Disk geometries

Some drives had hidden partitions and operating systems loaded onto them that ran on the drive itself. CP/M was common on ESDI drives.

I went from 750GB to 1TB to 1.5YB to 2TB, so the jumps are getting bigger.

Right about the time they start to top out the 3.5" form factor, they drop the whole line and have us all using 2.5" drives and SAS interfaces.

Perhaps you merely remember the old early MFM trick of setting the drive interleave right.

Early drives had some hard sector flags in them and you were not able to do what you describe... at all.

Though you may also be remembering early IDE drives and virtual cylinders, etc.

Kind of like the guy telling Vizzini that the word "inconceivable" was not a word.

You gained no additional drive space, unless you were performing the first mentioned method.

Overkill in every instance.

Seagate drives come thoroughly tested from the factory. I have 50 of them, so I should know.

Partition, format and use... forever.

Besides, the sectors get examined at format time, and while in use with S.M.A.R.T. on.

That was not the only thing that limited them. Duh!

Isn't that called "hard sector flags"? Used as markers for the servo smarts to calibrate 'as formatted' cylinder track center points and sector start points against.

CD players??!!!

You are an IDIOT. The argument was about HARD DRIVES!

CD players wear out quick, because they want you back buying the next, newer model each year or so.

The fix is to invert them, because they have a main optical lens that sags and loses its calibration window. Inversion sags it back the other way. Known fact. Known for years. Get back on the topic of the branch.

No, to both. The argument was about epistemology. Hard drives don't use 'absolute' track positioning any more, they are servo-driven to find and hold tracking, BUT we don't have easy access to that tracking system. The usual CD player layout (of yesteryear, at least) included convenient test points where one could monitor the error feedback loop to verify that the compensation was occurring. And that compensation signal told us the size and nature of the errors.

The tracking errors, either due to thermal expansion, position changes, accelerations, or spindle motion tolerances, are compensated successfully when a hard drive operates, and the only way you know about those errors is when they get too big and the drive fails.

I figure it is a simple commonality in disk drive production equipment. I'll bet there is no more than two manufacturers.

Id bet not. A simple glance inside two different drives makes that glaringly apparent.

This is aside from the fact that there is no specialized gear needed. All of the processes are normal, even the mag particle coating on the platters.

I think it's mostly an issue of "what the customers want" - which includes market predictability.

Consider that a high percentage of 2.5" and 3.5" disk drives are sold to big system-building companies (Dell, HP, etc.). These companies want to be able to build, ship, and support systems over a period of time. It's in their best interest to be able to maintain a stock of (e.g.) "120-gigabyte" drives, mastered with a specific set of software (e.g. Windows-de-jure, certain applications and drivers, etc.).

If the drive manufacturers suddenly stop shipping "120-gigabyte" drives, and start shipping "130-gigabyte" drives, the system builders may have to go through a relatively expensive process of compensating for this. They'll have to generate a new set of disk masters in their factory. They may have to change the packaging and advertising of their desktop or laptop systems to reflect the increased disk size. They may not be able to perform "exact" warranty repair or replacement on their systems.

The same consideration applies to the manufacturers' sales people, and to wholesalers and retailers. They can no longer maintain a stock of drives of predictable sizes, and there could be an increasing number of stock SKUs which have to be handled throughout the sales chain.

The preferred business model seems to be for the drive manufacturers to pick a fairly limited number of capacity-points, and ship drives which are configured to meet these. The vendors will end-of-life older models, and substitute newer-technology models configured for the same capacities (I think the average market lifetime of consumer-grade hard drive models is somewhere between three and six months these days). Over time, as capacities grow, the smaller- capacity niches are allowed to go empty (who makes a 40-gig 3.5" drive these days?) and increased-capacity drives are introduced with relatively large size increments (25%-50% is fairly common, I think).

One new computer I bought was advertised as having a 10 GB hard drive. There was a nice sticker on the shipping carton that said, "Free upgrade to a 20 GB hard drive" The stickers probably cost 10 cents per computer.

You can buy new 40 GB solid state drives.

I'm not sure about that. I.e., that would lock them into

Regardless, it doesn't explain why the "increments" are what they are *when* they do come along.

E.g., you can understand why RAM increases by multiples of

1. You would *never* expect to see a 5x increase! OTOH, the geometries may shrink by a factor of 5.10367 (or any other insane figure).

It just seems like too many *seemingly* arbitrary constraints are shared among all drive manufacturers. E.g., why don't we see more variation in platter speeds? Why don't we see

It's like deciding all tires should be inflated to 30psi (not 29 or 31...)

But that's true of everything. E.g., you don't find folks making

I just find it puzzling that there *appears* to be so little variation between the big players where there seem like there

Plot them on a graph (semi-log) and the slope of the line will be relatively constant. Incremental improvement, but everyone is constrained by the same physics and economics.

Cost doesn't follow the 4x quantum. A smaller than 2x feature size jump shows up in cost. Plot cost per bit on semi-log paper.

Platter speed and (linear) bit density are inversely proportional. You have high density and high speed market points (and turtle slow and cheap) in the mainstream. More differentiation isn't deemed important.

Because of the physical limitations. You have to fit into the 2.5" or 3.5" package so an odd sized platter has no advantage and loses capacity.

For good reason. The market doesn't want small variations if it sacrifices

If all cars were exactly the same size and weight, they probably would be.

The knobs are all interrelated. Tweak one and they all move. The market has decided on some standards (7200/5400RPM, 2.5"/3.5") and the technology curve does the rest.

software

their

replacement

relatively

same

shows

have

the

sacrifices

be.

and

drive

market has

curve

Say, Keith were you watching the computer technology changes back as far as the early 18" 1800 rpm drives. Did you get to notice the 14" drives come out (Whitney technology replacing Manchester tech)? What was the next platter diameter and when did the spindle speed step up to 3600 rpm?

It was WINCHESTER, dingledorf.

Before my time, but the first disk drive was the RAMAC(1956) with 24" platters. The 3330 was codenamed "Merlin" and the 3340 (1973) was the "Winchester", of which just about all drives today are descendants (fixed disk

- "sealed" HDA). We had tons of 3330s but 3340s sold too well. ;-)

There were a pile of special-purpose disk drives in the '70s, (Zeus/2305 was a fixed-head paging store) but I don't remember them all.

I don't remember any of them being >1800RPM until the 5-1/4" drives came out.

This came up on alt.folklore.computers just last month.

Little more than halfway down the page. I wonder what a head crash would sound like? :)

More words and pictures:

Fifty .30cal rifles? ;-)