When you could see the tracks with your own eyes
You already know the master idea of this rung: a detector reads the ionization track a charged particle leaves as it tears through matter, and from a curved track in a magnetic field it extracts charge and momentum. Before any electronics existed to do that reading, physicists used a gorgeous trick — they made the ionization trail visible enough to photograph. The three classic instruments that did this, grouped together as the historic detectors, are the cloud chamber, the bubble chamber, and the photographic emulsion, and for roughly half a century they were how the field saw.
Each makes the same trail visible in a different way. A cloud chamber holds a gas right on the edge of condensing into fog; the ions a passing particle leaves become seeds on which tiny droplets grow, drawing a thin thread of mist you can photograph. A bubble chamber does the reverse in a liquid — usually liquid hydrogen — held just past its boiling point: along the ion trail, tiny bubbles bloom and are snapped before they grow, giving denser, sharper tracks in a medium that doubles as the collision target. A photographic emulsion is a thick block of photographic film in which the particle's passage exposes a line of grains that, once developed, can be traced under a microscope to within a fraction of a micron. Put any of them in a magnetic field, and the curvature of the tracks hands you momentum, exactly as in the modern picture.
Why beautiful pictures were not enough
For all their elegance, the photographic detectors had a fatal pair of weaknesses, and naming them tells you exactly what the modern era had to invent. The first is speed. A human being had to sit and scan the photographs one frame at a time, measuring tracks with a ruler and a loupe. A bubble chamber might take only a handful of pictures a second, and an army of scanners might labour for months over a single run. The second, deeper flaw is selectivity: the chamber recorded whatever happened to drift through it, and could not be told to keep only the interesting collisions. If the process you wanted was one in a billion, you photographed the other billion too — and then someone had to look at them all.
The breakthrough that broke this logjam was the wire chamber. Instead of photographing droplets, Georges Charpak's 1968 invention strung a gas volume with many fine wires held at high voltage; when a particle's ionization frees electrons, they drift to the nearest wire, avalanche in its strong field, and produce an electrical pulse — telling you which wire fired, electronically. Suddenly a track could be read out as numbers and fed straight to a computer, with no film to develop and no human to scan. This single shift, recognized with a Nobel Prize, is what made the modern detector possible: a wire chamber (and its descendants, like the silicon trackers you met earlier) turns the soft, slow language of photographs into the fast electronic stream that everything downstream depends on.
The flood: forty million crossings a second
Electronics solved the bubble chamber's slowness, but it then created a problem of the opposite kind — far too much data, far too fast. At the Large Hadron Collider, the protons are not a smooth stream; they ride in tight clumps called bunches, and two bunches cross about forty million times a second. Each of those crossings is a potential bunch crossing in which not one but dozens of proton pairs may collide at once — the messy phenomenon called pile-up. Reading out the whole detector for every crossing would mean writing something like a petabyte of data every second. There is not enough storage on Earth, let alone enough bandwidth out of the detector, to keep even a small fraction of it.
There is a saving grace, and it is the whole reason this works: almost every one of those crossings is boring. The overwhelming majority produce only ordinary, long-understood physics — the well-mapped roar of the crowd. The rare event you actually came for, a Higgs boson or some hint of something new, hides among them at perhaps one in a billion or worse. So the task is not to record everything; it is to recognize, in the instant a crossing happens, whether it might be one of the precious few — and if not, to let it vanish forever. That recognizing-and-discarding is the job of the trigger.
The trigger and DAQ: deciding what not to keep
The trigger and data-acquisition system is the experiment's real-time filter, and it works in stages because a careful decision is slow and a fast decision is crude — so you do both, in order. A first level, the Level-1 trigger, is built from custom electronics sitting right on the detector; it looks at coarse, fast information — say, a high-energy lepton or a big lump of missing transverse energy — and within microseconds throws away the vast majority, cutting forty million crossings a second down to perhaps a hundred thousand. A second level, the high-level trigger, is software running on a large farm of ordinary computers; it reconstructs the survivors far more carefully and trims that down to a few hundred or few thousand events per second. Only those are written to disk. The DAQ is the plumbing that reads out every channel of an accepted event, assembles its millions of fragments into one coherent record, and ferries it to storage.
~40,000,000 crossings/s (every bunch crossing)
| Level-1 trigger (custom electronics, microseconds)
v
~100,000 events/s
| high-level trigger (software farm, full reconstruction)
v
~1,000 events/s --> DAQ assembles & writes to disk --> analysisThe giants: ATLAS and CMS
Wrap the onion of guide one — tracker, calorimeters, muon chambers, all inside a magnet — around an LHC collision point, scale it to the size of a cathedral, and feed it the trigger and DAQ we just described, and you have a general-purpose detector. The two giants at the LHC, ATLAS and CMS, are exactly this: building-sized instruments, tens of meters across and weighing thousands of tonnes, with tens of millions of readout channels each, built to study almost anything a collision might throw out rather than to chase a single effect. They are not specialists; they are layered detectives designed to reconstruct whole events.
Why two of them, doing the same job, at the same machine? Because the most important results in this field demand independent confirmation, and the surest way to get it is to build two detectors by separate teams with deliberately different choices. ATLAS and CMS use different magnets, different technologies, different layouts — CMS, true to its name (the Compact Muon Solenoid), is built around one enormous solenoid and is famously heavy; ATLAS is larger and uses a distinctive toroidal magnet system. The payoff was historic: in 2012 the two collaborations announced the discovery of the Higgs boson essentially simultaneously, and the fact that two differently-built machines saw the same bump made the claim far more convincing than either could have been alone.
Be honest, though, about what "general-purpose" does not mean. It does not mean these detectors see everything. Neutrinos pass through all those tens of thousands of tonnes entirely undetected, betrayed only as missing transverse energy. The instruments cover most but not all directions, so some particles escape straight down the beam pipe. And no detector ever sees a quark or gluon directly — colour confinement guarantees they dress into hadron jets first, so what is recorded is always the aftermath. "General-purpose" means broad, not omniscient.
What changed, and what stayed exactly the same
Stand back and the arc is clear. Across seventy years the technology was transformed beyond recognition — from a single chamber photographed a few times a second and scanned by eye, to a cathedral of silicon and gas reading a hundred million channels and a trigger sifting forty million crossings a second. Yet the underlying physics never moved an inch. Every one of these instruments, old and new, is doing the same thing: a charged particle ionizes the matter it crosses, a magnetic field bends its path to reveal momentum and charge, and dense material stops it to measure its energy. The bubble chamber and the ATLAS silicon pixel detector are reading the very same four clues you met in guide one — momentum, energy, charge, identity. Only the speed, the scale, and the cleverness of the bookkeeping have changed.
There is a quiet lesson here for what comes next on this ladder. The bubble chamber's downfall was that it kept everything and could choose nothing; the LHC's triumph is that it keeps almost nothing and chooses ferociously well. That shift — from passive recording to active, real-time selection — is why a modern experiment can find a one-in-a-trillion process at all. When you move on to the analysis side of the field — pulling a faint signal from a loud background, and bump hunting — remember that the data you get to analyse is never the whole story; it is precisely, and only, what the trigger decided was worth keeping.