• Updated 2023-07-12: Hello, Guest! Welcome back, and be sure to check out this follow-up post about our outage a week or so ago.

Hard Drive help!

macclassic

Well-known member
Sound of bells tolling slowly.

"My 6100/60's hard drive has a problem Huston" The drive starts to spin up on startup but then spins down again and I get the floppy with a ? on the monitor.

I tried swapping the power supply but it made no difference.

Has anyone had a similar experience with a 500mb 50pin SCSI drive 'and fixed it'!!!

It's obviously a physical problem so I'm after any tips about HD's, pinouts, motors, circuits, whatever.

I am willing to "boldly go" on this one as I'm used to fixing things, from gas regulators to chronometric rev counters.

 
Last edited by a moderator:

macclassic

Well-known member
It will start up with an OS on a floppy but I can't find my HD from the floppy.

I'm wondering if it could be a dry solder joint somewhere on the drives board?

 

wally

Well-known member
I have several untrustworthy 4 GB 50 pin SCSI external drives (the data is backed up!) that have similar unresolved symptoms if stored for a while. After some visual inspection yielded nothing I did some simple experimentation including power supply substitution and found that repeated power up cycles eventually would spin the disk up. Initially the spin up would terminate quickly. Subsequent power ups yielded slightly longer and longer spin up periods and higher rpms reached before drop out. Snap rotating the drive to relieve any stiction tendency also seems to help sometimes. Finally after enough of these prematurely terminated spin ups ( maybe 4 to 8 ) the disk would reach final speed and stay up. After leaving the disk on for a while, then turning it off, the next turn on cycle would come up to speed in a short time, more rapidly than any of the previous attempts. I haven't ruled out congealed ball bearing grease, bad electronics, electrolytics or a cold solder joint, but my two units act like the motor drive is not quite up to overcoming some friction source before some timeout or cutout occurs. When the disk comes up to speed and runs for tens of minutes, it is good for subsequent fast spin ups for many days thereafter. And once running, I have yet to see the disk drop out suddenly, as might happen if a solder joint was bad or connection intermittent... [?]

 

macclassic

Well-known member
Thank's Wally,

That's just the kind of info I was after (that gained from experience).

I will try the repeated power up cycle technique now!

 

macclassic

Well-known member
Tried the repeated power up cycle technique, no luck - not yet anyway.

More detailed account, the drive starts to spin at one speed for a few seconds and then makes a noise before "I think" attempting to go faster but it doesn't sustain this second speed or phase and slows to a stop, so I wonder if it's the switching done by the board which has gone bad.

I also had the drive out, took the circuit board off, looked all over it for a bad joint, no luck.

 

wally

Well-known member
If you like to really dig in, Google search for 3 phase brushless DC motor controller driver, and look at the Allegro and On Semiconductor (and your vendor, if you can identify) datasheets and application notes for examples of circuit operation. You will see that most of the chips require a low value external current sense resistor that determines the current limit for the high current needed to deliver startup torque. You can attempt to locate this resistor and verify it is intact, soldered well, and that the phase drive and hall effect sensors are connected well (might be good to unplug/replug the motor to controller board connections). You can also monitor the 5V and especially the 12V power to make sure they are not current limiting/folding back under the startup load, or substitute a really formidable power supply. No computer is needed, the question is whether the drive will just spin up, and most will try without any SCSI bus connection.

 

tomlee59

Well-known member
Be aware that the firmware of many drives terminates spin-up if some serious fault is encountered (for example, if it is unable to move the head to track 0). If it doesn't pass the basic self-diagnostics, it figures that there is no point in spinning up all the way. Seagate drives seem to have particularly fussy power-on diagnostics, but other makes behave similarly.

If it is just a case of stiction, then you may wish to find the current-limit resistor Wally mentioned, and reduce it in value. This hack was useful in getting some old 40MB and 80MB Quantum drives to work much longer than they had a right to. I never did any detailed experiments, but cutting the resistor value in half (by paralleling it with another of the same value) seemed to work ok.

 

macclassic

Well-known member
Is this the said informaton?

http://www.allegromicro.com/en/Products/Part_Numbers/8904/8904.pdf

http://www.onsemi.com/

The trouble is one of my favourite films is The Flight of the Pheonix in which a disparate group of men make a flyable plane from the wreckage of their crashed plane in the middle of a desert and an old hand played by James Stewart clears the engine's cylinders of fuel with one of the two cartridges left to start the engine, which works, as the engine starts on the last cartdrige.

***

 

wally

Well-known member
So you wanna fly out of the desert on a modified hard drive? Only in Hollywood!

Those indeed are two vendors of motor control chips, but there are many others and many variants. The newer ones may have internal current sense resistors and computer programmable current limits. I must point out that if you are crazy enough to try component level repair on an obsolete drive without schematics or code listings, you might consider coming to Silicon Valley where they pay you to do that kind of crazy, and you have access to both the schematics, PC board artwork/netlist, and controller code, and also more test equipment than you can fully understand in your life.

If you like, try visiting a great reverse engineering assistance site, http://www.plasma-online.de/index.html?content=http%3A//www.plasma-online.de/english/identify/picture/index_logochip.html

and see if you can identify the manufacturers of chips that you think could be the brushless spindle motor controller/driver and the voice coil driver from the chip logos. Then by reading the datasheet and application notes, you might get some ideas on overriding normal operation and manually ramping up the spindle rpms. On my 1993 vintage Conner 540 MB drive for example there is a row of SMT power transistors, and two nearby Silicon Systems LSI chips that are the likely suspects.

As a wild guess on non-electronic causes, I think over time the degraded lubricant along with abraded media from the head parking zone could plate out on the spindle bearing races and gradually contaminate them. It's a uncertain race between the bearings getting tighter from deposits and abrasion loosening the bearings with wear particles getting pushed to the side by the balls. The guys that know which happens first are not talking, and perhaps this is the kind of thing that varies greatly on whether the power is left on all the time, or if there are many power up/down cycles with consequent head contact in the parking zone.

 
Top