That's a very interesting servo arrangement. Linear voice-coil motors were pretty standard for "big" drives of that era, but it looks like two servo motors 180° opposed, presumably with two separate sets of heads. I can't recall ever seeing a hard disk like that before.
Any idea what the back-story is on that? Upper heads on one side, lower heads on the other? Two heads per surface for redundancy / speed / double the probability of a head crash?
Looks like a model 3380 or 3370, but I still can't find any specs.
I was working on CAD back in the early 80's. Hard drives were multiple platter things, about the size of a standard LP record, with 8-12 of them stacked up. They were insanely expensive. Damage to the platters when changing out the stack was a distinct possibility.
At one job in the DC area, we found that the drives were getting full (80% of capacity or so) and starting to have sector errors because of it. Nowdays one small HD (well under 500GB) would have handled the entire company's needs with plenty of rooom to spare. We had three stacks at that particular company, and that was considered excessive at the time. 3.5" floppies were a marvel of modern technology.
Seeing that it wrote twice the amount of data of the 3350, i'm guessing they are reading and writing at different sectors.
I'd still love to know details as to why (and how) they split the headstack. No reason you couldn't achieve the same density with only one armature. I can only imagine that maybe they were able to double the read performance by being able to have two different heads active at the same time, servicing two different requests. (Imagine taking apart two hard drives, and installing the entire head / electronics / controller package (except for the spindle drive) of one onto the other.
Just a guess.
Originally Posted by Full_Tilt_Boogie
To think now a SSD is ~1 $/GB
Less, actually. I just bought a new 120GB Kingston for $99, with a $39 rebate ($60 final price.)
I just wish that, leveredging today's cost / density factors, that somebody would make a small SLC drive and sell it for what an equivalent MLC drive cost 3 years ago.
Originally Posted by Pen2_the_penguin
cant imagine the ear piercing sound those hard drives wizzed at, probably sounded like a turbocharger sucking in random BBs
Honestly, they were not nearly as loud as you'd think. Moving the heads was pretty noticeable, but the spindle itself was damn near silent. Remember that these drives turned very slowly (3,600 RPM shaft speed was typical for 14" drives) and they were built with spacecraft precision. They were rugged, and yet exquisitely well made. Thy were also belt-drive, rather than direct drive.
Originally Posted by rleete
I was working on CAD back in the early 80's. Hard drives were multiple platter things, about the size of a standard LP record, with 8-12 of them stacked up. They were insanely expensive. Damage to the platters when changing out the stack was a distinct possibility. .
Reminds me of the "big" computer at UF when I was in college. It was a VAX 11/780 (actually a pair of them in a cluster configuration) and their primary storage was a row of RA-81 drive cabinets placed prominently in front of the machine room. 10 or so of these lines up side-by-side:
Each one of those cabinets houses three drive assemblies, roughly 450MB each. If I recall, a couple of them actually had RA-60s in the uppermost slot, which were the removeable cakepack drives you mention:
I have utterly no idea what the '60s were used for- I don't remember them being accessible from my account. Having all that on-line fixed storage was just incredible. Even though they still had the 9-track tape drives, I don't recall ever needing to request a tape-mount. Everything was right there on the hard drives, online 100% of the time.
RAMAC, the first commercial hard drive. Officially the IBM model 350, and forming the RAMAC 305 when paired with its matching CPU. This bad boy had fifty 24" platters turning at 1,200 RPM, and yet only a single pair of heads!
Unlike modern hard drives, with one (or two) heads permanently positioned on each surface, the RAMAC functioned a bit like a 1950s era jukebox. The single pair of heads was mounted on an elevator, and moved up and down to select a specific disk. The disks were kept rotating constantly, and the heads were loaded and unloaded dynamically as needed. You can see the head elevator to the left of the disk pack here:
This gave the RAMAC a very interesting metric. Unlike modern drives (everything from the 60s onward) in which seek times are measured only from track to track, the RAMAC also had to seek from disk to disk. Its capacity was the equivalent of 5 megabytes. According to legend, engineering felt that they had the capability to double this shortly after release, however the sales department didn't think they'd be able to convince any customers of the necessity to pay for a 10 megabyte drive, and the upgrade project was thus shelved.