How The Heck Do Touchscreens Work?

retained text

In the last entry in the series, we talked about locks, an entirely mechanical system that has remained fundamentally the same over thousands of years. This time around, we're going to explore a technology on the other end of the spectrum, something electronic that has changed significantly in recent decades: touchscreens!
I'm of the age where I vaguely remember the Palm Pilot in the late 90s, and vividly remember Steve Jobs' legendary unveiling of the iPhone in 2007. I've used modern touchscreens for more than half of my life, but I realized I didn't really understand how they worked. How was the iPhone different from the Palm Pilot? Why doesn't it work if I have gloves on? Why does it act so weirdly when it gets wet? Can I use a sausage as a stylus?
This piece will dive into all of those questions, and more.
The old way: resistive screens
Resistive screens trace back to a physicist named Sam Hurst. While at the University of Kentucky in the late 1960s, he and his students were stuck digitizing endless reams of data from chart recorders by hand, so they built a touch sensor called the Elograph to speed it up. He went on to start a company called Elographics, which produced the Accutouch, one of the first transparent touchscreens, in 1974.
The technology works by having two stacked conductive layers, with tiny spacer dots between them to keep an air gap. The device's touch controller applies a small voltage across one layer, high on one edge and zero on the other. Because the coating is slightly resistant, the voltage fades evenly from one side to the other, like a gentle slope.
The other layer is not powered and only listens. When you press down, the top layer touches the bottom one at a single point, connecting them right there. The listening layer picks up the voltage at that spot on the slope, which tells the screen how far across you pressed.
It takes two readings, one with the voltage running left to right and one top to bottom, to get both coordinates. This process has to be fast enough for there to be no noticeable gap between getting the X and Y position. The controller flips back and forth between the two measurements, often a hundred or more times per second.
The interactive below shows the process by letting you press down on a screen and see what's happening inside. The side view only shows a slice along one direction, but remember the same thing is happening in the other.
How does a voltage become a position?
The math is just a fraction. If the controller runs the slope from 0 volts on one edge to 3.3 volts on the other, and the listening layer reads 1.1 volts, you pressed one third of the way across, since 1.1 ÷ 3.3 = 0.33. On a screen 300 pixels wide, that's about pixel 100. The controller then does the same thing with the slope running top to bottom to get the other coordinate.
Because resistive touchscreens are based primarily on pressure, they work with anything that can press: a fingertip, a fingernail, a stylus, or even a gloved hand. That's why devices like the Palm Pilot came with a stylus (which Steve Jobs was famously not a fan of), and why you could use them in winter without taking your gloves off.
Who wants a stylus? You have to get 'em and put 'em away, and you lose them. Yuck. Nobody wants a stylus!
Capacitive screens make you part of the circuit
While resistive screens work by sensing pressure, capacitive screens work by sensing you directly. This is why pressing down harder on your smartphone doesn't do much.
Instead of two layers being pushed together, a capacitive screen has a grid of electrodes (thin, transparent strips of conductive material) under the glass. One electrode, the sender, constantly puts out a small electrical signal. Its neighbor, the listener, picks some of it up across the gap between them, through an invisible electric field that extends above the glass.
The trick that makes it all work is that your body, being mostly salty water, happens to be a good conductor of electricity. When your finger gets close, it pulls some of the field towards itself, so less of the signal reaches the listener electrode. In a sense, the screen only knows that some of its signal has gone missing.
The visual below shows the system for one pair of electrodes. Ordinary gloves don't conduct electricity and keep your skin too far from the field for the screen to notice much, and a plastic stylus doesn't conduct electricity at all. Switch to a bare finger and drag closer to see how distance and conductivity change the reading, including the drop you get before anything touches the glass.
In one of the more bizarre How The Heck facts, during a particularly cold winter in 2010, people in South Korea started using snack sausages as “meat styluses” so they could use their phones without taking off their gloves. One sausage company reportedly saw its sales jump by 40%. A sausage, also salty and full of water, is a good enough conductor to stand in for a finger.
Can touchscreens tell how hard you're pressing?
Resistive screens can, roughly. A harder press squeezes the two layers together over a larger area, which lowers the resistance where they touch, and some controllers take an extra reading to measure it.
Capacitive screens can't measure force directly. A harder press flattens your fingertip so it covers more of the grid, and some phones use that contact size as a rough stand-in for pressure.
