ST506 HDD woes

Arc/RPCs, peripherals, RISCOS operating system & ARM kit eg GP2x, BeagleBoard
steve3000
Posts: 1692
Joined: Sun Nov 25, 2012 12:43 am

ST506 HDD woes

Postby steve3000 » Sun Aug 11, 2013 3:24 pm

Hi,
So today I decided to plug in the old ST506 drive which came with my A410. It's been fitted in the computer since I bought it, a couple of months back, but was left unplugged from the power as the previous owner said it was faulty...

I was hoping it would spring into life at least and maybe have a few disc errors to contend with, but nothing - the drive makes a few quiet clicks, but it doesn't actually spin up at all :-(

So is there anything that can be done to save this ST506 drive? Or is it best placed in the bin?

Steve

steve3000
Posts: 1692
Joined: Sun Nov 25, 2012 12:43 am

Re: ST506 HDD woes

Postby steve3000 » Sun Aug 11, 2013 4:21 pm

Ok this gets worse...

Having no luck with the A410's ST506 disc, I decided to fire up the A420, which has my only known-good ST506 disc, and this did work :-)

So I popped in my IDE podule to make a back up of the A420's ST506 disc.

All going well, then a few disc errors (13 and 16) started to appear, which was a little concerning... But most cleared on repeating the operation. As it was approaching 75% backed up, the numbers of errors started increasing - with some failing to clear on a retry... Then there was a loud 'click' and the drive spun down, followed by Disc error 08.

Thankfully 75% was backed up ok, because the drive now spins up on switch on, makes a 'click' and spins down again...

Not a good day for my computers... I'd better not touch anything else... I'm off to mow the lawn...

User avatar
1024MAK
Posts: 6679
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...

Re: ST506 HDD woes

Postby 1024MAK » Sun Aug 11, 2013 4:59 pm

Some HDD fail when they get hot.
Try waiting until everything has cooled down...

If that don't work, there is always the freezer trick...

Mark
For a "Complete BBC Games Archive" visit www.bbcmicro.co.uk NOW!
BeebWiki‬ - for answers to many questions...

steve3000
Posts: 1692
Joined: Sun Nov 25, 2012 12:43 am

Re: ST506 HDD woes

Postby steve3000 » Sun Aug 11, 2013 5:06 pm

1024MAK wrote:Some HDD fail when they get hot.
Try waiting until everything has cooled down...

Yes, I'm hoping that might work, hence mowing the lawn...

I don't know much about the inner workings of ST506 drives, but from the sounds of it, there are relays and/or solenoids involved - and it's the solenoid type 'click' which appears to be associated with the previously working A420's drive now failing.

Could be the same fault for the never-working A410's drive...

ajw99uk
Posts: 58
Joined: Fri Sep 29, 2017 3:00 pm
Location: W Yorks, UK

Re: ST506 HDD woes

Postby ajw99uk » Tue Oct 03, 2017 11:26 pm

steve3000 wrote:
1024MAK wrote:Some HDD fail when they get hot.
Try waiting until everything has cooled down...

Yes, I'm hoping that might work, hence mowing the lawn...

I don't know much about the inner workings of ST506 drives, but from the sounds of it, there are relays and/or solenoids involved - and it's the solenoid type 'click' which appears to be associated with the previously working A420's drive now failing.

Could be the same fault for the never-working A410's drive...

Steve,
How would you characterise a "solenoid type click"?

My Rodime 3065 (with an installation of RISC iX 1.15) has sat in a box for several years; tried it today and get six clicks but no spinning-up. Each click ends with a sort of metallic ping. Have tried the usual anti-stiction trick of a quick half-turn around the spindle axis, but fear something more serious!
Running RISCOS: A5000, A540, R140, RiscPC, RPi B
Running *nix: SGI Fuel & Indigo2, RPi2, x86, amd64, RiscPC, A540, R140

User avatar
danielj
Posts: 5139
Joined: Thu Oct 02, 2008 4:51 pm
Location: Manchester

Re: ST506 HDD woes

Postby danielj » Wed Oct 04, 2017 6:20 am

Is the spindle bearing visible? If so, drop of 3 in 1 on it then turn it upside down so it's drawn in by capillary action. Wait an hour, wipe up the excess and give it another go?

d.

