From principles to real machines
By now you have all the moving parts. You know a telescope's aperture sets how much light it gathers and how fine its angular resolution can be; you know the diffraction limit ties that sharpness to wavelength; you know the atmosphere blurs and absorbs starlight, so we fight back with adaptive optics or escape it entirely with a space telescope; and you know a CCD turns the captured photons into numbers. This guide is where those ideas stop being abstract. We will walk past four of the most consequential observatories ever built and watch each one solve a specific problem you already understand.
Keep one organizing question in mind as we go: which wavelengths does this telescope hunt, and where must it sit to catch them? That single question decides almost everything about a facility's design — its mirror, its location, even whether it must leave the planet. A telescope built for one band of the electromagnetic spectrum often cannot see another at all.
Hubble and JWST: two eyes above the air
The Hubble Space Telescope, launched in 1990, is a fairly ordinary 2.4-meter mirror — smaller than many mountaintop telescopes — flung into orbit about 540 kilometers up. That modest size matters far less than its address. Above the atmosphere there is no seeing to blur it and no air to absorb the ultraviolet, so Hubble simply works at its diffraction limit all the time, delivering crisp images night and day for over three decades. Its deep-field exposures — staring at a single blank-looking patch of sky for days — revealed thousands of faint galaxies in what had looked like empty darkness, a vivid lesson in what patience plus a steady eye above the air can do.
The James Webb Space Telescope (JWST), launched in 2021, looks like Hubble's successor but answers a different question. Its 6.5-meter mirror — built from 18 gold-coated hexagonal segments that unfolded in space, because no rocket fairing is wide enough for a single piece — gathers far more light, but its real specialty is the infrared. Why infrared? Because the light from the very first galaxies, emitted billions of years ago, has had its wavelength stretched by the expansion of the universe until it now arrives in the infrared. To catch it, JWST must be brutally cold: it sits about 1.5 million kilometers from Earth, behind a tennis-court-sized sunshield, chilled to roughly 40 kelvin so that its own warmth does not drown out the faint cosmic glow it is trying to detect. A warm infrared telescope would blind itself with its own heat.
ALMA: an array on a high desert
Move now to far longer wavelengths — millimeter waves, sitting between the infrared and the radio you have already met. This is the home of cold things: the dusty disks where planets are forming, the molecular clouds where stars are born, gas in distant young galaxies. ALMA, the Atacama Large Millimeter/submillimeter Array, hunts exactly this band. It sits at about 5,000 meters altitude on a bone-dry plateau in northern Chile, because water vapor in the air greedily absorbs millimeter waves — so the higher and drier the site, the better. This is the same logic that pushes optical telescopes onto mountaintops, taken to an extreme.
But ALMA is not one telescope — it is 66 separate dishes spread across the plateau, and that is the crucial idea. Recall the diffraction limit: resolution gets *worse* as wavelength gets longer. Millimeter waves are thousands of times longer than visible light, so a single dish, even a big one, would give a hopelessly blurry image. The fix is to combine the signals from many widely separated dishes so that, for the purpose of sharpness, they act like one giant dish as wide as the whole array. That trick is called interferometry, and it is the heart of the rest of this guide.
Interferometry: many dishes, one giant eye
Here is the core idea, stripped to its bones. A wave of cosmic light arrives at two separated radio dishes at very slightly different moments, because one dish is a touch farther from the source than the other. If you record the wave at each dish with an extremely precise clock and then carefully combine the two recordings, the small timing difference tells you something exact about the direction the wave came from. Add more dishes, in more positions, and each *pair* contributes another clue. Software then weaves all those clues together into a single sharp image. The resolution you achieve depends not on the size of any one dish, but on the baseline — the distance between the dishes farthest apart.
angular resolution ~ wavelength / dish diameter (single dish) angular resolution ~ wavelength / baseline (interferometer) baseline = distance between the two most widely separated dishes -> sharpness scales with the SPACING, not the size of any one dish -> collecting area (sensitivity) still scales with TOTAL dish area
Notice the honest trade-off that the schematic makes plain. Spreading the dishes apart sharpens the image, but it does *not* gather more light — the total light-collecting power still depends on the combined area of all the dishes, not on how far apart they sit. An interferometer can therefore have superb resolution yet still need long exposures to detect faint things. This is also why ALMA's dishes are mounted on transporters: by physically rolling them closer together or farther apart, operators retune the array between a wide, sharp configuration and a compact, sensitive one, to match whatever they are observing.
VLBI and the first picture of a black hole
If resolution scales with baseline, an irresistible question follows: how far apart can we push the dishes? With very-long-baseline interferometry (VLBI), the answer is *across the whole planet*. Radio telescopes on different continents observe the same source at the same time, each stamping its recording with an atomic clock. There is no cable linking them; the data are stored and later brought together and combined. The effective baseline becomes the distance between the telescopes — up to the diameter of the Earth itself, about 13,000 kilometers. That synthesizes the resolving power of a single, frankly impossible, planet-sized telescope.
The Event Horizon Telescope (EHT) is exactly this: a global network of millimeter-wave dishes — ALMA among them — linked by VLBI into a virtual telescope the size of Earth. The goal was a target so absurdly small on the sky that nothing else could resolve it: the event horizon of the supermassive black hole at the center of the galaxy M87, and later Sagittarius A* at the heart of our own Milky Way. In 2019 the EHT released the first direct image of a black hole's silhouette — a dark central shadow ringed by glowing gas. The resolution required was roughly equivalent to reading a newspaper in New York from a café in Paris.
Be honest about what that image is, though. It is not a photograph in the everyday sense and the black hole is not 'sucking in' the surrounding gas like a drain — the dark patch is a *shadow*, the region from which light cannot escape to us, outlined by the glow of hot gas orbiting just outside it. The picture was not snapped in an instant; it was painstakingly reconstructed from data gathered across the planet over several nights, with the gaps between the few real dishes filled in by careful modeling. VLBI does not magically give you a complete Earth-sized mirror — it gives you a sparse handful of points on one, and the rest is reconstruction, done with great care and stated uncertainties.
Why we build so many different eyes
Step back and the pattern across all four facilities is the same. Each is a deliberate answer to two questions: which wavelengths, and from where? Hubble chose visible and ultraviolet, so it went to orbit to escape blur and absorption. JWST chose the infrared, so it fled far from Earth's warmth and chilled itself to a whisper above absolute zero. ALMA chose millimeter waves, so it climbed a dry high desert and spread into an array to beat the diffraction limit at long wavelengths. The EHT chose the sharpest possible view of a black hole, so it borrowed the whole Earth as a baseline. No single telescope can do all of this; the cosmos speaks in every band at once, and we need a different ear for each.
And the lesson of interferometry reaches well beyond radio dishes. The same logic — combine separated detectors and recover the precision of one enormous instrument — runs through much of modern astronomy, from optical interferometers that link mountaintop telescopes to the gravitational-wave detectors you will meet later, whose kilometers-long arms work on a kindred principle. You now have the whole observing toolkit: gather the light, sharpen it, place the instrument where the sky is kind, turn photons into numbers, and, when one eye is not enough, link many into a giant. With that, you are ready to leave the how of observing behind and start asking what the data actually reveal.