Transmission media & connectors
Which cable, which fibre, which wireless band — and what's actually wired to which pin once you get to the connector. Structured cabling, fibre types, wireless control, RS-485's "unit load," and pinouts for the plugs you carry every day.
Structured cabling — Ethernet Cat5e to Cat8
| Category | Max frequency | Max standard @ 100 m | Show-network notes |
|---|---|---|---|
| Cat5e | 100 MHz | 1000BASE-T (1 Gbit/s) | Legacy DMX/sACN-over-Ethernet still works; skip it for anything you are pulling new |
| Cat6 | 250 MHz | 1G full distance; 10GBASE-T to ≈37–55 m | Common general-purpose AV/lighting backbone |
| Cat6a | 500 MHz | 10GBASE-T full 100 m | Full-distance Dante/NDI/converged AV — see the network planner |
| Cat7 | 600 MHz | 10GBASE-T 100 m | Individually shielded pairs, but its native connector (GG45/TERA) is not RJ45 — rare on tour |
| Cat8 | 2000 MHz | 25/40GBASE-T to 30 m | Short data-centre-style runs — video-over-IP core switching, not front-of-house distance |
On a show network the failure that actually bites is almost always termination quality — a re-crimped end, a kinked pair, a run laid across a dimmer feed — not the category printed on the jacket. Per EIA/TIA-568.
Fibre — multimode OM1–OM5, singlemode OS1/OS2
| Type | Core / cladding | Jacket | ~10G reach | Notes |
|---|---|---|---|---|
| OM1 | 62.5/125 µm multimode | orange | ≈33 m | Legacy, LED source |
| OM2 | 50/125 µm multimode | orange | ≈82 m | Legacy |
| OM3 | 50/125 µm laser-optimised multimode | aqua | ≈300 m | Common video-over-fibre, 850 nm VCSEL source |
| OM4 | 50/125 µm laser-optimised multimode | aqua / violet | ≈400 m | Higher-bandwidth backbone runs |
| OM5 | 50/125 µm wideband multimode | lime green | ≈400 m | Carries several wavelengths at once (SWDM) over fewer fibres |
| OS1 | 9/125 µm singlemode | yellow | ≈10 km | Indoor, tight-buffered |
| OS2 | 9/125 µm singlemode | yellow | tens of km | Outdoor, loose-tube — site-to-site, OB truck runs |
Jacket colour is an industry convention (TIA-598), not physics — an unlabelled coil should be confirmed with the printed cable spec or a light source and power meter before it goes into a rig. Per ITU-T G.651.1/G.652 and TIA-492.
Wireless control — Wi-Fi, Bluetooth, Zigbee
| Generation | IEEE | Year | Max PHY rate | Band |
|---|---|---|---|---|
| Wi-Fi 4 | 802.11n | 2009 | ≈600 Mbit/s | 2.4 / 5 GHz |
| Wi-Fi 5 | 802.11ac | 2013 | ≈3.5 Gbit/s | 5 GHz |
| Wi-Fi 6 / 6E | 802.11ax | 2019 / 2020 | ≈9.6 Gbit/s | 2.4 / 5 / 6 GHz |
| Wi-Fi 7 | 802.11be | 2024 | ≈46 Gbit/s | 2.4 / 5 / 6 GHz, multi-link |
6E/7's 6 GHz band matters most in venues: it is far less congested than 2.4/5 GHz, so a show-network Wi-Fi doesn't have to compete with a room full of phones — but it does not remove the need to keep control-network Wi-Fi off the same spectrum as anything RF-sensitive nearby. Per IEEE 802.11.
Bluetooth is a short-range 2.4 GHz link that frequency-hops across 79 channels roughly 1,600 times a second to dodge interference. Classic Bluetooth carries streamed audio (some in-ear/monitor products); Bluetooth Low Energy (BLE) trades bandwidth for very low power, and is what most modern consoles and remote apps use to pair a tablet to a desk. Zigbee (built on the IEEE 802.15.4 physical layer, also 2.4 GHz) is a low-power mesh network — many battery nodes relay through each other rather than each talking directly to a hub — common in architectural/permanent-install lighting and building sensor networks.
Wi-Fi / Bluetooth — star
Every device talks only to the access point / hub (filled centre dot). Simple, but the hub is a single point of failure and every device needs it in range.
Zigbee — mesh
No hub — nodes relay through each other, with more than one path between any two. No single node losing power takes the network down.
Wireless DMX/RDM systems share the 2.4 GHz ISM band with Wi-Fi/Bluetooth/Zigbee but are not built on any of these standards — they are purpose-built proprietary RF layers carrying DMX512 data, which is why coordinating channel/frequency use across all of them at a venue is a real RF-planning task, not just "everything is 2.4 GHz so it's fine."
RS-485 and DMX512, as numbers
The electrical limits, for when the question is whether a run is legal rather than why. The DMX explainer covers what a unit load actually is, why the line needs terminating, and why an unterminated rig often works anyway.
| Property | Limit | What bites |
|---|---|---|
| Devices per line | 32 unit loads | A fixture is not necessarily one unit load; some draw two or a half. The count that matters is unit loads, not boxes. |
| Cable length | Up to 1200 m at low rates | DMX runs at 250 kbit/s, so the practical figure quoted is around 300 m before a splitter — and marginal lines fail RDM long before they fail DMX. |
| Characteristic impedance | 120 Ω | Microphone cable is 45–75 Ω. It works until it does not, and the failure looks like flicker rather than a fault. |
| Terminator | 120 Ω at the far end | One, at the end of the line. Two, or one in the middle, is worse than none. |
| Differential voltage | ±1.5 V minimum at the receiver | What long or thin cable eats into, which is why the symptom is distance-dependent. |
| Slots per universe | 512 | Plus the start code, which is slot zero and is not a channel. Frame time depends on how many you actually send. |
Connector pinouts
HDMI Type-A — single row, 19 pins
The cable is not automatically the generation number on the box. Full HDMI 2.1 bandwidth needs a cable certified Ultra High Speed — an older "high speed" cable will physically fit and often still show an image, just not at the resolution/refresh the source can produce.
DVI-D dual-link — 3 rows × 8, 24 pins
DVI-I adds 4 analogue pins plus a flat ground blade around this same 24-pin block for VGA compatibility; DVI-D omits them entirely, so a DVI-D plug physically cannot go into an analogue-only DVI-A socket. Per the DDWG DVI 1.0 specification.
USB-C — 2 rows × 12, mirrored so the plug works either way up
Simplified for clarity — see the USB-IF Type-C spec for the authoritative per-pin table. The point to see here is the mirroring: row B is row A's function set again, which is what lets the plug work flipped either way up. There is no separate "Thunderbolt connector" — Thunderbolt 3 and 4 run over this exact 24-pin USB-C connector; Thunderbolt 1/2 used the older 20-pin Mini DisplayPort connector instead.
Open an issue — this page is hand-authored reference content rather than data pulled from the dataset, so corrections go straight to a pull request against this file.