taps | Per-tap current | Inline fuse | Tap wire |
|---|---|---|---|---|---|---|
| Rose petal | 12V | 8.6 A (4.3 A @ 50% cap) | far end only | ~4.3 A | 5 A | 18 AWG |
| Front quad FR/FL | 24V | 20.4 A | both ends | ~10 A/end | 15 A | 12 AWG |
| Back-bottom quad BL/BR | 24V | 19.9 A | both ends | ~10 A/end | 15 A | 12 AWG |
| Back-top quad BL/BR | 24V | 10.6 A | both ends | ~5.3 A/end | 15 A (or 7.5 A / 14 AWG right-sized) | 12 AWG |
Fuse type: waterproof inline blade holders (ATO/ATC, 32 V DC — covers both 12 V and 24 V). Fuse at the tap origin (PSU-bus end). Back-top quads draw less, so their taps can be right-sized (7.5 A / 14 AWG) or kept uniform with the others (15 A / 12 AWG) for build simplicity.
This applies to every 24 V zone, not just quads
The quads were the first case worked out, but the same rule holds for all 24 V fixtures (arches, spires, canopy): the SR pigtail is data + GND only, and V+ is a fused feed from the 24 V PSU. The practical difference is scale, so the fusing hardware differs by zone:
- Quads — high current (~20 A), dedicated raw taps at both ends, individual inline blade holders (table above).
- Arches — already PSU-injected at the feet / T-connectors; the foot point labeled “SR pigtail” is a PSU V+ tap (data + GND from the SR), fused at the PSU. Each leg/half already has its own PSU.
- Spires / canopy — many small 24 V runs, low current. Each enclosure gets a fused distribution block off the tower 24 V PSU, with grouped fusing (one fuse for a tower’s 8 spire strings, one per canopy run).
- Spirelets + wash floods — 12 V (2026-07-15), so they’re not on the 24 V bus. Data still comes from the SR outputs, but V+ is fused off the tower’s separate 12 V flood PSU (see the Tower PSU section). Per-enclosure BOM + connector method for all of the above: SR enclosure design #53.
Common ground is mandatory everywhere: SR GND ↔ PSU GND ↔ fixture GND must be bonded, or the data signal has no reference. This is the silent failure mode of split power/data.
Power injection points — complete reference
All power injection points and PSU assignments by zone.
⚠️ On 24 V zones the SR carries DATA only — V+ always comes from the PSU
Falcon SmartReceivers accept 5–13 V and are bucked to 12 V on every 24 V tower, so anything driven off the SR board’s power would only get 12 V. The arches, spires, canopy, and quads are all 24 V — no LED current can flow through the SR. (Spirelets + wash are 12 V floods on their own dedicated 12 V PSU as of 2026-07-15 — still data-from-SR, but V+ from that flood PSU, not the SR buck.) The fix is done inside the enclosure, not at the strip:
- the strip’s data-in end still plugs into a normal, fully-populated 3-pin pigtail (V+/GND/Data) — do not pull the V+ pin at the strip; and
- inside the enclosure, that pigtail’s V+ conductor is landed on the fused 24 V bus, not the SR board’s output. Data comes from the SR channel, GND is common, V+ comes from the PSU bus. So the strip is powered through its pigtail as usual, but the current is sourced from the PSU and never passes through the SR.
- Dual-end zones (quads, rose) add a separate power-only tap (V+/GND, no data) at the far end.
The 24 V bus is fed from the tower PSU through a fused distribution block at the enclosure (see the fusing section above and SR enclosure design #53).
Reading the tables below: wherever a 24 V row names an “SR pigtail” injection point, read it as “the SR pigtail at that point, with its V+ sourced from the named PSU’s fused bus — not from the SR.” This corrects the earlier assumption (spires/spirelets/canopy) that small globe runs draw power through the SR; that would deliver 12 V to 24 V fixtures.
The rose window (12 V) is the only exception — its SRx1 board runs at 12 V, matching the WS2815 strips, so rose LED power does flow through the SRx1 pigtail (near-end), with a raw far-end tap.
