The world is slowly, apparently, going through a floppy disk apocalypse. Tom Persky, the founder of floppydisk.com and the last true major reseller of floppy disks, has been warning us for years that his finite supply of 3.5-inch disks is dwindling. The massive stockpiles once used to service Boeing 747 avionics and aging medical equipment are evaporating, leaving the retro-computing world staring down a supply chain crisis with no factory left to restart. While creators like Polymatt have recently experimented with baking their own disks to prove that the medium can be resurrected by hand, isolated breakthroughs often slip under the radar. We may be indeed entering an era where the only way to save the technology is to reinvent it in our kitchens. This is exactly what engineer Nina Kalinina did in late 2023 when she decided that if she could not buy the past, she would synthesize it herself using office supplies and a toaster oven.

Through a multi-part saga documented on Mastodon, she embarked on a project that seemed equal parts mad science and historical reenactment. Her goal was simple but incredibly difficult. She wanted to manufacture a functional 5.25-inch floppy disk using nothing but office supplies, craft store powders, and a kitchen oven.

The project began with a hypothesis that sounded like a lunchtime dare. A standard floppy disk is essentially a circular piece of Mylar plastic coated in a magnetic medium. If you could replicate that medium, could you replicate the data storage?

Kalinina started with the most basic materials available. For the substrate, she eschewed the high-grade polyester of the 1980s and opted for a simple A4 sticky label sheet. For the magnetic medium, she turned to Iron(II,III) oxide, known chemically as Fe3O4 .

. This is commonly found in black artist pigments (Mars Black) or toner powder. It is magnetic, but it is not the industry standard for data storage.

The initial construction process was crude. She printed the outline of a 5.25-inch disk onto the sticky label, cut it out with scissors, and punched an index hole. Then came the messy part. She applied the black iron oxide powder directly to the adhesive surface of the label, scrubbing away the excess with a sponge. To test this contraption, she gutted a commercial floppy disk, removing its magnetic “cookie” (the internal disk) and replacing it with her paper creation.

To communicate with this Franken-disk, she did not use a standard DOS computer. A standard PC controller is designed to reject bad data. It sees noise and simply reports a “General Failure.” Instead, Kalinina used a Greaseweazle, an open-source USB interface that allows a modern computer to control a floppy drive at the raw flux level. It simply reports the magnetic transitions it sees on the disk, no matter how chaotic they might be.

The first results were physically catastrophic. When she engaged the drive, the head loaded onto the paper surface and immediately began to scrape. The black oxide powder, held only by the weak adhesive of the sticky note, flew off the disk. This phenomenon is usually the death knell for old tapes.

The powder had coated the delicate read/write head of the drive, blinding it. A magnetic drive head works by detecting microscopic changes in magnetic polarity. If the head is covered in a layer of loose rust dust, the gap between the sensor and the disk increases, and the signal vanishes.

Despite the mess, there was a glimmer of hope. Before the drive choked on dust, the Greaseweazle detected magnetic activity. The disk could be erased, and it could hold a charge. The physics were sound, but the engineering was fatally flawed. The problem was not the magnetism. The problem was the binder.

In industrial floppy disk manufacturing, the magnetic oxide is not just dusted onto the plastic. It is mixed into a complex slurry containing a polymer binder, lubricants, and solvents. This slurry acts like a paint that hardens into a durable, smooth surface. It traps the iron particles so they cannot move physically, even while their magnetic orientation flips.

Kalinina needed a binder. For her second prototype, she looked around her home for a solution and settled on nitrocellulose. To the layperson, this is nail polish. To a chemist, it is a film-forming polymer that has been used in everything from guitar finishes to early motion picture film.

She diluted the nail polish and applied it with an airbrush, aiming for a layer less than one micrometer thick. The logic was sound, but the result was a failure. As the nitrocellulose dried, it shrank, causing the paper disk to curl and crack. Worse, the layer was too thick.

