The use of magnetic tape and other removable magnetic media is now on the wane, leading to scarcity in some cases where manufacture has ceased. Is it possible to produce new magnetic tape if you don’t happen to own a tape factory? [Nina Kallnina] took the effort to find out.
It’s probably one of those pieces of common knowledge, that magnetic media use iron oxides on their surface, which is the same as rust. But the reality is somewhat more complex, as there is more than one iron oxide. We follow [Nina] through this voyage of discovery in a Mastodon thread, as she tries first iron filings, the rust, and finally pure samples of the two iron oxides Fe3O4 and Fe2O3. She eventually achieves a working tape with a mixture of Fe2O3 and iron powder, though its performance doesn’t match manufactured tape. It turns out that there are two allotropes of Fe2O3, and she leaves us as she’s trying to make the one with better magnetic properties.
These results look promising, and while there is evidently a very long way to go before a home-made magnetic coating could replicate the exacting demands of for example a hard drive platter it’s evident that there is something in pursuing this path.
This may be the first time we’ve seen tape manufacture, but we’ve certainly seen extreme measures taken to rejuvenate old tapes.
I distinctly remember James Burke’s voice “This is recorded in sticky tape and rust” from secret life of machines.
That would be tim hunkin.
You can watch that on Tim’s YouTube channel.
I went to find the episode, and it’s actually Rex Garrod’s voice that they record:
https://youtu.be/g1JlUcFKm5o?si=tSJbYkIE6V__oB2x&t=600
Thanks for replies! At least I remembered a couple details right from decades ago. Those shows were pivotal to making me who I am today.
James Burke was Connections (season 1 was in 1978)
Interesting idea, I always assumed the best way to make magnetic tapes was vacuum deposition and wouldn’t have thought of just adhesive tape as a base material. Just seems like a recipe for lots of metal dust getting everywhere as it rubs off the tape. Rather impressive results though, and yet another fun idea I want to play with now. I’ve idly wanted to create a reel to reel backup tape drive as a really cool addition to another project I’ve still not finished yet for a while, now I might just have to try it.
Like the old joke says…”Turn it over”. Run it with the coated side away from the head.
If you try to make your own hard drive, I would imagine that more difficult than the coating would be getting the platter optically smooth and extracting all dust from the environment.
Amateurs regularly make excellent mirrors for telescopes. The techniques are well known and fairly easy, albeit tedious.
Walking around a barrel?
I can’t claim this as absolute fact, but I seem to recall reading someplace that the original “winchester” platters were coated by pouring a magnetic media/binder slurry over the blank platters as they were gently spun, resulting in a uniform coating. The platters were then baked.
Maybe someone else can confirm.
Look on YouTube for the film ‘IBM magnetic disk pack production ‘ from 1968.
It shows the magnetic slurry being poured on in a machine that also bakes the platter.
Another difficulty will be to achieve a “good” / “satisfying” storage density.
Let me explain, the storage area of a hard disk is close to an A4 sheet of paper (~0.06m²), for up to 20TB of data. A tape is 800 meters length, by 1.3 cm wide, which makes a bit more than 10m². With current tapes (LTO9) we can store 18 TB of data (uncompressed) (IBM TS1160 can deal with 20 TB tapes).
That makes hard drive 160 more dense than tapes. And the difference will increase (IBM demoed a 580TB few years ago: https://blocksandfiles.com/2020/12/16/ibm-and-fujifilm-580tb-tape-capacity-record/ ).
Personally, I’m more worried about reliability here. The higher the integration density, the lower the tolerances. Tape storage is traditionally being used for backups and long-term storage. HDDs aren’t quite up to that task. Also, capacity isn’t so much of a concern for tape storage. I mean, it is at some point, but the archival sites are usually very big, anyway.
The storage density is rather more defined by the read/write head than the medium isn’t it? While there has to be a limit to how small a spot you can magnetise stably and not overwrite the neighbours, and so how dense the data can be I don’t think any tape drive is even pushing close to the theoretical limit of the tape itself.
