# Sphere electronics — parts by topology and location

Compiled 2026-08-16 for the 33-panel HairCut / hairband-15 layout.

**Sources.** Topology and per-stage counts from the *Power flow* / *Data flow*
sections of [NYOL 2026 Sphere][doc]. Run lengths from Vals's layout legend
(2026-08-09). Aurica quantities from Ross's photo count (2026-08-10) in the
*Sphere tutorial* thread. Ross's corrections 2026-08-15: one Big Switch not two;
power primaries are made in 10 ft sizes only (20/30/40); a longer cable may serve
a shorter run; "12 units in theory, somewhere"; want ≥1 spare of most parts.

**Column meaning.** *Need* is for the built sphere. *Aurica* is what was
photographed. *Ross* is what he took home. *Unlocated* is believed-to-exist but
not seen — mostly Spire, which nobody has inventoried. *Still needed* assumes
unlocated stock is **not** found.

[doc]: https://docs.google.com/document/d/1kdji4jCu4IWLoeJxBEGNbrlMVP5zCv3QYEXdp_nLZ9E/edit

---

## POWER — base to strip

| # | Stage | Spec | Need | Aurica | Ross | Unlocated | Still needed |
|---|---|---|---|---|---|---|---|
| 1 | Batteries + solar | camp electronics | — | — | — | — | out of scope here |
| 2 | **Big Switch** | XL in, 48 V ≤60 A → 6 × XT60 | 1 | 1 (6-out) | — | — | **0** |
| 3 | **XT60 Y-splitter** | 1 → 2 | 1 (6-out switch → 7 feeds) | — | — | — | **1 — fabricate** |
| 4 | **Power primary 30′** | XT60, ≤10 A @ 48 V | 6 | 3 | — | — | **3** |
| 4 | **Power primary 40′** | ″ | 1 | 0 | — | — | **1** |
| 4 | *Power primary 20′* | ″ | 0 | 2 | — | — | idle — reaches nothing |
| 4 | *Power primary T10* | ″ | 0 | 1 | — | — | idle |
| 5 | **Tobsun 240 W DC-DC** | XT60 48 V ≤5 A → XT90 12 V ≤20 A | 11 | 11 | — | 1 | **0** (spare only if 12th found) |
| 6 | **Power octopus** | XT90 in → 8 × XT90 12 V ≤5 A | 11 | 11 | — | 1 | **0** (spare only if 12th found) |
| 7 | LED strips | fed from both ends | 33 panels | — | — | — | see panel fab |

The two 25′ runs round up to 30′ because primaries are only made in 20/30/40.
That is why *need* reads 6 × 30′ rather than 4 × 30′ + 2 × 25′.

## DATA — base to strip

| # | Stage | Spec | Need | Aurica | Ross | Unlocated | Still needed |
|---|---|---|---|---|---|---|---|
| 1 | **Control PC** | HIGOLE mini PC, Celeron N5095, 4 × USB 3.0, 1 × ethernet, fanless | 1 | **3 held** (Ross, 2026-08-16) | — | — | **0** (2 spare) |
| 2 | **USB hub** | active hub, not a Y-cable | see note | — | — | — | **1 × 4-port** (or 3 × 1→2) |
| 3 | **USB → Cat5/6** | USB over twisted pair | 7 | ✓ held (Ross, 2026-08-16) | — | — | **0** |
| 4 | **Cat6 25′** | RJ45, up truss | 2 | 4 | — | — | **0** (2 spare) |
| 4 | **Cat6 30′** | ″ | 4 | 3 × 35′ | +2 × 35′ purchased | — | **0** (1 spare) |
| 4 | **Cat6 40′** | ″ | 1 | — | 1 × 50′ | — | **0** |
| 5 | **Adafruit RP2040 Scorpio** | ethernet in, 2 × 8 pin out | 7 | — | — | 8 (see below) | **0 if found, else 7** |
| 6 | **Data splitter** | ribbon 2 × 8 in → 6–8 × JST SM4 | 7 | — | — | ? | **unknown** |
| 7 | LED strips | JST SM4 into upstream end | 33 | — | — | — | see panel fab |

Longer cable serves a shorter run, so the 35s cover 30′ runs and the 50 covers
the 40′. Only one 30′ run is left uncovered.

### The Scorpio lead

Ross's hand-drawn plan of 2025-07-05 (photo in #4πsr, 2026-07-06) lists, under
**"IN WHITE 0"**: *11 Power Unit, 11 Power Octopus, **8 Scorpio**, 2.5 Data
Octopus, 2 × 25′ Data, 2 × 35′ Data, 1 × 50′ Data, 3 × 30′ Power, 2 × 20′ Power,
2 × T5 Power, 1 × MVP set.*

The power/data cable counts in that list match Ross's 2026 Aurica photo count
almost exactly, which suggests White 0 **is** the bin that was photographed — and
therefore that **8 Scorpios and ~2.5 data octopi were in it**. They do not appear
in the 2026 count, so either they were missed in the photos or they have moved.
Worth opening White 0 before ordering seven Scorpios.

---

## Still needed — the short version

**Certain, fabricate:**
- 3 × 30′ power primary
- 1 × 40′ power primary
- +1 × 40′ if honouring the spare rule (a 40 covers any run; a 30 does not)

**Certain, buy:**
- **1 × 4-port USB hub** — not 4. The PC has 4 root ports and the chain needs 7
  data lines, so the deficit is 3. One 4-port hub on one root port yields 4 lines
  and leaves 3 root ports direct: 4 + 3 = 7. With 1→2 splitters instead you'd
  need 3 of them (each nets +1). Either way, **active hub, never a passive
  Y-cable** — those pass data on one leg only and fail silently.
  Worth testing one USB-over-Cat adapter *behind* the hub before committing;
  some extenders dislike being downstream of one.
  Bandwidth is not a consideration when choosing the hub — the chain's
  constraint is at the Scorpio exit, not upstream of it (Vals, 2026-08-16). A
  second hub is worth having purely as a spare.
  **Root-port count verified 2026-08-16:** the PC is a HIGOLE mini
  ([B0CZ3V9G48](https://www.amazon.com/dp/B0CZ3V9G48)) with **4 × USB 3.0**,
  1 × ethernet, Celeron N5095, passively cooled. Three are held. The 4-root-port
  assumption the hub arithmetic rests on is therefore confirmed, not assumed.

**Certain, fabricate (added 2026-08-16):**
- 1 × XT60 Y-splitter — the Big Switch has 6 outputs and the layout wants 7
  feeds, so one output splits.

**Unknown until someone looks:**
- 7 × Scorpio — probably in White 0, see above
- 7 × data splitter — no record
- 12th Tobsun and 12th octopus — the difference between a spare and none

**Resolved 2026-08-16 (Ross):** control PCs held; USB-over-RJ45 adapters held;
data cable complete after purchasing 2 × 35′ and 1 × 50′.

**Idle stock that fits nothing:** 2 × 20′ and 1 × T10 power primary.