🔌Computer Ports & Connectors

The flat rectangular one on every computer since the late nineties.

The classic USB port. It carries data and power for keyboards, mice, flash drives, printers, phone cables, webcams, headsets and just about everything else that plugs into a computer, and it has done so for nearly thirty years. Every desktop has a row of them and most laptops still have at least one.

USB 1.0 arrived in 1996 to replace the tangle of serial, parallel and PS/2 ports with one connector. USB 2.0 (2000) took it to 480 Mbit/s and made it universal; USB 3.0 (2008) added five extra contacts inside the same shape for 5 Gbit/s, and USB 3.1 (2013) doubled that again. The same shape is still on the back of every desktop.

What the colours mean. Black or white plastic inside means USB 2.0. Blue is the convention for USB 3.0 (5 Gbit/s) and teal for USB 3.1 (10 Gbit/s) — but it is only a convention. Plenty of makers use black for everything, and a red or yellow port usually means it stays powered for charging when the machine is asleep.

The squarish one on the back of printers and scanners.

The device end of a printer or scanner cable. It also turns up on external hard drives, audio interfaces and some monitors’ USB hubs — anything big enough not to care about a chunky port. The computer end of the same cable is a Type-A.

Part of the original 1996 USB standard: the idea was that a cable should have a different plug at each end so you could never plug a computer into a computer. USB 3.0 (2008) had to add contacts, and rather than squeeze them in it put a second storey on the plug.

What the colours mean. White or black. USB 3.0 versions are usually blue inside, same as Type-A.

The little trapezoid on old cameras, sat-navs and PlayStation 3 controllers.

The small USB before micro-USB existed. It charged and synced digital cameras, older GoPros, TomToms, PS3 controllers, MP3 players and some external drives, and it is still the port on a lot of Arduino boards and cheap gadgets because it is sturdier than micro-USB.

Added to the USB 2.0 standard in 2000 for cameras and MP3 players. Officially deprecated in 2007 when micro-USB arrived, but it never really left — PS3 controllers used it until 2013 and hobby electronics kept using it.

The thin wide one that charged every Android phone before USB-C.

The charger and data port on Android phones for a decade, and on Kindles, Bluetooth speakers, power banks, game controllers, headphones and most cheap electronics. Still very common on anything made before about 2019 and on budget gadgets today.

Announced in January 2007 to replace mini-USB with something thinner and tougher. In 2009 the phone makers agreed with the EU to standardise on it, which ended the era of a different charger for every phone. USB-C has been replacing it since 2015.

A micro-USB with a second plug bolted on the side.

The port on most portable external hard drives made between about 2010 and 2018 — the Seagate and WD Passport drives in every drawer. Samsung put it on the Galaxy S5 and Note 3 as a phone charger, and was roundly ignored.

Defined in the USB 3.0 standard in 2008. A normal micro-USB plug fits in the left half and works at USB 2 speed, which was the whole point: keep the old cables working, add speed for those who bought the new one.

Small, oval, reversible — and now the law in a lot of places.

One port for everything: charging up to 240 W, data, video to a monitor, docking a laptop, and (when the port supports it) Thunderbolt. It is on every current phone, tablet and laptop and most accessories. The catch is that the shape tells you nothing about what a given port can actually do — a USB-C port might be USB 2 speed and charge-only, or a full Thunderbolt port.

Finalised in August 2014, first seen on the 2015 MacBook and a Chromebook the same year. The EU made it mandatory for phones and most portable gadgets from the end of 2024, which is why the iPhone 15 (2023) finally dropped Lightning. Australia doesn’t have that law, but the phones are the same phones.

What the colours mean. No colour convention at all, unlike Type-A. The symbols printed beside the port are the only clue to what it can do.

Looks like USB-C. Is USB-C. Just much faster, with a bolt beside it.

The fast lane. Thunderbolt 3 and 4 move 40 Gbit/s — enough for two 4K monitors, an external graphics card and fast storage through one cable — and Thunderbolt 5 doubles that. It is what makes a one-cable docking station work: power, two screens, network and every USB device on the desk through a single plug. USB4 is the open standard built on the same technology.

Thunderbolt 3 (2015) moved Intel’s interface from the Mini DisplayPort shape onto USB-C. Thunderbolt 4 (2020) tightened the minimum requirements so every port actually delivers, Thunderbolt 5 (2023) reached 80 Gbit/s, and USB4 (2019) opened the technology to everyone. Every Mac since 2016 and most premium Windows laptops have it.

The Mini DisplayPort shape, on Macs from 2011 to 2016.

Monitors, fast external storage, Ethernet adapters and docks on Macs of that era, at 10 or 20 Gbit/s. A Mini DisplayPort monitor cable plugs straight in and works, which is how most people used it.

Thunderbolt 1 (2011, 10 Gbit/s) was Intel and Apple’s answer to FireWire’s decline; Thunderbolt 2 (2013) doubled the speed with the same connector. Thunderbolt 3 moved to USB-C in 2015, and Apple’s adapter between the two is still the way to run old Thunderbolt drives.

Video and sound in one cable. On every TV, monitor and console made this century.

Picture and sound together, digitally, down one cable — which is why it replaced the piles of leads behind the telly. TVs, monitors, projectors, game consoles, set-top boxes, laptops, graphics cards and soundbars all have it, and the same plug has carried everything from standard definition to 8K.

HDMI 1.0 was released in December 2002 by a group of TV makers, built on the DVI digital signal with audio added. HDMI 1.4 (2009) brought 4K, 2.0 (2013) brought 4K at 60 Hz, and 2.1 (2017) brought 8K, 120 Hz gaming and variable refresh rate. The connector has never changed.

HDMI shrunk for cameras and some tablets.

The HDMI output on DSLR and mirrorless cameras, camcorders, some older tablets and a few graphics cards. Same 19 signals as full HDMI, so a simple adapter or a mini-to-full cable is all it takes to get the camera onto a TV or a capture device.

Added in HDMI 1.3 (2006) for portable devices. Micro HDMI arrived three years later and took most of its market.

About the size of a micro-USB. Raspberry Pi 4, GoPros, action cameras.

The video output on action cameras, some phones and tablets of the early 2010s, and — the reason most people meet it now — the Raspberry Pi 4 and 5, which have two. Same signal as full HDMI; the cable or adapter does the size change.

Introduced with HDMI 1.4 in 2009, aimed at phones. Phones went with MHL and then USB-C instead; cameras and single-board computers kept it.

The computer industry’s monitor connector. One cut corner, a latch, and more bandwidth than HDMI.

The port between a PC and its monitor, and the one to use for high refresh-rate gaming screens — it has more bandwidth than HDMI and it can daisy-chain monitors. On graphics cards, docks and PC monitors. It does not appear on TVs, which is the quickest way to know which cable you need.

