The same plug is not the same signal

Stage 1 of 7 · Foundations6 of 6 in this stage9 min read

One assumption costs this industry more time and money than any other: it fits, so it works. A Mini DisplayPort socket might be DisplayPort or Thunderbolt. A USB-C socket might be a charging port or a 40 Gbit/s link. A 3-pin XLR might be a microphone or a lighting universe. None of that is visible from the outside.

The fix

Three layers that are routinely treated as one

Almost every connector argument dissolves the moment these are separated. They are independent, and any combination is possible.

The connector is a mechanical shape. It tells you what will physically mate and nothing whatsoever about what is behind it. The pinout is what each contact carries — and the same shape frequently has several different pinouts in service at once. The protocol is what the signal on those pins actually means.

So the useful question is never "does this fit". It is "what is on the pins, and does the thing at the other end speak it" — which is answerable, unlike the question people usually ask.

The interesting case

One connector, several protocols — USB-C

USB-C is 24 pins in a reversible shell, and it is the clearest example on earth of the pinout being negotiated rather than fixed.

A handful of pins never change: VBUS and GND for power, D+/D− for legacy USB 2.0, and CC1/CC2 — the configuration channel, which is where orientation is detected, power delivery is negotiated, and any alternate mode is agreed. Everything else is up for discussion.

The four high-speed differential pairs are the ones that change meaning. In plain USB they carry SuperSpeed data. In DisplayPort Alt Mode some or all of them are re-purposed as DisplayPort lanes — two lanes of DP plus USB data, or four lanes of DP and no USB at all. In Thunderbolt 3/4 and USB4 they carry a tunnelled mixture: PCIe, DisplayPort and USB multiplexed over the same wires.

That is why two identical-looking sockets on two laptops can be a 480 Mbit/s charge-only port and a 40 Gbit/s Thunderbolt port. And why the cable matters as much as the ports: a charging cable may have no high-speed pairs connected at all, and a passive cable rated for 20 Gbit/s will not do 40 over any length.

what the port is doing

power high-speed data negotiation & sideband unused in this mode

The other interesting case

One connector, two eras — Mini DisplayPort

Thunderbolt 1 and 2 used the Mini DisplayPort connector. Not a similar one — the same one. A DisplayPort monitor plugs straight into a Thunderbolt 2 port and works, because a Thunderbolt port also speaks DisplayPort.

The reverse does not hold. A Thunderbolt device plugged into a plain Mini DisplayPort socket does absolutely nothing, because there is no PCIe behind that socket. The plug seats perfectly, the connection is mechanically ideal, and nothing happens.

This asymmetry is the pattern to hold on to, because it recurs everywhere. The more capable port usually accepts the simpler signal; the simpler port almost never accepts the more capable one. And nothing about the outside of either tells you which one you are holding.

rule

When two things share a connector, compatibility is nearly always one-way. Ask which end is the more capable one, and expect the traffic to only flow down that hill.

The ones on a show

What is actually on the pins

Connector	What is on it	The thing that catches people
XLR5 (DMX)	Pin 1 shield, 2/3 the first data pair, 4/5 an optional second data link.	ANSI E1.11 specifies five pins. The industry uses three because it is cheaper, which is non-compliant and — much worse — makes a lighting universe and a microphone line physically interchangeable at 2 a.m.
XLR3 (audio)	Pin 1 shield, pin 2 hot, pin 3 cold. A balanced pair.	Microphone cable is not 120 Ω. Running DMX down it works until it does not: the terminator no longer matches anything and the reflections come back. See termination.
etherCON	An ordinary RJ45 Ethernet pinout inside a ruggedised locking shell.	The pinout is standard, so a normal patch lead plugs into an etherCON socket. What you lose is the strain relief and the lock — which is the entire reason the connector exists.
BNC	One coaxial signal path plus screen. Available in 50 Ω and 75 Ω.	The two impedances mate happily and look identical. At SDI frequencies the mismatch reflects; at low frequencies you will never notice. Video is 75 Ω, most RF and antenna work is 50 Ω.
powerCON	Mains line, neutral, earth, in a locking connector.	Standard powerCON is not rated for connecting or disconnecting under load. TRUE1 is. This is a safety difference, not a convenience one, and the blue/grey pair look reassuringly similar to the thing that is rated.
speakON NL4	Two loudspeaker pairs: 1+/1− and 2+/2−.	An NL2 plug mates with an NL4 socket and only picks up pair one. If a bi-amped box is quiet on the low end, check the plug before the amp.
USB-C	24 pins: power, CC negotiation, USB 2.0, four high-speed pairs, two sideband.	Everything about the high-speed pairs is negotiated. Ports and cables both vary enormously, and nothing on the outside distinguishes them.
HDMI / DisplayPort	HDMI carries TMDS; DisplayPort carries packetised micro-packets plus an AUX channel.	They are different signalling schemes. A passive DP-to-HDMI adapter only works because some DP sources can also output HDMI signalling (DP++). A source without that needs an active converter, and a passive adapter on it fails silently.

