Stoatworks Labs

Reference

What is actually inside a four-channel CAT5 stagebox

A balanced audio channel needs one twisted pair. A Cat cable has four. So a four-channel analogue stagebox over Cat5 is not a clever product — it is a wiring table. The trouble is that everybody wrote a different one.

There are nine pinouts catalogued in AES72-2019, and the standard deliberately recommends none of them over the others. Radial, SoundTools, Whirlwind, Neutrik and Thomann between them use five different maps, two of which are not in the standard at all. Every one of them works perfectly with itself and none of them is wrong. Plug a box from one into a box from another and you still get four channels of clean, quiet, full-bandwidth audio — just not necessarily the channels you asked for, and not necessarily the right way up.

This page is the map. Everything on it is derived from one table of pin assignments, the interoperability matrix is computed rather than typed, and each row carries the source it came from and how much that source is worth.

The shape of the problem

Eight conductors, four pairs, four balanced channels — and that consumes the cable exactly. There is no ninth conductor left for a ground, so XLR pin 1 on all four channels is commoned and rides the cable screen. Every system on this page does this, without exception.

Anatomy of a four-channel CAT5 audio breakoutAn eight-pin RJ45 connector on the left. Its four twisted pairs — pins 1 and 2, 3 and 6, 4 and 5, 7 and 8 — fan out to four XLR connectors on the right, one pair per channel carrying XLR pins 2 and 3. The cable screen runs separately to XLR pin 1, which is commoned across all four channels.RJ45 / etherCON12345678pair 1/2pair 3/6pair 4/5pair 7/8screen → pin 1, all fourChannel 1pin 2 hot, pin 3 cold, pin 1 screenChannel 2pin 2 hot, pin 3 cold, pin 1 screenChannel 3pin 2 hot, pin 3 cold, pin 1 screenChannel 4pin 2 hot, pin 3 cold, pin 1 screen
The physical arrangement is not in dispute and never has been. What varies between manufacturers is only which pair is called channel 1, and which end of each pair is called hot.
  • Every scheme keeps each channel on a true twisted pair. The pairs are pins (1,2), (3,6), (4,5) and (7,8) in T568A and T568B alike — the two standards differ only in which colour lands on which pin. This is why mixing brands degrades nothing: the balanced interface still gets its twist, so common-mode rejection and crosstalk are unaffected.
  • Shielded cable is mandatory, not advisory. The screen is the only pin-1 conductor, so it is also the only phantom power return. On UTP, 48V phantom cannot work at all — and neither can the ground reference that keeps the whole thing quiet.
  • T568A versus T568B does not matter, as long as both ends match. These boxes are wired to pin numbers. Terminate one end A and the other B and you have transposed the orange and green pairs, which swaps two channels.

The nine types in AES72-2019

The naming is more helpful than it first looks. The digit is the channel-to-pair order; the letter is the polarity convention — E puts every hot on an even-numbered pin, O puts every hot on an odd-numbered pin, and M mixes them. So 1E, 1O and 1M are one channel order wired three different ways, and the whole zoo collapses into two questions: which pair is channel 1, and which way up.

TypePin 1Pin 2Pin 3Pin 4Pin 5Pin 6Pin 7Pin 8Polarity
Type 1E1-1+2-3+3-2+4-4+Hot on even pins
Type 1O1+1-2+3-3+2-4+4-Hot on odd pins
Type 1M1+1-2+3+3-2-4+4-Mixed
Type 2O2+2-3+1-1+3-4+4-Hot on odd pins
Type 3E3-3+2-1+1-2+4-4+Hot on even pins
Type 3M3+3-2+1+1-2-4+4-Mixed
Type 4E1-1+3-2+2-3+4-4+Hot on even pins
Type 5E3-3+4-1+1-4+2-2+Hot on even pins
Type 6E4-4+3-2+2-3+1-1+Hot on even pins

Read 2+ as "channel 2, XLR pin 2, hot" and 2− as "channel 2, XLR pin 3, cold". Swatches show T568B colours. The standard assigns type numbers so a manufacturer can state which one a product uses — but stops short of preferring one, which is precisely why there are still nine.