User avatar
1024MAK
Posts: 6679
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...

Re: ST506 HDD woes

Postby 1024MAK » Wed Oct 04, 2017 7:25 am

danielj wrote:Is the spindle bearing visible? If so, drop of 3 in 1 on it then turn it upside down so it's drawn in by capillary action. Wait an hour, wipe up the excess and give it another go?

^ a small drop...

Mark
For a "Complete BBC Games Archive" visit www.bbcmicro.co.uk NOW!
BeebWiki‬ - for answers to many questions...

User avatar
danielj
Posts: 5139
Joined: Thu Oct 02, 2008 4:51 pm
Location: Manchester

Re: ST506 HDD woes

Postby danielj » Wed Oct 04, 2017 7:26 am

Yes, not a glug. :D

d.

Boydie
Posts: 118
Joined: Sat Oct 24, 2015 8:25 am
Location: Sunny Wigan

Re: ST506 HDD woes

Postby Boydie » Wed Oct 04, 2017 8:35 am

I had a Micropolis ST506 fail in a very ancient A310 a couple of years back. Can't remember the exact symptoms but it turned out to be one of the caps on the drive's PCB. Luckily, these boards are easily removed and have massive components for the benefit of the ham-fisted!

My Rodime 3065 (with RISCiX 1.13 on it) has died as well. Really ought to try fixing it some time.

philb
Posts: 90
Joined: Sat Aug 05, 2017 6:05 pm

Re: ST506 HDD woes

Postby philb » Wed Oct 04, 2017 10:00 am

steve3000 wrote:from the sounds of it, there are relays and/or solenoids involved - and it's the solenoid type click' which appears to be associated with the previously working A420's drive now failing.

Could be the same fault for the never-working A410's drive...


I don't think I've ever seen a hard disk with relays involved. The "clicking" noise is usually the head positioner banging into the end stop as the controller tries to recalibrate the drive. These old disks didn't have embedded servo data and they basically relied on dead reckoning to position the heads correctly. Older/cheaper drives used stepper positioners, slightly later/higher spec ones used voice coil actuators.

Disc error 13 is "error correction failed" (i.e. too many bad bits for the ECC to correct) and disc error 16 is "ID search timed out" (the appropriate sector didn't appear under the head after a complete rotation of the disk). Both of those are symptomatic of the disk not managing to read the correct bits off the platter.

If you're getting that for data from all the heads then that would suggest that maybe the positioner has drifted out of alignment and the heads are not lined up with the track properly. If one head is doing it consistently and the others are OK then that might indicate a problem with either the electronics or the mechanics of that one head.

timw
Posts: 26
Joined: Tue May 10, 2016 6:55 pm

Re: ST506 HDD woes

Postby timw » Wed Oct 04, 2017 5:40 pm

ajw99uk wrote:Have tried the usual anti-stiction trick of a quick half-turn around the spindle axis...

I think it's worth applying this rotational shock to any ST-506 drive that has been powered off for a while, before powering it on. I powered on such a drive and it didn't spin up, and it then proceeded to move the heads back and forth across the stationary disc. I freed off the disc with the spinning tactic (power off!), but one head was destroyed, possibly by the seek. Luckily it was head 4 of 4, and I can format the disc using only 3 heads, but losing a quarter of capacity on a 21MB drive doesn't leave much left.