Myths

Things everybody has been told that are not true

myth“If it fits, the pinout must match.”

actuallyMating is mechanical. USB-C, Mini DisplayPort, BNC and XLR3 each carry several completely different things on the same shape. The plug is not a promise.

myth“A cable is a cable — all USB-C leads are the same.”

actuallyA charge-only lead may have no high-speed pairs in it at all, and a passive lead rated at 20 Gbit/s will not do 40. Current capability is declared by an e-marker chip in the cable, which is why one lead charges a laptop and an identical-looking one does not.

myth“3-pin DMX is fine, everyone does it.”

actuallyEveryone does do it, and it is still outside the standard, still loses the second data link, and still makes a mic line and a data line indistinguishable in a dark truck. Label everything, and never buy 3-pin for a rig that also has audio.

myth“Gold contacts sound better.”

actuallyGold resists corrosion, which matters for a connector that is mated once and left in a damp venue for ten years. On a connector cycled nightly it is a durability choice, not an audio one.

myth“Cat6 will fix my network problem.”

actuallyCategory is a bandwidth-over-distance rating. If the fault is a switch flooding multicast, a duplex mismatch or a bad crimp, a higher category changes nothing. See show networks.

Adapters hide the question, they do not answer it. A USB-C to HDMI dongle contains an active converter or it does not, and the failure looks identical to a bad cable.
Test the cable you are going to use, on the port you are going to use. Capability lives in the specific combination of source, cable and sink — not in any one of them.
Label at both ends, by what is on the pins. "DMX" and "MIC" on identical XLR3s is worth more than any amount of care later.
Buy the locking version for anything that carries a show. An etherCON or a TRUE1 costs a few pounds more than the moment it stops a disconnection mid-performance.

The habit worth forming

Ask what is on the pins

Every connector question on a show reduces to the same three checks, in the same order.

Will it mate? Mechanical, obvious, and the least informative of the three. What is on the pins at each end? Voltage, impedance, which contacts, which direction. Do the two ends speak the same protocol? And if a negotiation is involved — USB-C, DisplayPort, PoE — has it actually succeeded, or is it silently falling back?

That is the whole discipline. It also happens to be the discipline the rest of this site is built on: know what you are carrying before you worry about what it is carrying it in.

Got a pinout that regularly catches people out?

Connector conventions vary by market and by decade, and the ones that cause trouble are exactly the ones nobody writes down. Open an issue and it goes on the page with a citation.

Check yourself

1 question on the mechanism rather than the trivia. Nothing is scored and you can answer again — the explanation under a wrong answer is the point of it.

1A Thunderbolt 2 device will not work on a Mini DisplayPort output, even though the connector fits. Why?

Answers are kept in this browser on this device. Nothing is sent anywhere.

You can now answer

Why does this cable not work?
What is on a USB-C pin?
Is 3-pin DMX fine?

Kept in this browser on this device. Nothing is sent anywhere.

Where to go next

These pages are a chain rather than a library — the next one along usually answers the question this one raised.

the practice Experience architecture What all of this is actually for. Analogue Before there was a computer in it What does a capacitor do? Sensing Turning the world into a voltage What does phantom power do? Digital Numbers that stand for signals What does 24-bit buy me? Colour & video How a colour becomes a number What is a hex code?

Everything on this page, in the index

Each of these has its own entry with sources, ports, gotchas and an open gap where a field is missing.

standard ANSI E1.11-2024 Entertainment Technology - USITT DMX512-A Asynchronous Serial Digital Data Transmission Standard…