Created by VESA, the monitor standards body, and approved in 2006 as a royalty-free alternative to HDMI for computers. DisplayPort 1.2 (2010) added daisy-chaining, 1.4 (2016) 8K, and 2.0/2.1 (2019–22) tripled the bandwidth. USB-C can carry it natively, which is how a laptop drives a screen through a dock.

Apple’s tidy little version, later borrowed by Microsoft Surface and every ultrabook.

The monitor output on MacBooks from 2008 to 2015, Surface Pro up to 2017, ThinkPads and a lot of thin laptops, and on many graphics cards to save bracket space. It carries exactly the same signal as full-size DisplayPort, so adapters to DisplayPort, HDMI, DVI and VGA are simple and cheap.

Apple designed it in 2008 and then handed it to VESA, who made it part of the DisplayPort standard in 2009. From 2011 the same port doubled as Thunderbolt 1 and 2 on Macs.

The big white plug with a grid of pins. Digital video before HDMI, no sound.

The monitor connection of the 2000s: a clean digital picture to the first flat-panel screens, before HDMI and DisplayPort took over. It carries the same digital signal as HDMI minus the audio, which is why a passive DVI-to-HDMI cable just works. Still on older monitors, graphics cards and KVM switches.

Published in 1999 by the Digital Display Working Group (Intel, Compaq, IBM, HP and others). It ruled PC monitors for a decade; HDMI and DisplayPort took over, and Intel and AMD stopped putting it on their reference hardware from 2015.

What the colours mean. Nearly always white plastic, which is the quickest way to tell it from a VGA port at a glance.

Apple’s small rectangular DVI, 2003–2008. Needs an adapter for anything.

The monitor output on the 12-inch PowerBook G4, the plastic MacBook, the Intel iMac and the Mac mini of the mid-2000s. It carried DVI (and on some models VGA and TV-out), but nothing plugs into it directly — you always needed one of Apple’s adapters.

Appeared in 2003 on the 12-inch PowerBook because a full DVI port wouldn’t fit. Replaced by Mini DisplayPort in 2008.

Even smaller, on exactly one Mac: the first MacBook Air.

The monitor output of the original 2008 MacBook Air, and nothing else. Two adapters existed — to DVI and to VGA — and that was the whole ecosystem.

January 2008 to October 2008, one model. Mini DisplayPort replaced it on the very next Air. The shortest-lived video port on this page.

The blue trapezoid with 15 pins in three rows and two thumbscrews. Analogue, and still everywhere.

The analogue monitor connection that every PC had for twenty-five years, and the one projectors, older monitors, office docks, tills, medical and industrial gear still use. Picture only — no sound, no digital — and the picture gets softer the longer and cheaper the cable.

Introduced by IBM in 1987 with the PS/2 range as the Video Graphics Array standard, and the connector has outlived the standard by decades. Microsoft and the PC makers announced its retirement in 2010; it is still being fitted to new projectors.

What the colours mean. Blue is the PC colour code for VGA (the same 1999 scheme that made PS/2 ports green and purple). Older gear used plain black or grey.

Apple’s small VGA on iBooks and eMacs, 2001–2005.

The monitor and projector output on iBooks, the eMac and the early iMac G5, plus a few Sony VAIO laptops. It carried an analogue VGA signal (and TV-out on later models) through an Apple adapter.

First on the 2001 iBook. Replaced by Mini-DVI from 2005 as the digital monitors arrived.

A round 4-pin plug for a better-than-composite analogue picture. VCRs, DVD players, old laptops.

A sharper analogue picture than the yellow RCA plug, by carrying brightness and colour on separate wires, without needing three cables like component. On S-VHS and DVD players, older TVs, camcorders and the TV-out of laptops and graphics cards up to the mid-2000s.

Introduced by JVC with S-VHS in 1987. PC graphics cards adopted it for TV-out in the late nineties, often as a 7-pin version that also carried composite. HDMI killed it in the 2010s.

The huge angled plug on European TVs and video gear. Rare in Australia.

One cable for everything between a European VCR, console or set-top box and the TV: composite video, RGB, stereo audio and control signals in a single plug. Australian gear mostly skipped it, so you meet it on imported equipment and PAL-region retro consoles, which retro gamers prize for its clean RGB picture.

Created in France in 1977 (as Péritel) and made compulsory on French TVs from 1980; the rest of Europe followed. HDMI replaced it through the 2000s and new TVs dropped it around 2015.

The clear clip plug for a wired network. Eight gold contacts, 11.7 mm wide.

A wired network connection — computers, modems, TVs, printers and game consoles to the router, and the router to the NBN box. Still faster and steadier than Wi-Fi, which is why the gaming PC and the TV should be on it. The socket on a computer has two little lights that flicker when it’s working.

The plug is really an "8P8C" modular connector from the Bell telephone system of the 1970s; the name RJ45 stuck from a similar phone standard and everybody gave up correcting it. Ethernet adopted it with 10BASE-T in 1990 and every speed since has kept the same plug, from 10 Mbit/s to 10 Gbit/s.

What the colours mean. The cable sheath colour means nothing — it’s just what the installer had. The category printed on the cable does: Cat 5e is fine for 1 Gbit/s, Cat 6 or 6a for 10 Gbit/s and longer runs.

The narrow clip plug for a phone line. Looks like Ethernet, isn’t.

The phone line: it connects a phone, fax or modem to the wall, and on an NBN box it’s the port the home phone plugs into. ADSL modems used it for the line side, which is why the lead that came with the modem still gets confused with the network cable.

Part of the same 1970s Bell "registered jack" family as RJ45. Australia used its own chunky three-pin 610 phone plug until the 1990s; RJ11 took over as modems and imported phones arrived, and the 610 socket is now something you find in older houses and adapt.

The RJ11 shape with all six contacts fitted. Business phones, tills, robots.

Three phone lines in office phone systems, and these days the drawer-kick lead between a receipt printer and a cash drawer at the till, some alarm panels, and the sensor cables on LEGO Mindstorms robots.

Same 1970s family as RJ11. "RJ12" and "RJ25" were two wiring standards for the same 6P6C plug, and the names are used interchangeably.

The tiny clip plug on the curly cord between a desk phone and its handset.

It joins the handset to the base of a desk phone, and that’s about all it does — it is not a phone-line plug and it won’t fit one.

Same 1970s Bell family. Called RJ9, RJ10 or RJ22 depending on who you ask; none of those names were ever official for this plug, and the correct "4P4C" never caught on.

The headphone plug. Count the black bands on the tip — that tells you what it carries.

Headphones, earphones, speakers, microphones and aux leads — the analogue audio plug on everything from a Walkman to a gaming headset. Phones have mostly dropped it; laptops, desktops, car stereos and hi-fi still have it, and on a desktop PC the coloured rings around the sockets say which is which.

A shrunken version of the 6.35 mm telephone switchboard plug of 1878. The 3.5 mm size arrived with pocket transistor radios in the 1950s and the Sony Walkman made it universal in 1979. Phones began dropping it in 2016.