What each box actually does

Radial Catapult

TX4 / RX4, TX4L / RX4L, Catapult Mini, Catapult Rack

No AES72 type

  • Ch 1hot pin 7, cold pin 8 · pair 7/8
  • Ch 2hot pin 4, cold pin 5 · pair 4/5
  • Ch 3hot pin 3, cold pin 6 · pair 3/6
  • Ch 4hot pin 1, cold pin 2 · pair 1/2

Type 6E channel order, mixed polarity — not one of the nine

Radial number their channels from the far end of the connector: channel 1 is on the brown pair, channel 4 on the orange. It is Type 1M with the channel numbering reversed, which is why it interchanges with a Catdusa at correct polarity but in reverse order.

Manufacturer documentation — Radial Engineering, TX4-RX4 specifications

SoundTools CAT Box

CAT Box, CAT Box MINI, CAT Rack, RATCAT series

AES72 Type 2O

  • Ch 1hot pin 5, cold pin 4 · pair 4/5
  • Ch 2hot pin 1, cold pin 2 · pair 1/2
  • Ch 3hot pin 3, cold pin 6 · pair 3/6
  • Ch 4hot pin 7, cold pin 8 · pair 7/8

The only one of these whose own documentation names its AES72 type: "All SoundTools Analog Audio Over Shielded Cat5e products are AES72-2O compliant". Note that channel 1 is on the blue pair, not the orange one — SoundTools start at pin 4.

Manufacturer documentation — SoundTools, Analog Audio over CAT wiring diagram (v3, 2025)

Whirlwind Catdusa

Catdusa, CatdusaDMX

AES72 Type 1M

  • Ch 1hot pin 1, cold pin 2 · pair 1/2
  • Ch 2hot pin 3, cold pin 6 · pair 3/6
  • Ch 3hot pin 4, cold pin 5 · pair 4/5
  • Ch 4hot pin 7, cold pin 8 · pair 7/8

The mapping shown here is what the hardware is reported to do and is the only reading consistent with the manual's own sentence about twisted pairs. The manual's wiring table says something different and impossible — see below.

Manufacturer documentation, self-contradictory — Whirlwind, Catdusa RJ45 Breakout Box manual

the sssnake Cat Snake

Thomann Cat Snake breakouts and inline adapters

AES72 Type 1E

  • Ch 1hot pin 2, cold pin 1 · pair 1/2
  • Ch 2hot pin 6, cold pin 3 · pair 3/6
  • Ch 3hot pin 4, cold pin 5 · pair 4/5
  • Ch 4hot pin 8, cold pin 7 · pair 7/8

Thomann publish this nowhere — not on the product pages, the spec tables, the downloads or the reviews, across the whole Cat Snake range. A community table is the best paper source that exists for it, so treat it as unconfirmed and meter a unit before you build anything to match it.

Community-sourced, unconfirmed — GroupDIY, "AES72… Manufacturers and their Pinouts"

Neutrik NA-4I4O-AES72

and Ward-Beck QTP, Stageeasy QuadCAT

AES72 Type 4E

  • Ch 1hot pin 2, cold pin 1 · pair 1/2
  • Ch 2hot pin 4, cold pin 5 · pair 4/5
  • Ch 3hot pin 6, cold pin 3 · pair 3/6
  • Ch 4hot pin 8, cold pin 7 · pair 7/8

Neutrik state the type in the specification: "Assignment according to AES72 (Type 4E)". This and the sssnake differ by nothing but a channel 2 / channel 3 transposition.

Manufacturer documentation — Neutrik, NA-4I4O-AES72 product specification

The Catdusa's manual is wrong

Whirlwind's Catdusa manual contains two statements on the same page that cannot both be true. The prose says there are "four circuits on the Catdusa box, using the four twisted pairs of the Cat cable". The wiring table immediately below it assigns the pins sequentially:

That table puts channel 2 on pins 3 and 4, and channel 3 on pins 5 and 6. Neither is a twisted pair. Pin 3 is twisted with pin 6, and pin 4 with pin 5, in both T568A and T568B — so a box built to the printed table would send two of its four channels down split pairs, with one leg from the green pair and one from the blue. That is not a subtle wiring preference; it would destroy common-mode rejection on those channels and make them a magnet for crosstalk from the other two.

Two independently hosted copies of the manual print the same sequential table, so it is a durable documentation error rather than a bad scan. The mapping given in the card above — channel 2 on the green pair, channel 3 on the blue — is the only reading consistent with the manual's own sentence about twisted pairs, and it is what community measurements of the hardware report. It also turns out to be exactly AES72 Type 1M, which a sequential table is not.

