• Please review our updated Terms and Rules here

Computer Memories Inc hard disk collection (Incomplete)

Thanks for posting those schematics, they will be helpful in my CM-6426S project.

My drive spins up, and appears to pass whatever power on self testing it does. The drive is ready, but when I try to low level format it, it completes cylinder 0, makes a very quiet click, then stops responding. As an experiment, I taped over the step line on the edge connector, and it happily reformatted cylinder 0 640 times without an error. At least this tells me all 4 heads are working.

Next, I tried momentarily grounding the step line while the drive select line was also grounded. I got the same quiet click, and the drive again stopped responding. Clearly it's not happy about something when it tries to step to cylinder 1. I will be examining the head servo circuit more closely and post back soon.
 
Since these are just electric motors, there is every possibility that it is just jammed or sticky. That, and I hear the glass wheel inside the optical encoder can break during rough transit if you're unlucky. Let us know how it goes.
 
The head actuator moves freely. It does a fairly long seek test on startup, which seems to pass, and sounds like the drive in the MP3. With the cover to the stepper arm removed, I can see that the click it makes when it tries to step is actually the head lock engaging. Clearly, it's upset about something, and locks the heads for safety. My drive has board number 100079; hopefully it's not too different from 100078.
 
"Fairly long" is not a good sign. These drives give up completely after I believe 3 to 5 cycles looking for track zero. You should not hear more than the initial two long, slow seeks on spinup. Any more is usually indication that the drive is having difficulty locating the embedded track zero, and while eventually it gives up entirely and proceeds to complete the rest of it's test, it will usually not respond to many commands. The head lock engaging would also be a sign of this, or spindle speed instability.
 
I'm not sure how long is too long, it takes about 7-8 seconds to do the seek test. It seems to successfully find track 0, and the spindle it on speed. There appear to be several blank (apart from the servo burst) tracks between track 0 and the rest of the recorded data tracks on the disk.

The head actuator seems to be locking onto the servo burst and keeping it on track. It takes a fair amount of force on the stepper arm (from under the drive) to push it off track, and then it immediately returns to tack 0. As soon as a step pulse is sent to the drive, it locks the head, and disables the head servo loop.

At that point, I can manually unlock the heads, and freely move the heads to any part of the disk. The entire disk looks about the same, apart from a few tracks near track 0 which have a different looking servo burst. I assume this is how it finds track 0.

In case anyone is curious, here are a few oscilloscope screen shots showing the servo burst, and the start of the data area. The servo burst is the part visible when the upper trace (USRMSK) is high. Servo burst A, or B is visible depending on which way it's off track. While, I have noting to compare it to, it all looks reasonable.

At this point, I'm going to wait until I find another drive. It's either an electronic problem which I can fix, or a problem with the heads/disks, which I can't fix.
 

Attachments

  • ins.jpg
    ins.jpg
    184.7 KB · Views: 2
  • on tracks.jpg
    on tracks.jpg
    181.8 KB · Views: 2
  • outs.jpg
    outs.jpg
    191 KB · Views: 2
Sorry to hear that your drive has an electronic problem. Other than the servo bursts and embedded track 0 there is nothing on the disk which might cause misoperation of the drive. Asking around I have heard that the optical servo motor uses an etched glass wheel, which can sometimes crack or become dirty and cause the drive to miss seek. Maybe it is worth pulling the lid off the servo to check?

Either way, beautiful pictures of the tracking system there. Few people believe me about the burst system because Wikipedia claims these drives have no on-disk feedback - Which is not true, as evidenced here.
 
I found someone selling a broken drive with the same version PCB locally, so I decided to give it a shot. The new drive has a broken encoder wheel, so it's definitely a parts donor.

I can report partial success. The drive will now seek, and even format with no errors, but it's not working reliably. Sometimes it works great, other times it has trouble reading certain sectors, and occasionally it freaks out and spins down. The power on seek test is also hit and miss. It reliably passed the power on test with the original board, so I think the new board may have a different problem. I'll make some measurements, and see if anything stands out.
 
That sounds like a capacitor failure to me. These drives ran hot when new, and probably cooked theirs. Some of my most reliable drives would quit working seemingly at random because of leaking electrolytics, like a Lapine Titan.

It's not too expensive or difficult, most of the big ones are on the spindle control board. You might swap that whole board as well if you can. If the drive detects underspeed it will immediately cut the spindle and lock the heads.
 
That's good advice, but this drive uses only ceramic, and tantalum capacitors. I did try swapping the spindle board. I'll post back when I have time to dig into it more. Hopefully the problem will become clear after making some measurements.
 
You must log in or register to reply here.
Back
Top