Apple's 3D Touch, introduced with the iPhone 6s in 2015, added a separate force-sensing layer under the display for exactly this reason. Apple later replaced it with a long-press feature called Haptic Touch.
From one crossing to a grid
One sender-listener pair is enough to know that a finger is nearby, but to find a finger anywhere on a screen, we'll need a lot of pairs.
In a capacitive screen, the electrodes run in rows and columns across the whole screen, with senders going one way and listeners the other. Every intersection forms a crossing, each one working just like the pair in the interactive above.
The controller sends on one row at a time and listens on every column at once before it moves on to the next row. This sweep happens many times per second, a bit like the resistive screen flipping between X and Y.
In practice, these strips are made of chains of diamonds that interlock side by side, as shown below. This keeps every sender right next to a listener, so the field can bulge up toward your finger instead of running straight between them.
From a grid to a location
My intuition was that there were probably several thousand crossings in my smartphone, so I was surprised to learn it's closer to a few hundred, formed by a few dozen sender rows and listener columns.
For the sake of example, let's consider a small 6 × 8 grid below.
 Hover over the grid to simulate hovering your finger.
Tap on the grid to simulate tapping your finger. 
Each crossing essentially gives us a reading of how much signal is being pulled away by a conductor like your finger, refreshed many times per second. But your phone still needs to turn all of those numbers into the location your finger is tapping.
The simplest approach is to find the crossing with the biggest change and say the finger is there. Try it below, and notice how the target jumps from crossing to crossing as you drag.
This approach was good enough for many early touchscreens. In 1972, the PLATO IV educational terminal used a 16 × 16 infrared touch grid so students could tap their answers, and CERN's first capacitive screens had just 16 fixed buttons. It's still essentially how many touch buttons on microwaves and stoves work today. But for a phone, where you're scrolling, dragging, and tapping small targets anywhere on the screen, jumping between crossings isn't good enough.
Instead of picking a single winner, your phone can blend the readings together using a weighted average. Although there are more than two crossings, you can conceptually think of it like finding the balance point of a seesaw. If two crossings changed by the same amount, your finger is halfway between them. If one changed more, the balance point tips toward it.
Real touch controllers combine both approaches: they first find the crossing with the biggest change, like before, then average just the crossings around it. The result is a position far more precise than the spacing of the grid itself.
I was also surprised to learn capacitive screens actually came before resistive ones: the first was invented in 1965 for British air traffic control. Resistive screens just won out for decades because they were cheaper and simpler to make.
Special cases
Multi-touch
Because every crossing is read independently, two fingers show up as two separate clusters, each with its own peak. The phone finds each peak and averages around it, giving one position per finger. If you bring two fingers close enough together, though, their clusters blur into one, and the phone sees a single touch.
The old resistive screens could never do this. Since the listening layer only reports one voltage at a time, two presses connect it to the slope at two spots, and the readings blend into a single value somewhere in between. Two fingers always look like just one finger in the middle.
Palm rejection
As screens and other touch surfaces got bigger, a new problem came up: your hand. On a tablet or a laptop trackpad, it's natural to rest your palm on the surface while you write, draw, or type. I'm doing it as I type this.
As far as the grid is concerned, a palm is just another conductor, and it changes the readings just like a finger does. The main difference is size.
A fingertip only affects a handful of crossings, while a palm creates a cluster far bigger than any finger could. So the software treats anything that large as an accident and ignores it.
Try dragging your finger around the grid below. The palm resting near the edge still shows up in the readings, but the phone never treats it as a touch.
Water
At the start of this piece, I asked why my phone's screen acts weirdly when it gets wet. Water is conductive, much like your finger, so you might expect every droplet to register as a touch.
In practice, a puddle sitting on the glass on its own usually only causes a weak, oddly shaped change, which the phone can learn to ignore. The real trouble starts when your finger touches the water. Now the water is connected to you, so it spreads your touch across the screen. The phone sees one big, smeared cluster instead of a fingertip, and its best guess at where you're pointing gets pulled off course.
The next time you're on your phone, or maybe as you read this on it, it's cool to think about what's really happening. Your phone is sweeping an invisible field hundreds of times a second, noticing where a bit of its signal goes missing on the grid, and then using a series of surprisingly intuitive algorithms to handle the annoying stuff the best it can!
We're just getting started.
Subscribe for more thoughtful, data-driven explorations.

document fetched 2026-10-28 02:54