Spires — 24 V; spirelets + wash — 12 V floods
Spires are 24 V globe strings. Per the rule above, each plugs into its normal SR pigtail but inside the enclosure the pigtail’s V+ is landed on the tower’s fused 24 V bus, not the SR — V+ from the tower 24 V PSU, grouped fusing (a whole tower’s spire ≈ 3 A). Spires are on group A of each tower’s SRx4.
Spirelets + wash are one fixture — the 12 V variant of the spirelet fixture (2026-07-15). Their data still comes from the same SR outputs (spirelets on the connectors in the table below; wash on the reserved 8E / 12E), but their V+ comes from the tower’s dedicated 12 V flood PSU (HLG-185H-12), fused off that rail — not the SR-board buck (kept isolated for logic). Same data-from-SR / V+-from-dedicated-PSU pattern as the 24 V zones, just on the 12 V rail. The PSU column below applies to the spires; spirelet rows draw V+ from the tower’s 12 V flood PSU.
| Tower | Component | SR Connectors | PSU |
|---|---|---|---|
| FR | Spire (8 strings, 13 nodes each) | 5A, 5A+1, 5A+2, 5A+3, 6A, 6A+1, 6A+2, 6A+3 | PSU-S |
| FR | Upper spirelets (FR-1–4) | 7A, 7A+1, 7A+2, 7A+3 | PSU-S |
| FR | Lower spirelets (FR-5–6) | 7E, 6E | PSU-S |
| FL | Spire (8 strings) | 9A, 9A+1, 9A+2, 9A+3, 10A, 10A+1, 10A+2, 10A+3 | PSU-T |
| FL | Upper spirelets (FL-3–6) | 11A+2, 11A+3, 12A, 12A+1 | PSU-T |
| FL | Lower spirelets (FL-1–2) | 10E, 11E | PSU-T |
| BR | Spire (8 strings) | 17A, 17A+1, 17A+2, 17A+3, 18A, 18A+1, 18A+2, 18A+3 | PSU-U |
| BR | Upper spirelets (BR-1–4) | 19A, 19A+1, 19A+2, 19A+3 | PSU-U |
| BL | Spire (8 strings) | 13A, 13A+1, 13A+2, 13A+3, 14A, 14A+1, 14A+2, 14A+3 | PSU-W |
| BL | Upper spirelets (BL-1–4) | 15A, 15A+1, 15A+2, 15A+3 | PSU-W |
Rose window petals — 12V (32 injection points)
Each petal is injected at both ends by the same PSU: the data-in end (pixel 0, at the 2a cell — SR xConnect pigtail) and a power-return end at the physical strip end (past the last mapped pixel / dark tail). Both ends sit near the hub center. Strips are left uncut (the ~0.4 m unaddressed tail stays dark). See Rose window PSU assignment — 4× SRx1, one petal per dedicated port (ports 33–48).
| Strand | Port | PSU | Data-in end (SR pigtail) | Power-return end (physical strip end) |
|---|---|---|---|---|
| RW-PETAL-1 | 33 | PSU-A | SR xConnect pigtail | Power-only Male pigtail |
| RW-PETAL-2 | 34 | PSU-A | SR xConnect pigtail | Power-only Male pigtail |
| RW-PETAL-3 | 35 | PSU-A | SR xConnect pigtail | Power-only Male pigtail |
| RW-PETAL-4 | 36 | PSU-A | SR xConnect pigtail | Power-only Male pigtail |
| RW-PETAL-5 | 37 | PSU-B | SR xConnect pigtail | Power-only Male pigtail |
| RW-PETAL-6 | 38 | PSU-B | SR xConnect pigtail | Power-only Male pigtail |
| RW-PETAL-7 | 39 | PSU-B | SR xConnect pigtail | Power-only Male pigtail |
| RW-PETAL-8 | 40 | PSU-B | SR xConnect pigtail | Power-only Male pigtail |
| RW-PETAL-9 | 41 | PSU-C | SR xConnect pigtail | Power-only Male pigtail |
| RW-PETAL-10 | 42 | PSU-C | SR xConnect pigtail | Power-only Male pigtail |
| RW-PETAL-11 | 43 | PSU-C | SR xConnect pigtail | Power-only Male pigtail |
| RW-PETAL-12 | 44 | PSU-C | SR xConnect pigtail | Power-only Male pigtail |
| RW-PETAL-13 | 45 | PSU-D | SR xConnect pigtail | Power-only Male pigtail |
| RW-PETAL-14 | 46 | PSU-D | SR xConnect pigtail | Power-only Male pigtail |
| RW-PETAL-15 | 47 | PSU-D | SR xConnect pigtail | Power-only Male pigtail |
| RW-PETAL-16 | 48 | PSU-D | SR xConnect pigtail | Power-only Male pigtail |
Major arches — 24V (20 injection points)
Four injection points per arch: right foot (SR pigtail), right-mid T-connector, left-mid T-connector, and left foot. V+ is cut at the crown — data and GND pass through uncut. One PSU per leg; same PSU feeds both the foot and its mid T-connector.