What the colours mean. On a desktop PC the rings follow the 1999 PC colour code: lime green is the main speaker/headphone output, pink is microphone, light blue is line in. Black, orange and grey are the rear, centre/subwoofer and side channels for surround sound. Laptops and phones use one combined headset socket with no colour.

The headphone plug’s little sibling, on old phones and two-way radios.

The headset socket on mobile phones before 2005 or so, and still on cordless home phones, two-way radios, some cameras’ remote shutter releases and older Bose headphones.

Miniaturised again from the 3.5 mm plug for 1990s mobile phones. Phones moved to 3.5 mm and then to nothing.

The big one: guitars, amps, mixing desks and studio headphones.

Electric guitars and basses, keyboards, amplifiers, mixing desks, powered speakers and studio headphones. Any hi-fi amplifier has one on the front for headphones, and big headphones ship with a screw-on adapter to fit it.

The oldest connector on this page: invented in 1878 for telephone switchboard operators to patch calls, and essentially unchanged since. It moved to audio because every telephone company already made millions of them.

A square port with a red glow inside. Sends sound as light to a soundbar or amplifier.

Digital sound from a TV, console or player to a soundbar or AV receiver, sent as light down a plastic fibre so it can’t pick up hum or interference. It carries stereo and compressed surround (Dolby Digital, DTS) but not the lossless formats from Blu-ray or the newer Atmos streams — those need HDMI eARC.

Toshiba created it in 1983 to link CD players to amplifiers (the name is Toshiba Link). It carries the S/PDIF signal that also travels down the orange RCA coax socket. HDMI ARC has mostly replaced it on new TVs, but soundbars still ship with the cable.

One orange RCA socket carrying the same digital signal as optical.

The same digital audio as the optical port, carried electrically instead of by light — same formats, one ordinary cable. On DVD and Blu-ray players, AV receivers, some TVs, desktop sound cards and a few soundbars.

S/PDIF (Sony/Philips Digital Interface) dates from the CD era of the mid-1980s, defined with both a coax and an optical connector. The coax one borrowed the RCA plug because everyone already had them.

The chunky three-pin round plug with a latch. Microphones, PA systems, stage lighting.

Professional microphones, mixing desks, powered speakers and PA systems. It carries balanced audio, which rejects hum over long cable runs, and the five-pin version runs DMX stage lighting. Unusually, the sockets on equipment can be either gender: inputs are female, outputs are male.

Cannon Electric’s "X" series from the 1950s, with a Latch added and then a Rubber boot — hence XLR. It has been the professional audio standard ever since, which is why a fifty-year-old microphone plugs straight into a new mixer.

The push-on round plugs in every colour. The colour is the whole story.

Analogue video and audio between TVs, VCRs, DVD players, consoles, turntables and hi-fi. Every piece of home entertainment gear made before HDMI has a row of these, and turntables and amplifiers still use the red-and-white pair. The colour says what travels down the wire: the plugs themselves are all identical.

Radio Corporation of America put it on its radios in the late 1930s so a record player could plug in, and it has never been replaced for what it does. The colour code came later, standardised for home cinema gear in the 1990s along with the five-channel colours for surround sound.

What the colours mean. Composite: yellow video, white left, red right. Component video: green (Y), blue (Pb), red (Pr) — three cables for one picture, audio separate. Digital coax audio: orange. Multichannel analogue audio on receivers: white and red front, green centre, blue and grey surround, brown or purple subwoofer. The plugs are all electrically identical; the colour is a label.

The twist-and-lock coax plug on security cameras, test gear and broadcast video.

The video connection on analogue and HD-over-coax CCTV cameras and recorders, on oscilloscopes and test gear, and on professional broadcast video (SDI). Old office networks ran on it too. Network (IP) cameras use RJ45 instead and get power down the same cable.

Designed around 1951 by Paul Neill and Carl Concelman (Bayonet Neill–Concelman). Thin-coax Ethernet used it in the 1980s and early 90s, and it never left the CCTV and broadcast worlds.

The screw-on coax plug on Foxtel, satellite dishes and cable modems.

Pay TV and satellite: Foxtel boxes, satellite dishes and LNBs, cable modems, and the antenna input of some TVs and amplifiers. It carries the RF signal from the dish or the street to the box, and it is the connector the whole world’s satellite gear standardised on.

Designed in the 1950s by Eric Winston at Jerrold Electronics for the first cable TV systems.

The smooth push-in plug for a TV aerial. On the wall plate and the back of every TV in Australia.

The TV aerial. It joins a TV, set-top box or recorder to the antenna wall point, and it is the plug on both ends of the fly lead behind every telly in Australia, the UK and Europe. The Americas use the F-type for the same job.

Belling & Lee designed it in England in 1922 for the first radio sets, which makes it the oldest connector still in daily use on this page after the ¼-inch jack.

A small gold screw-on antenna connector. GPS, 4G routers, radio modules.

External antennas on 4G/5G routers and modems, GPS receivers, ham and LoRa radio gear and test equipment — anywhere a small, reliable radio connection matters. Good to 18 GHz, which is why it’s the default on radio hardware.

Developed in the early 1960s as the SubMiniature version A connector for microwave work.

SMA with the pin and hole swapped. The screw-on antennas on a Wi-Fi router.

The removable antennas on Wi-Fi routers, access points and PC Wi-Fi cards screw onto this. It exists purely to be incompatible with ordinary SMA antennas, and a 4G antenna will screw onto a Wi-Fi router and not connect.

The Wi-Fi industry adopted it in the 1990s because US radio rules required a non-standard antenna connector on consumer wireless gear, to stop people fitting big antennas. Everyone promptly started selling big RP-SMA antennas anyway.

The big weatherproof screw-on coax for outdoor antennas.

Rooftop 4G/5G and Wi-Fi antennas, cellular boosters and the thicker low-loss cable that runs to them, with a pigtail down to SMA at the equipment end. It has a rubber seal inside so it survives outdoors.

Paul Neill designed it at Bell Labs in the 1940s, named for himself, as the first connector to carry microwave signals cleanly. Still the outdoor standard eighty years on.

Nine pins in a D. The port that ran mice and modems, still running tills, alarms and machinery.

Slow, simple serial data — which is exactly why it refuses to die. It was the mouse and modem port on every PC until USB, and it is still how point-of-sale printers and scanners, UPS units, alarm panels, CNC machines, network switches’ console ports and lab instruments talk. Modern PCs need a USB-to-serial adapter.

The D-sub family dates from 1952 (ITT Cannon). IBM put the 9-pin version on the PC/AT in 1984 as a smaller RS-232 port than the 25-pin, and it stayed on every PC until USB took over around 2000.

The wide 25-hole port that printers used before USB.

Printers, scanners, Zip drives and the odd software dongle, sending eight bits at once down a fat cable. It was the printer port on every PC for twenty years. The other end of a printer cable was usually the different-looking Centronics plug.