What happens when you mix them

Read the row as the box sending and the column as the box receiving. Because every scheme respects the twisted pairs, no combination produces noise, distortion or level loss — the failures are purely which channel and which way up.

sending ↓   receiving →Radial CatapultSoundTools CAT BoxWhirlwind Catdusathe sssnake Cat SnakeNeutrik NA-4I4O-AES72
Radial Catapult—Order and polarity wrong1→42→1⁻3→34→2Order wrong, polarity correct1→42→33→24→1Order and polarity wrong1→4⁻2→33→2⁻4→1⁻Order and polarity wrong1→4⁻2→23→3⁻4→1⁻
SoundTools CAT BoxOrder and polarity wrong1→2⁻2→43→34→1—Order and polarity wrong1→3⁻2→13→24→4Order and polarity wrong1→3⁻2→1⁻3→2⁻4→4⁻Order and polarity wrong1→2⁻2→1⁻3→3⁻4→4⁻
Whirlwind CatdusaOrder wrong, polarity correct1→42→33→24→1Order and polarity wrong1→22→33→1⁻4→4—Order correct, polarity wrong1→1⁻2→2⁻3→34→4⁻Order and polarity wrong1→1⁻2→3⁻3→24→4⁻
the sssnake Cat SnakeOrder and polarity wrong1→4⁻2→3⁻3→24→1⁻Order and polarity wrong1→2⁻2→3⁻3→1⁻4→4⁻Order correct, polarity wrong1→1⁻2→2⁻3→34→4⁻—Order wrong, polarity correct1→12→33→24→4
Neutrik NA-4I4O-AES72Order and polarity wrong1→4⁻2→23→3⁻4→1⁻Order and polarity wrong1→2⁻2→1⁻3→3⁻4→4⁻Order and polarity wrong1→1⁻2→33→2⁻4→4⁻Order wrong, polarity correct1→12→33→24→4—

1→4 means the sender's channel 1 arrives on the receiver's channel 4. A trailing ⁻ means that channel also arrives polarity-inverted. The matrix is symmetric in effect but not in labelling, so both halves are shown.

Three things worth knowing from that table

  • Radial and Whirlwind are electrically the same box with the channel numbers written backwards. Every pair matches on polarity; only the labels are reversed, so channel 1 arrives on 4 and 2 arrives on 3. Repatch and you are done — this is the most benign mismatch on the page.
  • The sssnake into a Catdusa is the dangerous one. All four channels arrive in the right place with the right names, and three of the four are upside down. There is nothing to notice.
  • The sssnake and Neutrik differ by one transposition. Type 1E and Type 4E are the same wiring with channels 2 and 3 swapped and identical polarity, so that pairing is annoying and completely harmless.

The failure that reaches site

A swapped channel announces itself. Put a signal into 1, watch it appear on 3, and you have found the problem in the first ten seconds of line check.

A polarity inversion does not. A single inverted channel sounds completely normal in isolation — same level, same tone, same everything. It only appears when it is summed against a correctly-polarised version of related material: a snare top against a inverted snare bottom, a DI against an inverted mic on the same cabinet, a stereo pair with one leg flipped collapsing to nothing in mono. It survives a bench test, it survives soundcheck on solo, and it turns up in the mix in front of an audience.

That is the real argument for knowing which map is in a box before it leaves the shop, and the reason the matrix above flags order-correct/polarity-wrong as worse than an obvious scramble.

Identify a box from four meter readings

The procedure in the box above, with the lookup done for you. Probe one pin of each twisted pair — 1, 3, 4 and 7 will do, since the other leg of each pair is whatever is left — and record which channel it reaches and whether it lands on XLR pin 2 or pin 3.

This needs JavaScript. The type table and the per-product cards above hold the same information.

A reading that matches nothing is worth having: it means either a split pair — check that the channel on pin 3 also appears on pin 6, not pin 4 — or a scheme that is not catalogued here. Both are more interesting than a match.

Sources, and how much each is worth

Not every row above is equally well attested, and flattening them into one table would imply otherwise. Radial, SoundTools and Neutrik all publish their own assignments and two of the three name their AES72 type outright. Whirlwind publish a table that contradicts their own prose. Thomann publish nothing at all.

Assembled 17 August 2026 with AI assistance, from the manufacturer documents linked above. The pin assignments were cross-checked between independent sources and are enforced at build time against the twisted-pair invariant, but nothing on this page has been verified with a meter on hardware. Before you cut a harness to any of it, measure the box in front of you.