Discussion about old computer accessories: the sound coupling device
Discussion about old computer accessories: the sound coupling device
I only noticed one of these gadgets during middle school in the computer lab. It allowed connecting to a remote host through a basic green-screen terminal. I have no idea how quickly the link worked, but it was definitely very slow—perhaps around 300 baud. Almost anyone could type faster than the system replied. Much more advanced modems appeared about 15 years later, capable of speeds up to 56.6kbs, and those devices could cost over $200—about $463 today. Acoustic coupler: https://www.computinghistory.org.uk/det/...c-Coupler/
Yeah, I remember using one of those back in the '70s. I'm pretty sure 300 baud was accurate.
Many people don't know that the well-known IBM Selectric typewriter, with its interchangeable "golf ball" type sphere, was actually built as a computer input/output device, capable of printing much quicker than a regular typewriter but still smaller and quieter than a line printer. I relied on terminals based on that design to connect to a system running APL on a central IBM 360/50 machine. That language was meant for interactive work, not just punch card submissions. Usually, several terminals were placed in one central room for general access. However, we had one dedicated terminal in a Chemistry Department undergraduate lab, using one of these modems and a landline phone to reach the mainframe.
I discovered touch typing on the IBM Selectric. I was one of just two students in my high school typing class. Those machines were quite costly back then.
An IBM 360 is quite old. Have you ever seen one of the well-known IBM 30/30 "Winchester" hard drives?
I recall the large industrial dot matrix printers the mainframe staff used, which always had jobs in progress.
Expensive, indeed! The selectrics delivered outstanding results and gained traction in both business and academic circles. This was largely due to the ability to switch fonts simply by altering the "golf ball" type head. However, the supplies could be costly since most relied on single-use ribbon cartridges that provided sharp type but only lasted through one use. Multi-use fabric ribbons were also an alternative option.
To be precise, the "Winchester" hard drive design emerged in the 1960s and remains a foundational concept today. Its distinctive aspect lay in the understanding that the gap between the head and the spinning hard drive disk wasn't created by rigid support, but rather by a dynamic airflow generated as the disk spun. This air movement could stabilize the head if it rested on a flexible arm. The small gap allowed for excellent magnetic coupling, enabling quicker data transfers. The main limitation was that this mechanism depended on the disk rotating, so starting and stopping had to occur only when the arm retracted to a safe position.
I don’t recall if this drive design was implemented in the mainframe systems of the late 60s. I am certain, though, that a previous random-access storage method was already in use. There were several large floor-mounted machines, each containing removable "disk packs." These consisted of rigid double-sided magnetic disks about 12 inches in diameter, comparable to an LP phonograph record. They were mounted on a vertical shaft roughly 10 inches high. The entire pack was housed in a protective case to prevent dust, and removed for insertion into the drive. The drive featured arms for each pack, keeping them at fixed clearances over the disk surfaces. As the disks spun, the arms tracked their positions, similar to modern hard drives. Unlike Winchester drives, these units didn’t employ the "air cushion" technology but offered interchangeable packs that allowed each drive to handle different data sets. For instance, the APL system couldn’t run normally on the mainframe’s OS, so it was temporarily swapped into a different OS for APL operations. Afterward, the system returned to standard multi-use functionality.
The storage capacity per pack wasn’t specified, but in the system I used, each user received several workplaces of 32 KB, storing a few hundred users’ worth of data. Within each workplace, software and files could be copied between locations.
For those unfamiliar, APL stands for "A Programming Language"—a term quite original at the time! It was heavily focused on mathematics. All data were essentially vector arrays, ranging from 0-dimensional constants to 2-dimensional flat arrays. Larger structures added more dimensions and indices. Many symbols, mainly Greek letters, represented pre-defined operators for custom or quick calculations. Most functions worked with scalar or vector variables, while built-in operations favored matrix methods suited to higher dimensions. These programs could be called subroutines by specifying their name and input arguments. They offered powerful capabilities that even non-mathematicians found useful.
By the way, the standard Line Printer used in computer centers back then wasn’t a dot matrix printer. It printed significantly faster and used a different technology. If you’d like a detailed explanation, feel free to share. This is unrelated to your original comment.
So the mainframers didn't use a dot matrix printer. That explains why it wore out paper much quicker than any single-head dot matrix could have. I'd like to learn more about the Line Printer. It definitely caused quite a stir — even when the cover was in place.
I once spoke with a legal word processor user who joked about my slow typing speed of 40 WPM compared to her 100 WPM. She charged $100 per hour, which was around $277 a day back in the late 1980s. I'm pretty sure she was using a computerized word processor by then, not a typewriter.
My knowledge about the mainframe systems Line Printer was this. It operated with tractor-feed paper, which consisted of strips with holes punched along the edges for quick paper handling. The sheets were generally around 15 inches wide. It could print an entire line in a single pass. There was a continuous band of cast metal blocks, each displaying one character, which could be pressed against the paper using an ink-coated ribbon to form a letter image. This method was similar to an older typewriter, though the characters were arranged on a continuous strip rather than individual arms. The metal band might have been composed of linked units or sliding blocks within a channel. This strip contained three full sets of uppercase letters plus one set of special characters, rotating continuously across the paper width and returning. Behind it, a head scanned the sheet forward and backward while carrying a hammer. Timed signals to the hammer would push the correct character block forward, ensuring accurate printing in the proper position. The horizontal spacing between letters was fixed at 10 characters per inch. The rapid hammer strikes and paper feeding generated considerable noise, so the printer featured a sound-absorbing enclosure to reduce the noise. It was truly incredibly fast and usually in active use. The operating system would load pages with labels at the start of each job. System operators would periodically remove the printed output, breaking it into user packets and delivering them for collection with the original punched cards submitted for the task.
Those "band" printers were placed in enclosures primarily for operator safety rather than noise control. If a band failed, it could instantly cause serious injury. The bands were very thin and moved rapidly, often several feet per second. I witnessed one fail; luckily the enclosure remained intact. The noise was unbearable and the internal damage required weeks of repair.