Centronics invented the parallel printer interface in 1970; IBM shrank the PC end onto a 25-pin D-sub for the first IBM PC in 1981. Standardised as IEEE 1284 in 1994 with faster modes, then killed by USB within a few years.

The original PC serial port — same size as the printer port, but with pins.

The full RS-232 serial interface with all its handshaking lines: the modem and terminal port on early PCs before the 9-pin took over. Old modems, plotters and some industrial equipment still have it on the device end.

RS-232 was standardised in 1960 for teletypes, and the 25-pin D-sub became its standard plug. The first IBM PC (1981) had one; the AT (1984) shrank it to nine pins.

The printer end of the old parallel cable, with the wire clips that snap over it.

The printer end of a parallel cable — the PC end is the 25-pin D-sub. Any laser or inkjet printer from before about 2005 has one, and industrial label printers still do.

Centronics Data Computer Corporation built it into their printers in 1970, borrowing the connector from Amphenol. The whole industry copied the printer, then the plug.

Fifteen holes in two rows, usually yellow, on old sound cards. Joysticks and MIDI keyboards.

Analogue joysticks and, through a Y-cable, MIDI music keyboards. It was on every sound card and many motherboards from the late 80s until about 2005, and it is the reason old flight-sim joysticks have that plug.

IBM’s Game Control Adapter for the 1981 PC defined it; Creative’s Sound Blaster (1989) put it on the sound card and added MIDI. Windows Vista dropped support in 2007, and USB joysticks had already won.

What the colours mean. Yellow, in the 1999 PC colour code. Older ones are plain.

The round 6-pin keyboard and mouse ports — purple for keyboard, green for mouse.

The keyboard and mouse ports on every PC for twenty years, and still on some motherboards because PS/2 keyboards never miss a key and work before the operating system loads — gamers and server admins like that. Two sat side by side: purple keyboard, green mouse; newer boards keep one split-colour port that takes either.

Arrived with IBM’s PS/2 computers in 1987 to replace the big 5-pin DIN keyboard plug and the serial mouse. The purple-and-green colours come from the PC 99 design guide of 1999. USB replaced it.

What the colours mean. Purple keyboard, green mouse, and a split green/purple port accepts either. Older machines have plain ports labelled with a tiny picture instead.

The big round 5-pin plug. Keyboards on PCs before 1990, and every MIDI cable since 1983.

Two very different jobs share it: the keyboard port on original IBM PCs and ATs (and their clones into the early 90s), and MIDI, the connection between music keyboards, synths, drum machines and interfaces. Same plug, unrelated signals — a MIDI cable is a perfectly good AT keyboard cable and vice versa.

A German industrial standard (DIN) from the 1960s for hi-fi and tape recorders. IBM used it for the PC keyboard in 1981; the MIDI standard of 1983 picked it because it was cheap and everywhere. PCs moved to PS/2 and USB; music gear still uses it, though USB and the 3.5 mm "MIDI TRS" jack are taking over.

A tall plug with a rounded top. Camcorders, external drives and audio interfaces, 1995–2012.

Fast data with power, before USB could manage either: external hard drives, DV camcorder capture, scanners and pro audio interfaces, especially on Macs. It is the reason people ring us about old mini-DV tapes — capturing them needs a FireWire port, an adapter, or us.

Apple developed it through the early 90s; it became the IEEE 1394 standard in 1995. Sony called it i.LINK, and every digital camcorder of the DV era used it. USB 2.0 undercut it on price, Thunderbolt replaced it on Macs from 2011, and Apple dropped it in 2012.

The small 4-pin FireWire on camcorders and Windows laptops.

The FireWire port on nearly every mini-DV camcorder and on Windows laptops of the 2000s — data only, no power. A 4-to-6-pin cable connects a camcorder to a desktop for capturing tapes.

Sony’s i.LINK version of IEEE 1394, from the late 1990s, sized for camcorders. Same signal as the 6-pin.

FireWire’s faster sequel with a bevelled top instead of a rounded one.

External drives, audio interfaces and video gear at 800 Mbit/s — the drive port on Macs from 2003 to 2012. An adapter cable connects the older 6-pin and 4-pin devices at their slower speed.

IEEE 1394b, 2002. Apple used it heavily on the MacBook Pro, iMac and Mac mini. Thunderbolt replaced it.

A flat slot for external hard drives, before USB 3 made it pointless.

External hard drives at full internal-drive speed, which USB 2 couldn’t manage. On drive docks and the back of desktops and laptops from about 2005 to 2012. It carries no power, so the drive needed its own supply — the "eSATAp" version on laptops combined it with a USB port to fix that.

Added to the SATA standard in 2004. USB 3.0 (2008) matched its speed and carried power too, and eSATA was gone from new machines within a few years.

Apple’s small reversible plug, iPhone 5 to iPhone 14.

Charging and syncing every iPhone from 2012 to 2022, iPads until about 2018, AirPods cases, and Apple’s keyboards and mice. Replaced by USB-C, but there are a lot of them still out there, and the port filling with pocket fluff is the single most common "iPhone won’t charge" fix we do.

Introduced September 2012 with the iPhone 5 to replace the 30-pin dock connector. Apple licensed it under the MFi program, which is why unlicensed cables get an "accessory not supported" message. The iPhone 15 (2023) moved to USB-C after the EU’s common-charger law.

The wide flat plug on iPods and iPhones up to 2012.

Charging, syncing and docking for every iPod with a dock connector, the iPhone up to the 4S and the iPad up to the third generation — and the speaker docks, car kits and alarm clocks of that era. Lightning replaced it.

Introduced with the third-generation iPod in 2003, replacing FireWire on the iPod. Ten years and hundreds of millions of devices later, Lightning took over in 2012.

The original magnetic MacBook charger, 2006–2012. Five pins in a thick rounded plug.

The charger on MacBooks, MacBook Airs and MacBook Pros up to 2012, held on by magnets so a tripped-over cable pulls free instead of dragging the laptop off the table. The cable fraying at the plug end is the famous failure; the plug itself is fine.

Introduced with the first MacBook Pro in January 2006, borrowing the idea from Japanese deep fryers. Replaced by the thinner MagSafe 2 in 2012.

Same idea, thinner and wider, for the 2012–2017 MacBooks.

The charger on the MacBook Air and MacBook Pro from mid-2012 until the USB-C models replaced them (2016 for the Pro, 2017–2018 for the Air). Same magnets, same five pins, flattened to fit thinner laptops.

June 2012, with the first Retina MacBook Pro. Apple dropped it for USB-C charging in 2016 and brought the magnets back as MagSafe 3 in 2021.

The magnets return: 2021 onwards, with a USB-C cable on the other end.

Charging on current MacBook Pros and Airs. The cable is detachable and USB-C at the charger end, so any USB-C power adapter of enough wattage runs it, and the laptops can also charge through their USB-C ports.

Returned with the 14 and 16-inch MacBook Pro in October 2021 after five years of USB-C-only charging, mainly to allow faster charging than USB-C could then deliver.