JonC
Posts: 609
Joined: Wed May 14, 2014 9:19 pm
Location: Wakefield

Re: ST506 HDD woes

Postby JonC » Wed Oct 04, 2017 6:10 pm

It's a shame there's no ST506 to CF Card converter out there, breath new life into those factory fit ST506 controllers in the A4x0's :mrgreen:
Jon
Image

philb
Posts: 90
Joined: Sat Aug 05, 2017 6:05 pm

Re: ST506 HDD woes

Postby philb » Wed Oct 04, 2017 8:55 pm

JonC wrote:It's a shame there's no ST506 to CF Card converter out there, breath new life into those factory fit ST506 controllers in the A4x0's :mrgreen:


There is, sort of. http://www.pdp8.net/mfm/mfm.shtml

Emulating an ST506 disk is harder than you might think (and somewhat harder than replacing an IDE or SCSI disk with solid state storage) because the disk interface is just a continuous stream of bits and although the data rate isn't that high by modern standards it's not entirely trivial either, a few megabits. The emulator needs to keep track of the rotational position of the virtual disk so that it can write the bits that the controller is sending it into the correct location.

Boydie
Posts: 118
Joined: Sat Oct 24, 2015 8:25 am
Location: Sunny Wigan

Re: ST506 HDD woes

Postby Boydie » Thu Oct 05, 2017 9:58 am

I've had another look at my Rodime 3065. As they say around here, it's reyt proper poorly.

Spins up okay, and accesses the root directory. Verifying it is fine up until &8000. From then on, it starts querying sectors, then giving Disc Error 10, then just Disc Error 16 from about &8B00 onwards. As it gets warmer, the Disv Error 16s spread down towards &8000 and previously good sectors become bad.

The only bits of the controller side that get vaguely warm are the voltage regulators and an IC with a massive heatsink on it (which gets seriously toasty). I know this a hitachi IC but can't remember the part number offhand. Is this likely to be a buffer chip and whilst it's clearly intended to get hot, should it get that hot?

philb
Posts: 90
Joined: Sat Aug 05, 2017 6:05 pm

Re: ST506 HDD woes

Postby philb » Thu Oct 05, 2017 10:41 am

Is there data on the disk that you want to retrieve, or are you just interested in making it work again?

It's possible that the head alignment has drifted a bit, in which case reformatting it might improve the situation. Some disks of that vintage needed to be reformatted if, for example, their orientation changed because gravity acting on the head positioner would cause the tracks to shift. But obviously, reformatting would be a dim plan if you want to get the data back.

Hard to say much about that IC without knowing more details. It might be a motor driver chip for either the spindle or the positioner. If it's getting too hot to touch then I think there's a fair chance that this might be "too hot". It's not inconceivable that the spindle might be getting a bit sticky and taking more motor current to make it spin. Can you take a picture of the PCB?

User avatar
1024MAK
Posts: 6679
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...

Re: ST506 HDD woes

Postby 1024MAK » Thu Oct 05, 2017 2:27 pm

Buffer chips (as in for data) are not normally fitted with heatsinks. A motor driver chip however, is likely to be fitted with a heatsink. As indicated above, if the motor and load are requiring more effort to move, then more power will be lost as heat via the driver chip, so it may well run hotter than intended. You could try some freezer spray on it. Of if you don't have any, use canned air duster with the can upside down. Just be careful of the cold liquid (you don't want it in your eyes or on your skin). Although I'm not convinced that this is the cause.

Mark
For a "Complete BBC Games Archive" visit www.bbcmicro.co.uk NOW!
BeebWiki‬ - for answers to many questions...

Boydie
Posts: 118
Joined: Sat Oct 24, 2015 8:25 am
Location: Sunny Wigan

Re: ST506 HDD woes

Postby Boydie » Thu Oct 05, 2017 3:40 pm

Looking at google images, it looks like it's a motor driver chip.

Unfortunately, the ridges of the heatsink are stopping my ir thermometer reading properly. However, the voltage regulators are reading 35-ish and it's considerably hotter to touch than those.

Boydie
Posts: 118
Joined: Sat Oct 24, 2015 8:25 am
Location: Sunny Wigan

Re: ST506 HDD woes

Postby Boydie » Thu Oct 05, 2017 8:52 pm

On the downside, I've just tried to get the RO3000S out of my A680 working, and the heatsink on that gets even hotter. So it's pretty unlikely the problem on the RO3065 is indicated by a hot heatsink.

