Ricochet is the best place on the internet to discuss the issues of the day, either through commenting on posts or writing your own for our active and dynamic community in a fully moderated environment. In addition, the Ricochet Audio Network offers over 50 original podcasts with new episodes released every day.
How to Build a Computer 24: Reading and Writing
Last time we discussed electricity, magnetism, and how you can generate a magnetic field with an electric current. You know what? Let’s jump straight to the kielbasa:
You could start with a salad, but why bother?
If you run a current through the wire you generate a magnetic field in the sausage (which ought to be a magnetizable metal, naturally), and a field between the prongs. Okay, we can use that thing to make a magnetic field, and use it to write fixed magnetic spots to a disk. The question we left off with is ‘how do you read it?’ Actually, I skipped a step. There’s still one important distinction to be drawn in how you write. Why does the magnetic field still write things if the platter isn’t in between those prongs?
You’ll have to forgive the crudity of the illustration, but most read/write heads aren’t actually made of kielbasa. I’ll give you another illustration that illuminates the point a little better. In this diagram, the blocks on the left and right are your read/write head (or the prongs on the kielbasa part. Called a ‘yoke’, incidentally). The little red lines are magnetic field lines.
I could draw diagrams for everything but they wouldn’t be as greasy. Then where would we be?
In the middle of the gap, you’ve got straight lines. And if I had my say in matters that’s where you’d stay; the math is much easier to solve when you don’t have to worry about edges and whatnot. But without actually looking at the math, see how it starts to curve near the edge, and how there’s a long fringe to the field? The fringe on the outside has much less strength than the magnetic field in the deep gap, but that fringe exists, and that’s the part that does the reading and writing. Here, let me draw the hard drive platter in.
Here, let me add a grey block to the previous picture and pretend that I’m doing something complicated.
You put a strong current in the wire around the yoke, it produces a strong magnetic field in the gap and a weaker magnetic field in the fringe. If it’s strong enough you’ll write to the platter. Okay, but how do you read from the platter? That was the question I asked at the top. Let me draw in the magnetic field lines from the written cells on the platter.
I’m going to level with you; I’m drawing diagrams out because by this point I’ve already eaten the kielbasa. And the spaghetti.
Blue lines represent the boundaries between cells. The field generated by the permanent magnets within the cell curves around from north to south. You can see the arrows poorly drawn onto the green field lines. I probably should have put some arrows on the red lines too, but it’s already hard enough to read. You’ve got a constant current running through your spaghetti, so the red lines don’t change. Where the red lines intersect with the green lines though, that’s where you get electricity.
That is, you get it if the read/write head is moving across the platter. The red lines don’t change shape at all, but they’re moving with respect to the green lines. Since the red lines are moving they’ll cross the boundary from one cell to the next, in which case you get a sharp change in the number of lines crossing one another.
Well, you would if the lines were a real thing; remember they’re representations of the magnetic field. Really we’re talking about the flux density (more technobabble that actually means something!) varying with respect to position, as well as the direction vector changing. Sometimes I like to remind myself what the approximations I’m using are concealing. As a rule, it’s much easier to think about these things in terms of the lines.
So what do your red lines see? (Well not “see” per se… wait, didn’t we just do this?) The direction of the green field is changing from side-to-side in the middle to up-and-down at the edge. The flat middle portions induce relatively little voltage in your wire. The high degree of change as you get to a cell boundary registers as a spike in the voltage. Measure the voltage and you can see what the underlying magnet looks like, that is, you can read the information off of the disk.
Okay? Let’s recalibrate our perspective for a minute. What do you care about in your hard drive? That’s right; how much data it can store. That’s dependent on how small you can make those spaces. Your credit card’s magnetic strip holds some information in it. Intuitively, if you bring your credit card next to a strong magnet then the strip goes from recording your card number to storing exactly one piece of information; a one or a zero (depending on how you were holding the magnet). Going in the opposite direction you can store more data on your disk the smaller you make your spaces. As you might imagine that’s easier if you’re not waving giant magnets at it. The word for that, by the way, is “areal density.” That is, how much information you can store on a given square inch of drive. To get a higher areal density you’ve got to decrease the space it takes to store your information.
To make your footprint smaller you’ve got a couple options. You can make your gap smaller between the sides of the yoke, which gives you a stronger magnetic field. You can hover the thing closer to the platter, which lets you use a more dense part of the field to do your writing. You can make your recording layer on your disk thinner, which actually works much like bringing your head closer. You can change the material in your magnet; sputtered layers of permalloy (81% nickel 19% iron) work better than a straight iron core. You can fill the gap with a nonmagnetic material. And you can do all of that at the same time.
That’ll get you about to the disks we were using in 1990. To get further you’ve also got to change the magnetic properties of the material you’re recording on. For all that excitement join us fortnight next to get to what I thought we’d get to today for “The futility of magnetoresistance” or “Seriously, this guy is long-winded. How long can he drag this out?”
This is part twenty-four of my ongoing series on building a computer, the hung-over and strung-out way. You may find previous parts under the tag How to Build a Computer. This week’s post has been brought to you by Red Beer. Don’t be a dummy. Red Beer!
[First – Silicon] [Previous – Magnets] [Next — Magnetoresistance]
Published in Science & Technology
MOAR CHEEZ!
Flash is actually a lot simpler, mechanically. Just harder to actually do in real life.
They should be.
Fortnight next. I like that.
Next time, Gadget! Next time!
Floppy disks had a tiny hole in the mag surface (not the “don’t write” slot) that I’m guessing had something to do with synchronization. What does the hard drive use to orient itself as to where the head is writing? In television we use voltage to distinguish between frames of a moving image. Low voltage is picture “black”, but really really low voltage means “jump back up to the top and start again”. But since the hard disk signal is binary, you can’t use voltage. Is one of the concentric tracks a control track?
Also, how wide are the tracks? What is the precision required of the servomechanism to position the head correctly?
From “Basics of Digital Computers”, 1958. When I was ten (in 1962) this is the first book I ever bought with my own money. Saved up for quite a while for it; it cost $2.80
Servo positioning is checked by specific bits at specific points in the hard drive. At times past that was a specific track, yeah, or sometimes even a whole platter surface (if you have four platters you can dedicate one surface without losing too much space. You keep your read/write heads fixed to one another so if one knows where it is, they all do.) These days there are specific bits embedded in each track that contain that information.
I don’t know how wide tracks are on current hard drives. I’ll have to look that up.
I think it was 40 tracks on the old floppies, meaning that including the space between, they were close to a millimeter wide.
@Hank Rhody, Acting on Emotion:
Nice explanation of the recording process, HRAoE. I actually know how the sausage is made as I’ve worked in hard drive development for the last 30 years.
The read-back process which you describe very well here is known as “reciprocity” — the magnetic flux pickup from external field sources has the same shape as the magnetic flux produced by the head when current flow through the coils. This read-back method became unusable in the mid-90s, because we needed too many strands of spaghetti to get a decent signal. At that point, magneto-resistance was not futile.
Present day hard drives have track densities of around 400,000 tracks per inch. Each track is about 300 atoms wide.
Appreciate it. I work in a factory that makes stuff a little bit up the line. This stuff I’ve been having to read up on.
Utile?