The Mac’s keyboard and mouse port for twelve years. Same shape as S-Video.

Keyboard, mouse, graphics tablet and the odd dongle on every Mac from 1986 to 1998, daisy-chained — the keyboards usually had a second port on the side for the mouse.

Steve Wozniak designed it for the Apple IIGS in 1986. The 1998 iMac replaced it with USB and the rest of the Mac line followed by 1999. Griffin’s iMate adapter kept old ADB keyboards alive on USB for years.

The two round 8-pin ports on the back of old Macs, marked with a printer and a phone.

The printer and modem ports on every Mac from 1986 to 1998 — one each — which also ran LocalTalk, Apple’s early office network. USB-to-serial adapters with this plug still exist for old MIDI and printer gear.

Introduced on the Mac Plus in 1986 in place of the DE-9 serial ports, and dropped with the 1998 iMac.

A DVI plug with USB and power added, so an Apple monitor needed one cable. 2000–2004.

The DVI picture, a USB connection for the monitor’s hub and 25 volts to power the screen, all in one cable — so Apple’s Cinema Displays of that era have no power lead. Only on Power Mac G4s, the G4 Cube and early G5s, and their monitors.

Introduced with the Power Mac G4 Cube in July 2000. Apple went back to plain DVI in 2004, and the ADC-to-DVI adapter needed to run those monitors on anything else is now a collector’s item.

The round power plug, in a dozen sizes that all look the same from across the bench.

Low-voltage power from a plug pack or laptop charger into the device: routers, modems, monitors, LED lighting, CCTV cameras, speakers, and every laptop before USB-C charging. The laptop makers each picked their own size, and some (Dell, HP, Lenovo, Samsung, Sony) added a third pin down the middle that tells the laptop which charger it is.

Around since the transistor-radio era; the EIAJ standard of the 1980s tried to tie plug size to voltage and everybody ignored it. Laptops used them until USB-C charging took over from about 2018, and small electronics still do.

What the colours mean. HP’s 4.5 mm "blue tip" is the only colour that means anything. Otherwise black, and the size is stamped on the charger label if you’re lucky.

The yellow rectangular plug with a pointed end on ThinkPads and IdeaPads, 2012–2020.

The charger on Lenovo ThinkPads, IdeaPads and Yogas for most of a decade, in wattages from 45 W to 230 W through the same plug, until USB-C charging arrived.

Introduced in 2012 with the ThinkPad X1 Carbon to replace the round 7.9 mm plug. Most Lenovo laptops moved to USB-C from 2018 onwards.

Microsoft’s magnetic charging and docking port on Surface tablets and laptops.

Charging and docking on every Surface Pro, Laptop and Book since 2014. It also carries data to the Surface Dock. Newer models charge through USB-C as well, and Microsoft has been quietly shifting that way.

Arrived with the Surface Pro 3 in 2014, replacing the earlier five-pin magnetic connector.

The three-pin power lead on every desktop, monitor and printer. Not actually a kettle lead.

Mains power into a desktop PC, monitor, printer, amplifier, server or UPS — the one lead every piece of IT gear shares, so a cable from any of them works on all of them. Rated 10 A, which is fine for a gaming PC with a 1000 W supply and not fine for a rack of them on one power board.

Standardised in 1970 as IEC 60320, so one cable design could serve every country’s wall plug. The nickname "kettle lead" is wrong: real kettles use the C15/C16 version with a notch, rated for the heat. A kettle lead fits a PC; a PC lead does not fit a kettle.

The two-pin lead shaped like a figure 8. Consoles, TVs, radios, laptop power bricks.

Mains power for anything double-insulated that doesn’t need an earth: PlayStations, Xboxes, small TVs, radios, printers and a lot of laptop power bricks. Rated 2.5 A — fine for a console or a laptop brick, not for a heater.

Also part of the 1970 IEC 60320 standard, as the unearthed 2.5 A member of the family. The polarised variant came later for equipment that cared which pin was live.

The three-lobed "Mickey Mouse" plug on laptop power bricks and projectors.

The earthed mains lead on most laptop chargers, projectors and some monitors and printers — small and earthed, which is why the laptop makers chose it. It is completely standard: any cloverleaf lead with an Australian plug will do.

The earthed 2.5 A member of IEC 60320 from 1970. Some newer chargers use the figure-8 instead.

Two angled flat pins and a vertical earth. Australia, New Zealand, Papua New Guinea, Fiji, Argentina and (upside down) China.

The plug on the wall at home, 230 V and 10 A. Two-pin versions leave the earth off; the 15 A version has a wider earth pin and needs a matching socket, which is why the caravan lead doesn’t fit the house. China uses the same layout, mounted with the earth at the top, so the plugs interchange.

Adopted in Australia in 1937 and based on a 1916 American Hubbell design that the US itself never used.

Two parallel flat pins, plus a round earth on Type B. North America, Japan, Central America, Taiwan.

The wall plug of the Americas and Japan, at 120 V (100 V in Japan). That voltage matters: laptop and phone chargers are nearly always 100–240 V and only need a plug adapter, but a hair dryer or heater bought in the US will not survive an Australian socket.

Harvey Hubbell patented the two-blade plug in 1904; the earth pin was added in the 1920s and required in US homes from 1962. Japan uses the same shape at its own voltage.

Three big rectangular pins and a fuse inside. UK, Ireland, Singapore, Malaysia, Hong Kong, the Gulf.

The wall plug of Britain and the former British territories in Asia and the Middle East, at 230 V, so most UK gear works in Australia with an adapter. It carries its own replaceable fuse, and a UK appliance that has gone dead may just need a new 3 A or 13 A cartridge — in the plug, not the appliance.

Designed in 1947 for post-war Britain’s ring-main wiring, which is why the plug carries its own fuse.

Two round pins, no earth. Fits most of Europe, and most of the rest of the world’s round-pin sockets too.

The plug on phone chargers, shavers and small double-insulated appliances across Europe, and it fits the Type E, F, J, K and L sockets across Europe and much of Asia, Africa and South America. Europe is 230 V like Australia, so European chargers work here with a simple adapter.

Standardised in 1963 as a plug that would fit every European round-pin socket at once, and it does. Rated 2.5 A, which is why anything bigger uses the earthed Type E or F instead.

The round earthed European plug with metal clips on the sides. Germany, Spain, the Netherlands, Scandinavia, Russia, Korea.

The earthed wall plug of most of continental Europe, rated 16 A. Nearly every European appliance ships with the hybrid CEE 7/7 version, which also has a hole for the French earth pin, so one plug fits both the German and the French socket.

Germany’s "Schutzkontakt" (protective contact) plug of 1926, one of the first earthed domestic plugs anywhere.

Like Schuko, but the earth is a pin sticking out of the socket. France, Belgium, Poland, Czechia, Slovakia.

The earthed wall socket of France, Belgium, Poland and their neighbours. Modern European appliances carry the hybrid plug that fits this and the German Schuko socket, so in practice the two are interchangeable for anything made this century.

