VirtualAcorn Technical support:
Recovering data from a RISC OS hard disc
Over the years there have been a number of hard disc standards. It's important to understand that these standards relate to the connection of the hardware, not to the way data is stored on that hardware. A hard disc with an IDE interface can be physically connected to an IDE interface on a computer motherboard. So a RISC OS formatted IDE hard disc can be connected to a Windows PC. However the PC won't be able to make sense of the data stored on the drive because RISC OS stores data (called formatting) in different way to Windows. To read a hard disc you need both physical compatibility and formatting compatibility.
Types of interface
There were several different types of hard disc fitted to 32 bit Acorn computers over the years. The early 32bit machines (A400 series) used ST-506. The A540 and R260 series used SCSI. From the A5000 onwards Acorn machines were fitted with IDE interfaces. In addition a number of different manufacturers made hard disc interface Podules (the official Acorn name for an expansion card). An Acorn A300 series computer didn't have a hard disc interface when new, but could be fitted with either ST-506, SCSI or IDE from a dozen different manufacturers. To make matters more complex a drive formatted and used with one manufacturer's Podule would be unlikely to work using a different manufacturer's Podule.
As well as problems with incompatible podules and different types of interface the size of drives supported by RISC OS also changed. Early versions of RISC OS only supported drives up to 20 MB in size. As RISC OS was developed the size of drives that were supported gradually increased. A big change happened with RISC OS 4.02 which was the first version of RISC OS to support long file names. RISC OS 3.7 and earlier only supported a maximum of 77 files per directory with 10 character file names.
The ST-506 hard disc format was introduced in 1980 by Shugart Technology (now known as Seagate Technology). When introduced it supported drives up to 5 MB in size. A number of manufacturers produced ST-506 compatible equipment but the interface was expensive to make and slow. Acorn introduced it's own ST-506 podule in 1988, drives formatted using the Acorn ST-506 podule cannot be read using any other ST-506 interface. If you have an ST-506 hard disc the only way it can be read is using a first generation 32 bit RISC OS computer (A3x0, A4x0 and A540) running RISC OS 2 and fitted with an Acorn ST-506 podule.
SCSI, which stands for Small Computer Systems Interface, was developed in part because of the limitations in the ST-506 standard. It was developed over a number of years but a fully documented standard was finalised in 1986. SCSI supports a number of devices, typically up to 7 plus the interface, in a 'chain' with a physical terminator block on the final device in the chain. If you have a SCSI system on a 'real' RISC OS computer and it's not working correctly check to see if a terminator is fitted. SCSI was designed as a general purpose interface that wasn't just for hard discs. Other common SCSI devices include tape backup machines and early flatbed scanners.
Acorn released their own SCSI podule with the A540 in 1989. SCSI was also used by Commodore and Apple. It was an easy interface to make as everything was documented and the required controller circuitry was cheap compared to alternatives. Adoption by large companies also saw the price of drives rapidly pushed down during the late 1980's and early 1990's. A number of different manufacturers made RISC OS compatible SCSI podules using either the cheaper 8bit or faster 16bit interface. Whilst SCSI defined a standard for the hardware devices the actual format for storing date on the drive varied between different manufacturers.
A SCSI formatted drive will need to be read using the same type of Acorn machine and the same type of SCSI podule that was used to format the drive. If you have a SCSI hard disc but don't have the computer that was used to format/read/write to the drive the best bet is to try reading the drive on a RiscPC with an Acorn (or Castle Technology) SCSI interface and don't forget the termination. If the drive cannot be read by An Acorn interface it's going to be all but impossible to recover anything from the drive.
IDE, which stands for Integrated Drive Electronics, was originally known as Parallel ATA. It first appeared in Compaq PCs in the late 1980s. An IDE interface was cheaper to make than a SCSI interface as the 'intelligent' part of the system was on the storage device, not on the computer interface. A number of computer manufacturers adopted IDE, including IBM, Apple and of course Acorn. The Acorn A3020, A4000, A5000, A7000, A7500 and RiscPC all have IDE interfaces on the motherboard.