On the upside, I've just managed to successfully format and verify the RO3000S, which I had previously believed to be dead. :D

So it's not all bad.

philb
Posts: 90
Joined: Sat Aug 05, 2017 6:05 pm

Re: ST506 HDD woes

Postby philb » Thu Oct 05, 2017 9:32 pm

This page:

http://stason.org/TULARC/pc/hard-drives-hdd/rodime/RO3065-53MB-3-5-HH-MFM-ST506.html

has some interesting info about the RO3065. In particular, it seems that it does have embedded servo data so what I said earlier about the positioner is probably not relevant. For some reason I had been visualising a much older drive than this one actually is.

It also seems that the drive has some limited amount of onboard diagnostics and can actually flash error codes through its status LED. Presumably that's not happening in your case?

Boydie
Posts: 118
Joined: Sat Oct 24, 2015 8:25 am
Location: Sunny Wigan

Re: ST506 HDD woes

Postby Boydie » Fri Oct 06, 2017 7:04 am

I found that page a couple of days ago too, and was hopeful about the status led.

Only problem is that it doesn't actually have an led, and I can't see where on the pcb one is supposed to be mounted (whereas I've found and use this mounting on other drives). I'm assuming that the drive in use led on the arc is driven solely by the arc's controller and doesn't receive feedback from the drive itself?

I've noticed a nut on the underside of the drive, roughly where the head positioner should pivot. The bolt this nut screws onto looks as though it may be hollow. Could my problems be because the pivot's becoming sticky and the head can't be positioned fast enough? It would appear that the outermost cylinder(s) can be read but anything further in is unreliable. Since this is temperature-dependent, could thermal expansion of the pivot lead to binding and inadequate movement?
In which case, what are the odds that this hole in the nut is for addition of extra lubrication?

edit: I've found what I believe to be the connecting pins for the led. It lights constantly, which apparently means drive selected and ready.

ajw99uk
Posts: 58
Joined: Fri Sep 29, 2017 3:00 pm
Location: W Yorks, UK

Re: ST506 HDD woes

Postby ajw99uk » Sat Oct 07, 2017 10:17 am

Hi Boydie,
Where are the LED pins, please?

I think my drive may have a startup calibration problem, though even once I know what the matter is, then fixing it is quite another matter!

Thanks
Running RISCOS: A5000, A540, R140, RiscPC, RPi B
Running *nix: SGI Fuel & Indigo2, RPi2, x86, amd64, RiscPC, A540, R140

Boydie
Posts: 118
Joined: Sat Oct 24, 2015 8:25 am
Location: Sunny Wigan

Re: ST506 HDD woes

Postby Boydie » Sat Oct 07, 2017 4:51 pm

I must admit I'm guessing, but all my Rodime drives have a pair of pins, with collars on them, about 5mm apart, with blobs of solder on them. On the 3065 and 3085 they're in the corner near the two voltage regulators.

My guess is that they had the led soldered to them for post-manufacture testing and it usually got remived after this.

Boydie
Posts: 118
Joined: Sat Oct 24, 2015 8:25 am
Location: Sunny Wigan

Re: ST506 HDD woes

Postby Boydie » Sat Oct 07, 2017 5:35 pm

In the end, I've had to reformat it.

Irritatingly, it's just reformatted, short soak tested, then verified, all without a single error. So it must gave been the head alignment.

Bit of a bummer because it contained a completely virgin, only been looged into once since it left the factory, version of riscix 1.13. :cry:

ajw99uk
Posts: 58
Joined: Fri Sep 29, 2017 3:00 pm
Location: W Yorks, UK

Re: ST506 HDD woes

Postby ajw99uk » Sat Oct 07, 2017 10:39 pm

Boydie wrote:I must admit I'm guessing, but all my Rodime drives have a pair of pins, with collars on them, about 5mm apart, with blobs of solder on them. On the 3065 and 3085 they're in the corner near the two voltage regulators. My guess is that they had the led soldered to them for post-manufacture testing and it usually got remived after this.