Magnetic storage relies on the “head gap,” which is the distance between the recording head and the magnetic particles. This gap must be infinitesimally small. The layer of nail polish acted like a brick wall. The drive spun the disk without shedding, but it read absolutely zero signal. The protective coating was effectively a magnetic insulator.

Prototypes 1 and 2 proved that she needed a surface that was stable enough not to shed, but thin enough to allow magnetic flux to pass through. For Prototype 3, Kalinina abandoned the paint and returned to mechanics.

She applied the black oxide powder to the sticky label again, but this time she used a roller to physically crush the particles into the adhesive. She scrubbed the surface aggressively to remove any loose material. Then came the secret weapon: silicone lubricant.

Friction is the enemy of data. A floppy disk spins at 300 revolutions per minute inside a plastic jacket. If the friction is too high, the disk drags, the speed fluctuates, and the data becomes gibberish. Kalinina applied a thin layer of silicone to the disk to help it glide.

Initially, it failed. The disk was too slippery and too thick with goop. But then she wiped it down with isopropyl alcohol, removing almost all the visible lubricant. What remained was a microscopic film. This was the “Goldilocks” zone. The film was thin enough to let the magnetic head get close to the iron, but slick enough to prevent the head from ripping the powder off the paper.

She spun up the drive and attempted to write a pattern.

On the screen, through the visualizer of the HxC floppy emulator software, a pattern emerged from the static. Buried in the chaos was a repeating ASCII string saying: “testtesttest.” The drive was now seemingly reading data from a piece of paper coated in toner powder.

Having proved the mechanical concept, Kalinina turned her attention to the chemistry. Her previous prototypes used Fe3O4 (Magnetite), which is black. However, the gold standard for magnetic storage in the 1970s and 80s was Gamma-Ferric Oxide, or gamma-Fe2O3. This compound has higher coercivity, meaning it holds a magnetic charge more strongly and reliably.

You cannot easily buy Gamma-Ferric Oxide at a craft store. But Kalinina found a research paper from Cardiff University (The Triopolitan has been unable to source said paper, not for a lack of trying.) suggesting a method to synthesize it. The paper allegedly claimed that by heating standard Magnetite to a temperature between 200°C and 300°C in an oxygen-rich environment, it would oxidize into the desired Gamma state.

Kalinina became a kitchen alchemist. She sprinkled her black toner powder onto a baking tray and placed it in her oven at 250°C. An hour later, the transformation was visible. The powder had shifted from a deep jet black to a rich, earthy brown (the exact color of a vintage floppy disk.) She had successfully synthesized magnetic recording medium where she likely cooked her dinner.

Now armed with her homemade “premium” rust, she began work on Prototypes 4 and 5. The process had become refined by this point, and the performance jump was immediate. Unlike the black powder disks, the new brown powder disk hit 50% signal strength.

Looking at the raw flux data, the success was undeniable. She was recording roughly 300,000 flux transitions per track, matching the density of a commercial 5.5″ floppy disk – she even tested the disk without its protective square sleeve (the “cookie” only). It spun, and it read. This success was matched by her attempts to make audio tape using a similar method, applying oxide to strips of Scotch tape to run through a reel-to-reel player.

This experiment yielded a different, sadder lesson. She had created a working audio tape using cobalt-doped magnetite, but she made the mistake of spooling it tightly while the adhesive was still active. Over a few days, the “wet” adhesive of the Scotch tape fused together, resulting in one very large block of sticky magnetite.

The saga of Nina Kalinina’s DIY floppy disk is a triumph of curiosity. She proved that the barrier to entry for creating digital storage is surprisingly low. With a sticker and some rust, you can store a bit. Unfortunately, it seems we still have to wait for a more quality controlled method to arrive to abate the predicted floppy disk apocalypse, but it is certainly a great example that magnetic media revivals are indeed something not completely out there, as the efforts to restore Polaroid production have shown. Perhaps, one day, we will get home-cooked magnetic disks drives in your local store.