You are quite right. The real challenge with tape drives is to make the heads follow the tape. A tape such as the one we use fly over the heads at something like 5 m/s (18 km/h or 11 miles per hour), and the tape wobbles. So the head must moves up and down to follow the tape movements (with adaptative and predictive behaviour). I had the chance to meet the IBM guys that work on these problems at the Zürich research centre, and they explained well this.
On the hard drives side, the problems are more challenging to enhance storage and requires complicated processes that increase the drive price or lower the endurance of such drives (see Heat Assisted Magnetic Recording (HAMR) or Microwave-assisted magnetic recording (MAMR)).
Where HDDs size do not follow a Moore Law anymore, the tape volume still increase in a quite linear manner.
The tape drives we use in our data centre can deliver 400 MB/s (read / write). So with 10 drives in parallel, we achieve a sustained 3 GB/s for data movements (sequential R/W of course ;).
Probably make a better tape using chromium and neodymium instead of plain ol iron
How easy is it to use vapor/ion/electrostatic deposition to form a nice magnetic surface so smooth you can see your face when you look at it…..and maintain a precise average thickness across the entire surface, could make your own floppy disk that way, that can actually hold over 1.44 meg, depending how good the r/w head is in the drive
My 90’s thinkpad could already hold 2.88MB
You could already do 32MB with regular old 1.44MB floppies; old LS-240 drives could write to regular floppies with 32MB capacity (obviously only readable by those drives) tho.
I only skimmed through the article. But Don’t forget to include some sort of surface lubricant. No idea of what it needs to be made of though.
I think without that, you’ve largely made the equivalent of Crocus cloth strips.
Sony Metal SR Master.
I remember reading about an early video recording setup at the BBC that used steel tape, and because it wasn’t helical scan it whipped the tape through so fast it was extremely dangerous to be around it when running.
The end-of-article comment that resonated with me most is “You’re nuts Nina, I love it.” Not quite as challenging as semiconductors, thin film emulsions are super hard to manufacture with consistency. But okay, if this wasn’t hard enough, try making photographic film. At least magnetic tape doesn’t have to be manufactured in total darkness. In the waning days of Kodak, they gave a tour of a film factory to [clears throat] a photojournalist, with the lights on… [one eyebrow raised] Fascinating! Explanation: every once in a while, they need(ed) to clean and do regular factory maintenance. “Please don’t take my Kodachrome away” (Paul Simon)
More recently, Kodak gave a lights-on factory tour to Destin from the Smarter Every Day youtube channel. It’s a great series.
Very impressive work. I’m surprised she didn’t also try metal foil as used in the first Univacs.
Earliest not-yet-tape recorders used steel wire. That could be easier as it is already off the shelf.
Or just chop up plain steel wool with scissors and let that rust, surely less effort and more consistent than scraping rust off a pan.
Paging the bunny electric man!
A lot of these coatings sound pretty abrasive, possibly leading to excessive head wear; if you think magnetic tape is hard to make, try fabricating recording/playback heads.
How were floppy disks manufactured? I would love to be able to make new 3″ disks, for Amstrad CPC…
If I was making tape or disks or whatever form of magnetic medium tickles your fancy, I would skip the “making iron oxide” part. Very fine iron oxides—red, black, yellow, brown—are readily available and very cheap; shipping will cost more than the material itself. Other materials such as oxides of chromium, manganese, nickel, cobalt, and various powdered metals are also available and cheap. Try fireworks-making and ceramics-making supply houses.
I am reminded of the guys in the rocketry community who want to make a liquid-propellant rocket engine, using hydrogen peroxide as the oxidizer. Why H2O2? Because they can make and concentrate it at home. It’s always pointed out that they’ll not really be working on making a rocket engine. They’ll be working at manufacturing high-quality H2O2, as that will consume almost all of the time spent on the project.
Likewise, if the objective is to make magnetic media, and if cheap, readily available starting materials can be purchased, it makes sense to start with such materials. Spend the time and effort on the main quest—making useful magnetic media— and don’t get caught up in a time-consuming side quest.
I love this so much. I hope they continue. I’d love to see someone make drum memory.