France standardised it in the 1950s, choosing a pin in the socket rather than Germany’s side clips. The hybrid plug of the 1960s ended the argument for appliance makers.

Three round pins in a straight line, earth in the middle. Italy and Chile.

Italy’s wall plug, in a 10 A size with thin pins and a 16 A size with fatter pins further apart; Italian sockets are often the "bipasso" kind that takes both, and new Italian homes usually have sockets that take Schuko as well. A Europlug fits the 10 A version.

Italy’s own standard from the 1930s, kept while the rest of Europe converged on Schuko.

Three very fat round pins in a triangle. South Africa, and high-power sockets in India.

South Africa’s everyday plug (slowly being joined by a new compact standard), rated 15 A, and the high-current plug for air conditioners in India and the UK’s old theatre lighting.

The British 15 A round-pin standard of 1934 (BS 546), which the UK replaced with Type G in 1947 and South Africa kept.

Three round pins in a triangle, smaller than the South African. India, Sri Lanka, Nepal, Namibia.

India’s standard plug for lamps, chargers and small appliances, rated 5 A; bigger loads use the Type M. Indian sockets are frequently the universal kind that also swallow a Europlug.

The 5 A member of the same 1934 British round-pin standard as Type M, retained in India after the UK moved on.

Two round pins with the earth pin offset below. Switzerland and Liechtenstein.

Switzerland’s wall plug. A Europlug fits; a Schuko doesn’t. The neat, small plug is the reason Swiss power boards fit so many outlets.

Switzerland’s own standard, current form from 1959. It looks close to Brazil’s Type N but the earth sits at a different offset, so they don’t interchange.

Two round pins and a half-round earth below — the socket looks like a smiling face. Denmark, Greenland.

Denmark’s wall plug. Europlugs fit; Schuko plugs fit but don’t earth, which is why Denmark has been quietly allowing Schuko sockets since 2008.

Denmark’s standard since the 1930s, and one of the few European plugs still incompatible with Schuko for earthing.

Three round pins in a shallow triangle, in a round recess. Brazil, and South Africa’s new standard.

The international standard plug that the IEC designed in 1986 hoping the world would adopt. Brazil did in 2007, South Africa is phasing it in, and nobody else has. Europlugs fit.

IEC 60906-1 was meant to be everyone’s plug. Brazil made it compulsory in 2007–11, which mainly meant Brazil went from two incompatible plugs to three for a while.

Three pins in a Y: two on top, earth below. Israel and the Palestinian territories only.

Israel’s wall plug, used nowhere else on Earth. Europlugs fit the round-pin sockets.

A local standard from the 1950s, redesigned with round pins in 1989 so European appliances could plug in.

The flat L-shaped plug that connects a hard drive or SSD to the motherboard.

Data between a hard drive, 2.5-inch SSD or optical drive and the motherboard, at up to 6 Gbit/s — plenty for hard drives, and the reason every SATA SSD tops out at the same speed. A separate, wider L-shaped plug from the power supply feeds the drive.

Serial ATA replaced the wide grey ribbon cables of IDE from 2003. SATA 2 (2005) doubled the speed, SATA 3 (2009) doubled it again to 6 Gbit/s, and there it stayed — NVMe drives on the M.2 slot took over the fast end.

The wider L-shaped plug from the power supply that feeds a SATA drive.

Power to a SATA drive — 3.3, 5 and 12 volts — off a lead from the power supply with several in a row. It plugs into the drive right next to the data connector.

Introduced with SATA in 2003 to replace the old four-pin Molex drive plug, which stayed on for years via adapters.

The white four-pin power plug from the days of IDE drives. Still on fans, lighting and adapters.

Power for every IDE hard drive and optical drive until SATA arrived, and still the power plug for case fans, LED strips, water pumps and as the adapter source for everything else in the case. 12 V on the yellow wire, 5 V on the red.

Made by Molex, hence the name, and used in PCs since the original IBM PC of 1981. SATA power replaced it on drives from 2003.

The 6-pin and 8-pin plugs into the top edge of a graphics card.

Extra power to a graphics card on top of what the slot provides: 75 W per 6-pin, 150 W per 8-pin. Big cards take two or three. Use the power supply’s own leads — splitters and adapters from Molex or SATA are how cards get starved and connectors melt.

The 6-pin arrived with PCI Express in 2004, the 8-pin with PCIe 2.0 in 2007. The 12VHPWR connector is meant to replace them on the biggest cards.

The compact 16-pin plug on the newest graphics cards. 600 W through one connector.

Up to 600 W to a graphics card through a single plug, on NVIDIA’s RTX 40 and 50 series and some others. The revised 12V-2×6 version has shorter sense pins so the card refuses to draw full power unless the plug is fully seated — the fix for the early ones melting.

Introduced with the ATX 3.0 standard in 2022. Early cards had connectors melt when the plug wasn’t pushed all the way home or the cable was bent hard at the plug; the 12V-2×6 revision of 2023 was the fix.

The smaller of the two graphics-card power plugs from the power supply.

Feeds a graphics card the power the slot alone cannot: 75 watts on top of the 75 the slot provides. Mid-range cards take one; bigger ones take an 8-pin, or two, or both. The cable comes off the power supply, usually labelled PCI-E or VGA.

Arrived with PCI Express in 2004, when cards first started needing more than a slot could deliver. The 8-pin version followed in 2007 as cards kept growing.

The little square four-pin plug that powers the processor on older boards.

Runs 12 volts straight to the processor, separately from the big 24-pin. Modern boards use the 8-pin EPS instead, and plenty take both — an 8-pin and a 4-pin side by side on the same board is normal, not a fault.

Introduced with the Pentium 4 in 2000, which is where the name comes from. The 8-pin EPS replaced it on anything with a serious processor from the mid-2000s.

The small white four-pin plug left over from the floppy drive era.

Powered 3.5-inch floppy drives, and now mostly powers nothing — though a few fan controllers, card readers and older graphics cards still use one. Most power supplies ship with one on the end of a Molex lead out of habit.

On every PC from the mid-80s until floppy drives disappeared in the mid-2000s. It survives because it costs nothing to leave on a cable.

The four-pin lighting header. The whole strip is one colour at a time.

Drives plain RGB lighting — fans, strips and coolers where everything on the lead shows the same colour together. Four pins: 12 volts, then red, green and blue.

Appeared on enthusiast boards around 2015 as case lighting became standard. It is slowly being replaced by the 5 V addressable header, but plenty of current boards carry both.

What the colours mean. Usually labelled 12V · G · R · B on the board.

The three-pin lighting header. Every LED can be a different colour.

Drives addressable lighting, where each LED on the strip or fan is controlled individually — rainbows, chases and the effects that sweep along a fan ring. Three pins: 5 volts, data and ground, in a four-wide housing with one position blanked.

