⏺️ Cylinder: The Invisible Elevator in Your Hard Drive
A hard disk cylinder is not a physical object—it is a logical collection of all the tracks that lie directly on top of each other across multiple platters. When the read/write heads are stationary, every track they can reach forms one cylinder. Because the heads move together, accessing data from the same cylinder means zero seek time. This article explores the geometry of a hard drive, the meaning of CHS (Cylinder‑Head‑Sector)[1], how engineers use cylinders to speed up data access, and why this concept still matters even in the age of SSDs. Keywords: platter track seek time actuator.
🥞 Platters, Tracks & the Magic of Alignment
Imagine a stack of metal pancakes, each coated with a magnetic layer. That is exactly what a hard disk drive (HDD) looks like inside. Each “pancake” is called a platter. Every platter spins at thousands of revolutions per minute. Both the top and the bottom of each platter have a read/write head that hovers a few nanometres above the surface. The heads are all attached to a single arm (the actuator) and move together—they cannot move independently.
Now imagine drawing circles on every platter with a compass. Each circle is called a track. A track is simply a ring where data can be stored. If you look from above, a track on the top platter has the same diameter as a track on the bottom platter. Because all heads are aligned vertically, when the arm moves to a specific radius, every head sits exactly above the same track number on its respective surface. The set of all tracks that can be read or written without moving the arm is one cylinder.
| Part | What it looks like | Job |
|---|---|---|
| Platter | Rigid disk (aluminium/glass) | Magnetic surface that stores bits |
| Track | Concentric ring on one surface | Path where data is written |
| Cylinder | Stack of same‑numbered tracks | All data reachable without moving heads |
| Sector | Tiny arc of a track (usually 512 B) | Smallest storage unit |
📍 CHS: The Postal Code of Old Hard Drives
In the 1980s and 1990s, every sector on a hard disk was located using three numbers: Cylinder, Head, Sector (CHS). The operating system would say: “I need cylinder 5, head 2, sector 10.” The disk controller knew exactly which track (cylinder), which surface (head), and which block (sector) to read. Because the actuator moves all heads together, picking a cylinder means the mechanical arm has to travel only once. Once the arm is on cylinder 5, every head is already over the correct track, and the drive just waits for the spinning to bring the right sector under the head.
Mathematically, the capacity of a disk using CHS was simple:
$Capacity = Cylinders \times Heads \times Sectors \times 512\ bytes$
If a drive has 1024 cylinders, 16 heads, and 63 sectors per track, total capacity is $1024 \times 16 \times 63 \times 512 = 528\ MB$.
Notice the word Heads in the formula—it is actually the number of recording surfaces. Since one head reads one surface, Heads equals the number of platters multiplied by 2. So a cylinder includes exactly Heads tracks (e.g., 8 surfaces = 8 tracks per cylinder).
⚙️ How Windows 98 Used Cylinders to Boot Faster
Let us travel back to 1998. A computer has a 4.3 GB hard drive. When you save a very large file, the operating system tries to place all pieces of that file inside the same cylinder—or as close as possible. Why? Because reading the first piece requires a seek (moving the arm). Once the arm is on cylinder 120, all the heads are ready. If the next piece is also on cylinder 120 (but on a different head), the drive can switch electronically from one head to another in microseconds. But if the next piece is on cylinder 121, the arm must move—a slow, mechanical operation.
Modern hard drives use a different addressing scheme called LBA[2] (Logical Block Addressing), which hides the cylinder geometry. However, inside the drive’s firmware, the concept of cylinders still helps to arrange data. The drive’s controller knows which physical cylinders exist and stores frequently accessed data together in the same cylinder to minimise seek time. That is why after “defragmenting” a disk, files feel snappier—they have been rearranged so that their parts lie in the same cylinder or neighbouring cylinders.
| Operation | Typical time | Compared to CPU cycle |
|---|---|---|
| Switch between heads (same cylinder) | 1–2 µs | ≈ 2000 CPU cycles |
| Seek to next cylinder | 0.5–1 ms | ≈ 1 million cycles |
| Seek across full disk | 10–15 ms | ≈ 20 million cycles |
❓ Questions Students Always Ask About Cylinders
No. It is an imaginary tube. The platters are separate disks with a gap between them. There is no metal or plastic cylinder connecting them. The word “cylinder” describes the shape you would get if you connected all the tracks of the same radius—a hollow virtual cylinder.
No. SSDs have no moving parts, so they have no tracks, heads, or cylinders. They use NAND flash chips. However, the term “cylinder” is still used in some legacy software for compatibility. When you see CHS values in an SSD utility, they are simulated, not real.
Exactly the number of recording surfaces. If a drive has 3 platters, there are 6 surfaces; therefore one cylinder contains 6 tracks (one on each surface). Some very old drives used only one side of a platter, but virtually all modern HDDs use both sides.
🎯 Why the Cylinder Still Matters
The cylinder is one of the most elegant ideas in computer engineering: it uses the physical alignment of heads to create a zero‑cost data group. By storing related data within the same cylinder, old hard disks could work almost as fast as the electronics allowed. Even though modern operating systems no longer expose CHS to programmers, the principle lives inside disk firmware. When you hear a hard drive clicking softly, the actuator is moving from cylinder to cylinder—searching for your pictures, your music, or the game you love. The cylinder is the reason that moving the arm just one step can unlock dozens of tracks at once.
📚 Footnote
[1] CHS (Cylinder‑Head‑Sector): An early method of addressing data on a hard disk. Each sector is uniquely identified by its cylinder number, head number (which surface), and sector position on the track.
[2] LBA (Logical Block Addressing): A simpler, linear addressing scheme where each sector is numbered from 0 to N‑1. The disk controller internally translates LBA to physical cylinder, head, and sector.