An IDE disc should be the easiest to read, but the devil's in the detail. A drive of 40 MB or less in size will probably be from a first generation machine. An A5000 or A4000 should be able to read the drive. A drive under 100 MB in size could be from a second generation Acorn computer, such as a RiscPC, or could have been fitted to a first generation machine with an IDE Podule. A drive over 100 Mb but under 500 MB will probably be from a RiscPC or A7000 running RISC OS 3.7 or earlier. A drive over 500 MB will probably be from a Risc PC or A7000/A7500 running RISC OS 4.02 or later and formatted for long file names.
As well as the official Acorn 'IDE' interface a number of third parties made IDE podules, some of which ran their own filing systems. Drives formatted with an Ian Copestake Software (ICS) IDE podule used their own IDEFS format, which were incompatible with other IDE Podules (and often incompatible with itself - ICS products did not have a good reputation). APDL produced several different types of interface, including the 'Blitz', which supported much larger capacity drives than the Acorn produced podules.
Only a 'real' Acorn computer can read an ST-506 or SCSI hard disc. However support for Acorn formatted IDE hard discs does exist for Linux. This article http://www.filecore.net/riscos/public/filetransferlinux.html has step by step instructions for setting up Ubuntu to recover files from a RISC OS formatted IDE hard disc.
Once the drive is working
Once you have the drive working the data still has to be extracted in a form that can be moved. If you are reading an IDE disc using Linux then moving the recovered data to a different storage device should be a matter of drag and drop. Otherwise it will be necessary to move the files without losing any of their unique RISC OS information, such as the filetype. This is easily achieved by storing the files in compressed RISC OS format archives created with ArcFS, SparkFs or Zipee, Acorn Squash is not suitable for this purpose. Once the data has been safely stored in compressed archives it then needs to be moved.
If your 'real' Acorn computer has a CD Writer you can create an ISO image of the data and burn it to a CD (or several CDs if necessary). If your Acorn computer has a USB interface you can copy the compressed archives to a PC (FAT) formatted USB thumb drive.
If your Acorn computer has a network interface it should be possible to connect it to your home network and then transfer the compressed archives to a PC for storage. An in depth guide to networking Acorn machines is beyond the scope of this article as it would take a small book to cover the topic in detail. The best on-line source for RISC OS networking information is http://www.riscos.org/networking/.
There are several different types of Acorn network interfaces which can use either 10BASE2 or 10BASE-T cables. If you are using 10BASE2 don't forget that you will need a terminator on the spare end of the T piece connected to the network interface. You may need a converter as very few modern devices support 10BASE2. If your machine doesn't have a network interface you may be able to find one on eBay, but make sure it comes with the software to make it work.
The final option, if you can't use any of the above is the good old DOS formatted floppy disc. Earlier first generation Acorn computers fitted with RISC OS 3 can read/write 720K discs. From the A5000 onwards 1.44Mb (High Density) discs can be used. An advantage of compressing the files before copying them to a DOS floppy is that more data can be fitted on each disc.
First generation machines with either Arthur or RISC OS 2 may need an extra utility to read/write DOS formatted discs. If you don't already have this utility then obtaining a copy may be impossible (we don't have it).
Regardless of the method used to transfer files everything should be compressed into archives. It's possible to open RISC OS format .zip files in Windows, but don't do it. If you decompress a RISC OS format archive using a different OS you will lose the filetype information. Only decompress the archives inside VirtualRPC. Copying and moving the archives around can be done with any OS, it's the decompression that's important.
A final note
This article only covers a small part of what is a very complex subject. Given the complexity that surrounds recovering data from hard discs that could be 25 years old there is one further alternative that should be considered, paying someone to recover data for you. Only one company still offers this service, CJE Micros in Worthing. There is never any guarantee that a hard disc is intact and working, but if you don't have the time or experience to recover data from a drive it might be worth talking to CJE Micros. Just remember that commercial data recovery is not cheap.