Arrived around 2017 and has largely taken over from the plain 12 V header, because it can do everything that one can and a great deal more.

What the colours mean. Usually labelled 5V · D · ✕ · G on the board.

The big two-row plug from the power supply into the motherboard.

Every voltage the motherboard needs, through the thickest bundle of wires in the case. A PC that won’t power on after a build is this plug not fully seated more often than anything else. Most supplies make it as a 20-pin plus a detachable 4-pin, so it fits older 20-pin boards too.

Intel’s ATX standard of 1995 defined a 20-pin plug; ATX12V 2.0 (2004) added four pins for PCI Express graphics cards, and the 24-pin has been standard since.

The 8-pin (or 4+4) plug in the top corner of the motherboard that powers the processor.

12 V for the processor alone, into the connector near the CPU socket. High-end boards have two. A lot of "PC turns on but no display" jobs are this one left unplugged in the corner where it’s hard to reach.

The 4-pin ATX12V connector came in 2000 with the Pentium 4; the 8-pin EPS12V version came from servers and became normal on desktops from about 2007 as processors got hungrier.

The flat slot on the motherboard that a stick-of-gum SSD (or a Wi-Fi card) screws into.

NVMe SSDs — the fast ones — and, in a shorter slot, the Wi-Fi card, on every current motherboard and under the cover of every laptop. The card slides in at an angle and screws down at the far end. On many boards, filling a certain M.2 slot switches off two SATA ports; that "missing drive" after an upgrade is the manual’s fine print.

Standardised in 2013 (originally "Next Generation Form Factor") to replace mSATA and mini-PCIe. NVMe drives on it took over from SATA SSDs from about 2017.

The small row of pins on the motherboard that a case or CPU fan plugs into.

Power and speed control for case fans, CPU coolers and water-cooler pumps. Three pins power the fan and read its speed; the fourth carries a PWM signal so the board can control it precisely. The CPU_FAN header is special: many boards refuse to boot if nothing is plugged into it.

The 3-pin header dates from the first ATX boards of the mid-90s; Intel added the fourth PWM pin in 2004 so CPU coolers could be throttled quietly.

The nine pins the case’s power button, reset button and lights plug into.

Every wire from the front of the case lands here as its own tiny two-pin plug: power switch, reset switch, power LED and drive-activity LED. It is the single most fiddly connector in a build, and the one most often blamed when a finished machine will not turn on.

Standardised on the ATX board in 1995 and essentially unchanged since. Intel published a recommended pin order that most, but not all, board makers follow.

Runs the headphone and microphone sockets on the front of the case.

Carries the front panel’s headphone and microphone sockets back to the board’s sound chip, along with the sense wires that tell Windows something has been plugged in. One plug, all of it.

Replaced the older AC’97 header from about 2004. The plugs look identical and physically interchange, which is why front audio so often works at the wrong volume or not at all on older gear.

What the colours mean. Commonly a green header.

One header, two USB 2.0 ports — front panel, card readers, AIO coolers.

Provides two USB 2.0 ports from a single block of pins. Case front panels use them, and so do a surprising number of internal devices: water-cooler controllers, RGB controllers, card readers and the dongles for wireless keyboards mounted inside the case.

On boards since USB arrived in the late 90s, and still on every current one — internal devices keep needing it.

What the colours mean. Usually blue on the board.

The big blue shrouded block that feeds the case’s fast front USB ports.

Two USB 3 ports from one header — the 5 Gbit/s blue sockets on the front of the case. Much larger than the USB 2 header and on a finer pin pitch, with a solid plastic shroud around it.

Arrived with USB 3.0 in 2008 and is still current, though newer cases increasingly want the USB-C header instead.

What the colours mean. Almost always blue.

Feeds the USB-C socket on the front of a modern case.

A small, dense connector that carries one USB-C port to the case’s front panel at 10 or 20 Gbit/s. Nothing else on the board uses this connector, so if a case has a front Type-C socket and the board has no Key A header, that socket simply cannot be connected.

Standardised around 2016 as USB-C reached the front of cases. Common on mid-range and above boards from about 2019.

A small digital audio output header, mostly used by graphics cards once.

Carries a digital audio signal off the board — historically to feed a graphics card so it could send sound out over HDMI, and sometimes to a case bracket with an optical or coaxial socket on it.

Common through the 2000s. Graphics cards have handled HDMI audio themselves since about 2010, so the header is now rare and rarely used when present.

A serial port on the board, waiting for a bracket to bring it out the back.

Provides an old-style RS-232 serial port from a header rather than a socket. Still genuinely used: tills, scales, label printers, alarm panels, industrial gear and plenty of workshop equipment talk serial and nothing else.

Once a rear socket on every PC. As the back panel filled with USB, board makers kept the circuit and moved it to a header, where it remains on business and industrial boards.

Where a security chip plugs in — the thing Windows 11 wants.

Takes a small Trusted Platform Module, the chip that holds the keys for BitLocker and satisfies Windows 11’s security requirement. Most processors since about 2016 provide one in firmware instead, so the header usually sits empty.

On business boards since the mid-2000s. Windows 11’s launch in 2021 briefly made the modules sell for ten times their worth.

The long slot the graphics card goes in, usually the top one.

The widest and fastest expansion slot, and where the graphics card lives. Also takes capture cards, fast network cards and adapters that carry several NVMe drives at once.

PCI Express replaced AGP and PCI from 2004. The slot has not changed shape since — a card from 2004 fits a board from today — while the speed has roughly doubled with each generation, through to PCIe 5.0.

Half-length PCIe — mostly on servers and workstation boards.

Eight lanes rather than sixteen. Rare as a physical slot on a desktop board: makers usually fit a full-length x16 slot and wire only eight lanes to it, because a longer slot accepts more cards.

Part of PCI Express from 2004.

Short PCIe slot — NVMe adapters, fast network and storage cards.

Four lanes, which is exactly what an NVMe drive uses — so an M.2-to-PCIe adapter card in one of these runs at full speed. Also common for 10-gigabit network cards and RAID controllers.

Part of PCI Express from 2004.

The little one — sound cards, USB cards, capture and wifi cards.

A single lane, and plenty for most add-in cards: sound, extra USB, wifi and Bluetooth, TV tuners, serial and parallel brackets. Usually there are one or two on a board, often buried under the graphics card.

Part of PCI Express from 2004, and the direct replacement for the old PCI slot.

The cream-coloured slot that PCI Express replaced.

The expansion slot on every PC from the mid-90s until PCI Express took over. Sound cards, modems, network cards, TV tuners and capture cards all lived here. Still turns up on industrial and point-of-sale boards where a card has never been redesigned.

Intel introduced it in 1992, it replaced ISA through the 90s, and PCI Express replaced it from 2004. New consumer boards stopped fitting it around 2012.

What the colours mean. Traditionally cream or white.

The current desktop memory slot. The notch sits in its own place.

