Re: [midatlanticretro] Re: How can I tell if my SMD disk head is crashing?
- It wound up being an actual head crash, and once I figured out how to get the dust cover off to look at it, I could see a ring scraped into the bottom-most data surface. I wound up being able to successfully clean the head and get it functioning again.
As for the disk drive - I'll do a longer write-up on my website at some point, but a quick summary:
-The control electronics were basically shot, and I didn't have schematics for the drive, so I wound up disconnecting everything but the analog read-amplifier/head-select board. I built a robot that clamped onto the positioning voice-coil and manually stepped the read-head assembly forward (at 5200 steps/inch, vs. 400 data tracks/inch), and built a simple ADC connected to an FPGA that buffered copies of the data tracks for every head on every step.
I now have a super high-resolution 'magnetic image' of the disk, and I'm going to try to do some signal processing to see if I can recover anything useful from it.
On Tue, Aug 30, 2011 at 10:24 PM, s100doctor <hjohnson@...> wrote:> disk drive [many platters and heads]
--- In firstname.lastname@example.org, c f <christopher.h.fenton@...> wrote:
> I've been working for a while on reviving/rehabilitating an old CDC 9762if a head had made physical contact with a platter, there would be some kind of circular trail on the platter. The heads are designed to float above the platter ALMOST like an airplane wing - hydrodynamic it's called, the air is dragged under the head and lifts it up. More like a surfboard on a wave.
> When I loaded it, everything was fine for a minute or so, but then I began
> to notice a high-pitched whining noise coming out of the drive.
It's possible what you are hearing is bearing noise - some bearing at speed, having lost its lubrication decades ago, is now whining.
So the diagnostics are
1) look very very carefully at the disk pack for any circular patterns. I'd photograph the hell out of it, then compare it after some use to see if there are changes. YOu might look for a lousy signal from one head, assume that's dragging, but that's tricky.
2) see if you can spin the platter at speed without loading the heads. The bearing should not care if the heads are there, I think. But it may make some difference in loading the motor. Hard to guess...
3) find someone who ran (or runs now) these suckers, see what they remember about noise. There's something about the whine of a hard drive with noisy bearings, versus the screech of a head as it scrapes across the platter until it's flung into orbit....(sigh) those were the days...
- --- In email@example.com, c f <christopher.h.fenton@...> wrote:
>That's VERY impressive work. Any chance you can put it up on a Web site and show it? If you put a few tracks of binary data up on the Web, some number of us might advise as to how to decode it. At that point it's an exercise in software, and some people do a LOT of that sort of decoding, just for fun. Certainly, more people that those who can jump-start a hard drive as you have. Congratulations!
> -The control electronics were basically shot, and I didn't have schematics
> for the drive, so I wound up disconnecting everything but the analog
> read-amplifier/head-select board. I built a robot that clamped onto the
> positioning voice-coil and manually stepped the read-head assembly forward
> (at 5200 steps/inch, vs. 400 data tracks/inch), and built a simple ADC
> connected to an FPGA that buffered copies of the data tracks for every head
> on every step.
> I now have a super high-resolution 'magnetic image' of the disk, and I'm
> going to try to do some signal processing to see if I can recover anything
> useful from it.