Amstrad 2086/30 computer possible hard drive fault

[high_vacuum_house]

Good evening,
I have just purchased one of these old models to have a playabout with as we had a near identical model that my parents had probably late 80's early 90's

I think that it has a fault with the hard disk as all I get is "insert a SYSTEM disk in drive A and press any key" and will not go through the boot up sequence. I haven't got a system disk for it, but by putting a non system disk in the drive, it reads it and tells me this isn't a system disk.

Also, it recognises if the keyboard is disconnected.

There is an int/ext switch located on the right hand side of the unit, but this doesn't change anything. The hard drive LED flashes 3 times in a row pauses and repeats this sequence. I cannot hear any noise from the hard drive which I remember is quite noticeable on boot up.

I have opened the machine and it is quite clean with no corrosion or dirt which is a good sign. The hard drive and it's controller PCB plus all of the ribbon cables have been removed and all the connections are seated fine.

Does anyone have an idea what to try next?

Many thanks,
Christopher Capener

[high_vacuum_house]

Good evening,
I have had another look this evening at what is going on and have re checked all of the connections. It appears the hard disk is actually doing something on switching the machine on. The stepper motor on the side is rotating backwards and forwards before stopping. I think there is some mechanical issue with the hard drive, I think the heads are scuffing against the platter as there is definitely that sort of noise from within. It does occasionally come up with "cannot read operating system" on switching on before freezing

What is the reliability like on these old hard drives? This one is a Western Digital model WD383R. Wonder where I would find a replacement ??

Many thanks,
Christopher Capener

[ortek_service]

I wonder if it's lost it's 'CMOS' setup?

IIRC, some early (PC1512/1640?) had AA-batteries in the top. But later ones had a DS1287 RTC IC wth a lithium battery on top in its epoxy-potted case. These are no-doubt rather obsolete (I recall paying £5 for these at radio rallies, afterI got lots of Astrad PC Motherbiards with these missing) and decades past their shelf-life. But the top battery can be cut-off and replaced by a lithium coin-cell holder (with instructions on the 'net).

If the hard-disk type no. in the CMOS isn't correct then it will fail to boot (and could also have incorrect boot device). Although I'd expect to get a Date & Time error come up. However there may not be a press F1 / F2 / Del etc to enter setup, as many early PC's rather unfriendly lacked a buitl-in CMOS setup program, so you had to run one from a system disk.

I do also recall some Amstrad PC HDD's suffering from 'Stiction' where the heads stick to the platters (they do normally rest on these, when not spinning, hence need to park them to an unused bit - Which you had to do via a WDPARK command)
I thinks it's only later 'moving-coil arm' ones that could return to park position (+ often had auto-locking with solenoid release). Whereas original stepper-motor driven head-positioner wasn't able to move quick enough once power had died.

One apparent 'fix' for a stuck HDD, was to give it a sharp tap on the top, whilst powered-off. But with such relatively-low capacities, you could probably get away with just removing the lid in a fairly-clean environment and seeing what it does, giving it a gentle push - Just don't touch the disk surface, as I was never able to remove fingerprints etc from these, even with usual IPA etc cleaning solution.

Regarding making a 'System' floppy-disk, you really need access to an old PC with a 5.25 FDD, that boots to DOS so you can do a format with /s extension / use Sys command on a blank formatted disk (or tick the 'make system' option under Win 9x etc).
There's no doubt some disk image files for FreeDOS (or MS-DOS) on the net, to recreate one. But you'll need a PC with a FDD interface (and still supports 360KB 40T D/S 5.25" drives in the BIOS), in order to connect the Amstrad drive to (assuming it sill works).
- Not sure if you can do it via USB and something like a greaseweasel?

You may well find images of the original Amstrad System disk on the 'net
(IIRC, I've got some PC1640 etc ones, that came with GEM, before Windows 3.x etc had caught on)

[paulsherwin]

The disk may simply be corrupt rather than broken. If so, you should be able to boot from a floppy and reformat it.

That said, some of the HDs fitted to Amstrads were real shockers and failure was quite common even when they were new. They were noisy and ran *extremely* hot - I remember that one got so hot that you couldn't touch the front of the drive bay.

[high_vacuum_house]

Ortec Service is correct, I have carefully opened the top cover of the hard disc and have discovered that one of the heads is missing and has become detached from its arm possibly getting stuck to the discs surface. I found the head in the cavity below the discs.

You can see that the head is missing on the upper surface of the lower disc and the score on the disc where the sprung arm has rubbed against its surface.

Oh well that looks like the end of this! I have had a lot of bad luck with things the last few weeks! Looks like finding a replacement HDD is out of the question.

Christopher Capener

[paulsherwin]

I think they are standard ST506 drives, so you should be able to find a replacement, though obviously they're all very old now. You'll also have to track down a setup floppy so that you can tell the BIOS about the disk geometry (sectors, heads etc).

[ortek_service]

Yes, I think Amstrad may have just used a 'HDD on a full length ISA Card' on some of these - I'm sure I've got one of these somewhere, where a 3.5" HDD is mounted on the end of the ISA Controller Card.