Good call! On mine, there is a pair of pins in that corner, with collars, but ending in loops rather than blobs of solder (various other pins sticking out of the PCB are the same, as if designed for ad hoc threading of a wire through, such as for factory testing as you suggest, without having to solder.
Anyway, crushed the collars (very brittle by now) and the loops, attached the power-indicator LED of a SCSI enclosure and used its power to fire up the drive.
Long solid light (during which six loud ping-clicks, but no spinning up; then the LED goes out, then a 1010 error code. "No INDEX" apparently - and no indication of a solution in the Product Specification document. I presume I need to get the platters spinning before there's any hope of finding an index on them :) so I'll have to try some light oil on that nut/bolt in the morning.
Running RISCOS: A5000, A540, R140, RiscPC, RPi B
Running *nix: SGI Fuel & Indigo2, RPi2, x86, amd64, RiscPC, A540, R140

Boydie
Posts: 118
Joined: Sat Oct 24, 2015 8:25 am
Location: Sunny Wigan

Re: ST506 HDD woes

Postby Boydie » Sun Oct 08, 2017 7:44 am

The clicks are the mechanism to unlock the head actuator. When the head parks, it gets latched behind a little lever (that's the "relay" click when it spins down). When the drive's up to speed, a magnet pulls this lever out of the way so the head can move (the "relay" click on spin-up).

I'd guess in your case, either the head's stuck on the lever or the actuator assembly can't move the head.

When my 3065 failed even to catalogue the disc last night, in desperation I opened it up and put a small blob of oil on the actuator spindle, checked the arm could move freely, then closed it up again. Hence my new understanding of the mechanism!

I'm not recommending this, except as a desperate last resort before binning the drive. However, in my case it then took a format and 4-hour soak test without a single error, so I suspect I got away with it. Any dust on the platter (there were three specks) showed up really clearly and just blew off.

The nut/bolt thing may (or may not) be for lubricating the head assembly. The lubricating point for the platter spindle is the one in the middle, with the three motor wires coming out of it.

Boydie
Posts: 118
Joined: Sat Oct 24, 2015 8:25 am
Location: Sunny Wigan

Re: ST506 HDD woes

Postby Boydie » Sun Oct 08, 2017 8:31 am

It's odd that your drive's giving a "no index" error code, given that it's not even managing to spin the platter. You'd have thought it would give one of the motor errors in preference, as this is the more fundamental problem.

Almost as if the controller doesn't know it's not spun up. Particularly since head head release click doesn't normally come until it's up to speed.

Wonder if this is a mechanical drive issue, or a busted controller pcb. In any case, I certainly wouldn't try anything too extreme (such as opening it) until you're sure it's a mechanical issue and are happy to sacrifice the drive and its contents.

Perhaps the logical next step is to try your drive mechanism with a known working pcb. At the very least, it will show which element is broken and guide repair. If we're lucky we may be able to get the data off your drive, and actually get an archive of riscix 1.1x.

ajw99uk
Posts: 58
Joined: Fri Sep 29, 2017 3:00 pm
Location: W Yorks, UK

Re: ST506 HDD woes

Postby ajw99uk » Wed Oct 18, 2017 8:35 am

Put a drop of light oil at each of the two points, and put the drive aside. Now, the drive gives two different behaviours:

Powered by R140 or a SCSI enclosure, it still gives the six clicks then the 1010 error (where I can see it using the enclosure's LED connected to the drive's PCB - I'm assuming I'd see the same with the Arc PSU if I could get an LED connected), as before.

Using a different enclosure (made for Syquest 3.5", so perhaps a less potent PSU?) I get continuing, more rapids click and the LED flickers in time.