| Strand | SR Connector | Right foot | Right-mid T | Left-mid T | Left foot |
|---|---|---|---|---|---|
| MA-MAJARCH-1 | 29A | PSU-E (SR pigtail) | PSU-E (power arm) | PSU-F (power arm) | PSU-F (Male pigtail) |
| MA-MAJARCH-2 | 30A | PSU-G (SR pigtail) | PSU-G (power arm) | PSU-H (power arm) | PSU-H (Male pigtail) |
| MA-MAJARCH-3 | 31A | PSU-I (SR pigtail) | PSU-I (power arm) | PSU-J (power arm) | PSU-J (Male pigtail) |
| MA-MAJARCH-4 | 32A | PSU-K (SR pigtail) | PSU-K (power arm) | PSU-L (power arm) | PSU-L (Male pigtail) |
| MA-MAJARCH-5 | 29A+1 | PSU-M (SR pigtail) | PSU-M (power arm) | PSU-N (power arm) | PSU-N (Male pigtail) |
Minor arches — 24V (30 injection points)
Three injection points per arch: right foot (SR pigtail), peak T-connector, and left foot. All three share the same PSU — no V+ cut needed. One PSU serves 2–3 arches per group; arches within a group are wired in parallel from that PSU.
| Strand | SR Connector | PSU | Right foot | Peak T-connector | Left foot |
|---|---|---|---|---|---|
| FL-MINARCH-1 | 21A | PSU-O | SR pigtail | PSU-O power arm | PSU-O Male pigtail |
| FL-MINARCH-2 | 22A | PSU-O | SR pigtail | PSU-O power arm | PSU-O Male pigtail |
| FL-MINARCH-3 | 23A | PSU-P | SR pigtail | PSU-P power arm | PSU-P Male pigtail |
| FL-MINARCH-4 | 24A | PSU-P | SR pigtail | PSU-P power arm | PSU-P Male pigtail |
| FL-MINARCH-5 | 21A+1 | PSU-P | SR pigtail | PSU-P power arm | PSU-P Male pigtail |
| FR-MINARCH-1 | 25A | PSU-Q | SR pigtail | PSU-Q power arm | PSU-Q Male pigtail |
| FR-MINARCH-2 | 26A | PSU-Q | SR pigtail | PSU-Q power arm | PSU-Q Male pigtail |
| FR-MINARCH-3 | 27A | PSU-R | SR pigtail | PSU-R power arm | PSU-R Male pigtail |
| FR-MINARCH-4 | 28A | PSU-R | SR pigtail | PSU-R power arm | PSU-R Male pigtail |
| FR-MINARCH-5 | 25A+1 | PSU-R | SR pigtail | PSU-R power arm | PSU-R Male pigtail |
Quad arches — 24V (chained, 2026-07-12)
Each quad is now ONE chained data string (4 faces, Front→Left→Back→Right, pixel 0 = front-right foot) on one SR output — not four per-face connectors. The string loops clockwise back to its entry corner, so data-out meets data-in at the front-right foot (~0.46 m jumper). Power is injected at the SR corner (front-right foot) and, for the longer quads, at the opposite feet-pair corner across the quad. Both injections use the same tower PSU — no V+ cut. See wiring.md for the authoritative per-connector channel map.