Where the memory sticks go on a current desktop board. Four slots is typical, two on cheaper boards. DDR5 moved the voltage regulation onto the module itself, which is part of why the sticks cost what they do.

Launched with Intel’s 12th generation in 2021 and AMD’s AM5 in 2022. The current standard on new machines.

The memory slot on nearly every machine built between 2015 and 2022.

By far the most common desktop memory slot in service today — the great majority of working machines that come across the bench take DDR4.

Arrived in 2014, went mainstream in 2015, and was the standard until DDR5 took over on new boards from 2021.

The memory slot on machines from roughly 2008 to 2015.

Still extremely common on older machines in daily use, and the reason a cheap memory upgrade is often still worth doing on a ten-year-old desktop.

Introduced in 2007 and the standard until DDR4 replaced it around 2015.

Mid-2000s memory. Same length again, notch in a different spot.

The memory slot on machines from about 2004 to 2008. Rare to see a working one now outside industrial gear and machines kept running for one piece of software.

Introduced in 2003 and replaced by DDR3 from 2007.

The oldest memory slot on this page, and deliberately so.

Memory for machines built around 2000 to 2004. If a board takes this, anything that goes wrong with it is usually the end of that machine — the parts are long out of production.

Arrived in 2000 and was replaced by DDR2 from 2003.

Laptop memory: half the length, lies flat, clips at the sides.

The memory slot in almost every laptop, plus mini PCs and some all-in-ones. The stick goes in at an angle and clips down flat rather than standing up.

In laptops since the mid-90s, through every DDR generation. DDR5 SO-DIMMs are current; CAMM2 is beginning to replace them at the top end.

The postage-stamp memory card with one corner cut off. Cameras, laptops, card readers.

Storage for cameras, and the card slot on many laptops and every card reader. Capacity classes (SD, SDHC, SDXC) and speed classes are printed on the label, and the same slot has taken every generation since 1999. Deleted photos are usually recoverable if you stop using the card immediately.

SanDisk, Panasonic and Toshiba launched it in 1999. SDHC (2006) took capacities past 2 GB, SDXC (2009) past 32 GB.

The in-between SD card. In phones for about four years, then gone.

The memory card in mid-2000s phones — Nokia, Motorola, Sony Ericsson — before microSD took over. It was sold with an adapter shell so it could go into a full-size SD slot, and that shell is usually where you find one today.

Added to the SD standard in 2003 to get a smaller card into phones. microSD arrived two years later at a third the size and finished it off; new phones had stopped using it by about 2008.

The fingernail-sized card in phones, dashcams, drones, the Switch and the Raspberry Pi.

Storage in phones, dashcams, security cameras, drones, action cameras, the Nintendo Switch and the Raspberry Pi. An adapter shell turns it into a full-size SD for a camera or card reader. Dashcams and security cameras kill them by rewriting constantly — use a "high endurance" card and expect to replace it every couple of years anyway.

SanDisk’s TransFlash of 2004, adopted as microSD in 2005. The miniSD in between them died out by 2008.

The thick 43 mm card with two rows of pin holes on the end. Professional cameras and old industrial gear.

The professional camera card for two decades — older Canon and Nikon DSLRs still use it — and the storage in industrial machines that use it as a hard drive.

SanDisk introduced it in 1994, built on the PC Card interface. CFast and then CFexpress replaced it in cameras from 2014; the last CF DSLRs were sold around 2020.

The paper-thin black card from 1990s cameras and music gear.

The memory card in Olympus and Fujifilm cameras before xD, and in a great deal of music gear — Yamaha and Roland keyboards and samplers loaded their sounds off these. It is barely thicker than a business card, which is also why so many of them have been bent.

Toshiba introduced it in 1995 as the "solid state floppy disk card". It never went past 128 MB, and CompactFlash and SD had pushed it out of cameras by about 2005.

What the colours mean. Which corner is cut says what voltage it wants — 5 V and 3.3 V cards are cut on opposite sides and are not interchangeable. A card that will not go into a camera the right way round is usually the wrong voltage, not a broken slot.

The small thick card only Olympus and Fujifilm cameras ever used.

Storage for Olympus and Fujifilm digital cameras through the 2000s, and nothing else — no other maker ever adopted it. Almost every one that comes over the counter belongs to somebody trying to get a decade of family photos off a camera that still works fine.

Olympus and Fujifilm launched it in 2002 to replace SmartMedia. It never went past 2 GB, cost far more than SD for the same capacity, and both makers had moved their cameras to SD by about 2010.

Sony’s own card, the long violet one. Sony gear, and nothing else.

The memory card in Sony cameras, camcorders and Walkmans. Sony put it in nearly everything they made for well over a decade and almost nobody else ever used it, so a Memory Stick in a drawer is very nearly always out of a Sony camera.

Launched in 1998, with Memory Stick PRO (2003) lifting the 128 MB ceiling. Sony finally gave in and fitted SD slots to its cameras from about 2010, and the format was done.

What the colours mean. The original Memory Stick is violet and holds 128 MB at most. Memory Stick PRO is usually black or white with PRO printed on it, and holds more. The colour is the quickest way to tell which you have.

The half-length Sony card — PSP, Cyber-shot, Handycam.

The card in the PlayStation Portable, most Sony Cyber-shot cameras and the Handycam camcorders of the late 2000s. It shipped with an adapter shell that turned it into a full-size Memory Stick, so plenty of them are still sitting inside one.

Memory Stick Duo arrived in 2003 as a shortened Memory Stick, and PRO Duo lifted the capacity limit the same way PRO did. Sony moved its cameras to SD around 2010; the PSP used it to the end.

The wide slot on the side of 2005–2012 laptops for add-in cards.

Add-in cards for laptops of that era: USB 3 ports, eSATA, card readers, sound cards and 3G modems. Often shipped with a blank plastic dummy card in it, which people think is a card.

Launched in 2003 to replace the PCMCIA/CardBus slot, and mostly gone by 2012 once USB 3 and Thunderbolt made add-in cards pointless.

The credit-card-sized slot on 1990s and early-2000s laptops. Modems, network cards, Wi-Fi.

How a laptop got a modem, a network port, Wi-Fi, a SCSI port or extra storage before those were built in. The cards had a little pop-out dongle for the cable. If an old laptop’s Wi-Fi lives in this slot, a modern USB Wi-Fi adapter is a better fix than hunting for a replacement.

The PCMCIA association defined it in 1990 for memory cards; the faster CardBus version came in 1995. ExpressCard replaced it from 2003 and it was gone from new laptops by about 2008.

The small oblong hole with a padlock icon, for a cable lock.

A cable lock’s anchor on laptops, monitors, docks and desktop cases: the lock’s T-shaped head goes in, turns, and can’t come out without the key, so the machine stays chained to the desk. It won’t stop a determined thief with tools; it stops the opportunist in a café.

Kensington introduced it in 1992 and most makers adopted the slot. Thin laptops have shrunk it (the Nano slot, 2017) or replaced it with the wedge-shaped Noble slot.