Would that suggest it's more likely a mechanical issue?
Running RISCOS: A5000, A540, R140, RiscPC, RPi B
Running *nix: SGI Fuel & Indigo2, RPi2, x86, amd64, RiscPC, A540, R140

Boydie
Posts: 118
Joined: Sat Oct 24, 2015 8:25 am
Location: Sunny Wigan

Re: ST506 HDD woes

Postby Boydie » Wed Oct 18, 2017 2:27 pm

Not sure. Since the controller board powers and controls the platter motor and head actuator, presumably it 's conceivable that a dodgy controller board will behave differently with inadequate power.

Since I have a known working drive PCB, the next logical step is to try swapping my PCB onto your drive. If it works, we'll be able to archive the contents of your drive and salvage riscix 1.1x. If it doesn't, we'll know for certain it's a mechanical fault. At which point, presumably we'll have nothing to lose from opening the drive and fiddling.

ajw99uk
Posts: 58
Joined: Fri Sep 29, 2017 3:00 pm
Location: W Yorks, UK

Re: ST506 HDD woes

Postby ajw99uk » Fri Oct 20, 2017 10:27 am

Boydie wrote:Not sure. Since the controller board powers and controls the platter motor and head actuator, presumably it 's conceivable that a dodgy controller board will behave differently with inadequate power.

Since I have a known working drive PCB, the next logical step is to try swapping my PCB onto your drive. If it works, we'll be able to archive the contents of your drive and salvage riscix 1.1x. If it doesn't, we'll know for certain it's a mechanical fault. At which point, presumably we'll have nothing to lose from opening the drive and fiddling.

Thanks for the offer - I've been hesitating as I would be anxious that the electronic and mechanical might interact to the detriment of your working PCB (was it Zip drives where a dodgy drive could damage a cartridge in such a way that the cartridge would damage other drives?). I have also been playing with the mknewfs script, so as to skip the "newfs st0a" lines and run the rest on an existing filesystem set up under 1.21, and got as far as a multi-user boot from 1.15 kernel, system and archive floppies (after a while spent swapping FDs from R140, A540 and a spare!). I had the new partition set up on a Syquest cartridge, but this morning could not get it to boot. Then some alarming noises - seems one of the pads at the end of the arm has partly detached :-(
Image
So I shall not be using that any more. May also explain why some of the cartridges I was trying to read the other day errored - luckily I'd only looked at a few before a stuck platter seemed to put the drive out of operation. Pleased to revive it yesterday (the arm had slipped off its guide/rest, so I poked gently to realign), but maybe I was too clumsy in doing so.
Anyway, the point of working on the floppy-install to SCSI is that I have found among the packageadmin disks a set of "dump" disks - if I can restore those to a SCSI drive, clean out stuff that seems personal to the previous owner and image that installation, we'll have a clonable version of 1.15. Using a relatively small partition, it might even be restorable to a 3065 set up with a matching RISC iX partition table, using iX 1.21 as a host system, for those who hanker for a more authentic 1988/9 experience.
My other notion, a dual-ST506 system with / on one 20MB drive and /usr on another, hopefully providing enough space for 1.15, is a possible (though more complex to clone) fallback. The Arc doesn't seem to like two drives on two controllers so I would need to find/make a dual-drive control cable.
Attachments
Syq_heads.jpg
(5.09 KiB) Not downloaded yet
Running RISCOS: A5000, A540, R140, RiscPC, RPi B
Running *nix: SGI Fuel & Indigo2, RPi2, x86, amd64, RiscPC, A540, R140

Boydie
Posts: 118
Joined: Sat Oct 24, 2015 8:25 am
Location: Sunny Wigan

Re: ST506 HDD woes

Postby Boydie » Fri Oct 20, 2017 12:31 pm

The documentation that came with the R140 told you that the first thing you should do with a new machine was make a complete set if backup floppies (I never did so because I only powered it up twice from new, and the second time the HD was broken) so you could restore the machine to its vanilla state if necessary.

If you're very lucky, this will be what your backup floppies are and there won't be any personal data.


Return to “hardware”

Who is online

Users browsing this forum: No registered users and 3 guests