| Quad | Chained connector | PSU | String length | Pixels | Injection |
|---|---|---|---|---|---|
| Front-top-right (FR) | 5E | PSU-S | ~34 m | 340 | SR corner + opposite corner |
| Front-top-left (FL) | 9E | PSU-T | ~34 m | 340 | SR corner + opposite corner |
| Back-top-right (BR) | 20D (spire SRx4 spare port) | PSU-U | ~17.6 m | 176 | SR corner (add opposite if far globes dim) |
| Back-top-left (BL) | 16D (spire SRx4 spare port) | PSU-W | ~17.6 m | 176 | SR corner (add opposite if far globes dim) |
| Back-bottom-right (BR) | 17E | PSU-V | ~33 m | 332 | SR corner + opposite corner |
| Back-bottom-left (BL) | 13E | PSU-X | ~33 m | 332 | SR corner + opposite corner |
Tower PSU assignments — PSU-S through PSU-X also power spires, spirelets, and (for PSU-U and PSU-W) canopy runs. See Tower PSU — per tower for full tower load breakdown.
Canopy — 24V (fed from the FRONT towers, #95 — 2026-07-15)
Canopy is fed from the front towers (data + power), because all AC originates at the front-right-tower generator (see Generator location & AC distribution). Each run’s data-in is at its front-tower end: the 2 upper runs on the spire SRx4, the 1 lower run on the restored low SRx2 (F). V+ is a fused feed from the front tower’s 24 V PSU. If the far (back) end dims, add a back-end tap — no extra PSU.
| Strand (was) | Front SR connector | Front PSU | Front-tower injection |
|---|---|---|---|
| BR→FR rising (was 20B) | 8A (spire SRx4) | PSU-S | SR pigtail at front-right tower |
| BL→FR diagonal (was 16B) | 8B (spire SRx4) | PSU-S | SR pigtail at front-right tower |
| BR→FR straight (was 20C) | 8F (low SRx2) | PSU-S | SR pigtail at front-right tower |
| BR→FL diagonal (was 20A) | 11A (spire SRx4) | PSU-T | SR pigtail at front-left tower |
| BL→FL rising (was 16A) | 11B (spire SRx4) | PSU-T | SR pigtail at front-left tower |
| BL→FL straight (was 16C) | 12F (low SRx2) | PSU-T | SR pigtail at front-left tower |
⚠️ Front-tower 24 V PSU load — needs a decision (Alex/electrician). Adding ~3 canopy runs (~3.8 A full white) to a front tower already at ~23.5 A (spire + quad) pushes it to ~27 A at full white, over the HLG-600H-24B’s 25 A rating. Under the mandated 50 % cap it’s ~13.6 A — fine — but the “size to full white” rule is violated. Options: accept it under the 50 % cap, add a small dedicated canopy PSU at the front, or up-size the front-tower PSU. The tower-PSU load tables above still show canopy on the back (PSU-U/W) and need re-derivation once this is settled.
Open questions
- Globe string wattage — Verify 0.72W/node against actual AliExpress product listing (ID: 3256805045342158).
- Other electrical loads — Sound, crew lighting, and charging on the generator circuit. Finalize once full crew power needs are known.
- Generator fuel — At ~10,000W continuous (full white), a pair of 10kW generators burn ~0.7–1.0 gal/hr combined. 8 hrs/night × 7 days = ~56–80 gallons total. Plan fuel accordingly.
- HLG-320H-12 (rose): 4 units in hand. 4 units (PSU-A through PSU-D), one per rose-window SRx1. Supports ~74% max brightness; held at the global 50% cap (F48V5 port level, not the xLights model).
- Arch PSU count — proposed; see Full PSU Summary above.
- Canopy injection — PSU plan assumes single-end injection from back towers. Test on-site: if far-end globes are visibly dim at show brightness, add a second injection wire run to the front towers. No extra PSU needed — tap the front tower PSU.