And back then, may well be an 'ST-506' (Technically the model of an HDD, whose interface was widely-copied) interface, originally using MFM, then RLL for greater capacity.
And no doubt could pick up an old but still working drive off eBay etc.

A bit less original, you could always put in one of those 'Intelligent' IDE ISA Cards, with its own auto-detecting BIOS, so avoids need for PC's BIOS to support the drive type / copes with much larger ones.
As old IDE drives should be easier to obtain / may be able to use a Compact Flash adaptor, to make it less of an issue.
Although maybe Amstrad had included an IDE interface on the Mobo, on the 2086?

[pbengtech]

Hi Christopher,
I should have some Amstrad system discs on 5 in floppies if they are of use to you. Should have some ide hard drives somewhere as well.
The Western digital hard drives were so unreliable Amstrad took western Digital to court over them.
All the Amstrad PC’s with the twin 5 in floppy drives I owned all had faulty hard drives in them.
PM me if what I have is of any use.
Phil b

[Reelman]

In my experience hard drives do not grow old gracefully, and should never be considered anything but a temporary storage facility at any age. Any drive over 20 years old should be expected to fail very soon. Maybe I’ve just been unlucky with my ones. Actually, I’ll have to power up my original AT sometime to see what happens, be very surprised if the hard drive still works.

Peter

[SiriusHardware]

I found a couple of pages of info which themselves may have further useful links.

https://oldcomputer.info/pc/ams2086/index.htm

https://www.seasip.info/AmstradXT/scsihd.html

The second page was written in 2004 when 8-bit SCSI controller cards and SCSI drives (which they are suggesting should be used to replace the original drive) were possibly a thing which could still be found for a reasonable cost. No idea whether that would be feasible now.

[ortek_service]

Amstrad in their x086 ranges (many of these from 2086 all the way to 9086?) used a 8086 CPU (with its 16bit external data bus, although multiplexed with address bus) - about the only ones who did, so maybe they could call it a '16bit PC' ?
- although it didn't have extra instructions etc of the 286 in a 'true PC-AT' etc

But I'd forgot that strangely they only had 8bit ISA slots - as used on original 8088 CPU (with their 8bit external databus) based "PC's".
Which rather complicates the use of IDE, that was basically a breakout adaptor of the 16bit ISA slot's data bus etc, so IDE was usually 16bit data.
So presumably that's why that article: https://www.seasip.info/AmstradXT/scsihd.html
used an 8bit SCSI interface.

However, I do recall their being an 8bit version of IDE, for use with 8bit ISA slot PC's. But there may have not been many IDE HDD's that supported 8bit mode (Not sure if any supported both 8 & 16bit modes?).
But I wonder if there was 8bit ISA card that converted to a standard 16bit IDE, so most HDD's could be used?

[ortek_service]

Originally Posted by Reelman

In my experience hard drives do not grow old gracefully, and should never be considered anything but a temporary storage facility at any age. Any drive over 20 years old should be expected to fail very soon. Maybe I’ve just been unlucky with my ones. Actually, I’ll have to power up my original AT sometime to see what happens, be very surprised if the hard drive still works.

Peter

Well unfortunately, there's not too many alternatives that last any longer!

FLASH Memory cards are often only rated for 10yrs retention (although you may be able to do a 'refresh' rewrite every few years, if you remember).

SSD's seem to be classed as 'worn-out' when retention drops to only 2years (but how would you know, unless you carefully monitor the number of writes to each location - which may be masked by built-in wear-levelling, as on 'managed FLASH' SD-cards etc)

With CD-R's, original Gold-layer ones were meant to last 30years, but I had an unbranded one's gold top surface peel-off in much-less time. IIRC, Silver ones were 10years.
And dye ones were usually only around 2yrs (if not left in Sunlight).
Although Verbatim did some 'Super-AZO' ones with a claimed 100yr archival life? - But Verbatim DVD-R's didn't seem to have any specified retention periods.
Although I think DVD-RAM's, with their higher amount of error-correction, were meant to be 30+yrs, but these seemed to have rather fizzled-out these days (although not many shops selling DVD-R(W)'s either)
Plus it seems CD-RW's & DVD+/-RW's will be even worse than non-rewriteables.

Hard-drives are probably most reliable if kept spinning - with fairy-high MTBF figures (although then problems with constant heat degrading things) - to avoid startup-up wear on heads (They used to be often only rated for a few thousand start-stop cycles, which could equate to only a few years typical use).
I think the retention-time, if not in use, could well be many decades (especially on older lower data-density ones).
But the biggest risk on these is possibly that the disk platters / head materials start to break down / react with other / stick together.
- And when powered-up again, end-up damaging the heads (as in this case). It might be safest to remove the HDD cover (although not normally advised), first, to check the platters will spin-freely before applying power and these trying to run at full-speed.
(A bit like not trying to start engines in old cars that haven't run for years, before turning by hand / checking the engine thorough / replacing cam belts etc. )

I have quite a few very old HDD's in the odd PC-AT, and a few Acorn Archimedes etc, that I'll have to try again sometime. Although it seems the external 'Winchester' unit I have for the Beeb/Master (for Econet file-server etc) still runs (although it seems those are infamous for cracking then smoking Rifa clear-Epoxy cased mains-filter capacitors in the PSU hadn't fared as well).