<Crass Commercial request> Is there anyone with experience doing hardware floppy emulation here? I may have a project that could use some paid expertise in this area... </Crass Commercial request> Peter Wallace Mesa Electronics

Tony wrote: Back in the day, I used to buy these drives at the Tektronix Country Store for pretty cheap...something like $35 or so. This was in the day when 5 1/4" ST-506 interface har disk drives were EXPENSIVE. In those days, I built my own computers, based on an internal Tektronix "development" platform called the "Board Bucket". These were a quite diverse, multi-CPU platform based on a common bus that had various CPU boards (6800, 6809, Z-80), and variou I/O boards (floppy interface, serial-parallel I/O board, RAM boards (dynamic and static), ROM boards (using 2708's), and even later things like ROM burners, Video Boards (I designed one based on the TI Graphics Chip used in the TI 9900 home computer), Modem boards (Bell 103), and various other custom boards. There was a "blank" prototyping board available for these systems. I did build an interface board for a Micropolis 1203, so I could use a hard disk on the "FLEX" 6809-based operating system. It was a pretty trivial parallel interface, with, IIRC, a line that controlled reading or writing from the onboard controller, and a "data or status/command" signal that told the controller where you are tell the controller whather you are reading or writing control commands, or data. It was an 8-bit data interface, parallel, with a strobe signal to read/write data to/from the controller. There were a couple of status signals that also came out to tell the computer the whether data was ready to be strobed out or pushed in. This was a long time ago, so my memory is hazy. The one thing that I do remember is that there was a weird little jumper block on the controller board that controlled something to do with the geometry of the disk as presented by the controller. It could specify the size of a sector...essentially, the drive could have a selection of sector sizes. As I remember, to get the 256-byte sectors that the FLEX operating system wanted, I had to figure out how to re-wire this jumper, as by default in the 8560's that these drives were used in, the sector size was either 512 or 1024 bytes/sector. Other than these drives being rather noisy, and requiring quite a bit of power, they were rock solid reliable, and very easy to interface to. They were NOT tolerant to shock, though. Bump the drive too hard while it was running, and they'd crash pretty easily. I wrote a driver for FLEX that made this drive appear as a number of logical drives, because FLEX had a limitation as to the number of sectors that a drive could have. It was pretty cool to have a system that had a "big" (I think that these were 30 Megabyte drives) hard disk on it in the day when most folks' machines were ucky to have a pair of 8" or 5 1/4" floppy drives. very straight forward, and the controller was pretty smart, providing commands for low-level formatting the drive, standard "read" and "write" sector commands, as well as extended read/write commands that would read/write the sector header information, pretty robust status reporting, and pretty robust status/error reporting. It was similar to SASI, but "simpler", and as I recall it took me virtually no time at all to get one of the drives up and running on my system. I'm sure that if someone can find the technical docs for the 1203, it'd be a rather simple proposition to build a simple interface for a PC or even a small microcontroller) that could allow a drive to be imaged. My "system" also had two DSDD 5 1/4" floppy drives, and a "RAM Disk" that I made out of a slightly modified 64K RAM board that had a 256-byte "window" into the address space, and a parallel port that would select which "sector" would appear. Later, I bumped the capacity of the board by using 64K DRAMs instead of the stock 16K DRAM chips, so I had a 256K RAMDisk (really fast). Later, I built a SASI (purely I/O driven, no DMA) interface that I could hook up 5 1/4" hard disk drives to, using an early Xebec SASI to "ST-506" controller. I bought a very early Tandon 5MB 5 1/4" full-height ST-506 drive, and got it running on the system. I still have a number of the two types of Board Bucket chassis (nice aluminum chassis with a backplane board in the bottom, with card guides) ...there was one with just a backplane, and an external power supply that connected to the chassis by banana plugs. There was also a smaller backplane all self-contained with its own power supply. There was a ROM monitor that talked through one of the serial/parallel I/O boards (two serial ports and two parallel (Motorola 6821's, I believe) ports. The CPU boards varied a bit...some had ROM on board for the monitor, and some you had to put in a plug-in ROM board to host the monitor code. Some CPU's had TIL-311 displays on them to display the address and data bus, with a single step button that would allow you to step through the code, as in a simple front-panel, though not possible to load data or address. That was one board (a 6800-based board) as I recall that could be used for in-circuit emulation. I've also got a bunch of the boards, including RAM, ROM, I/O, floppy, my homebrewed SASI interface, the prototype of the video board and a "production" version (which included a real-time clock chip also), a couple of Bell 103 modems, a ROM blaster (I think it could program everything from 2708's through 2764's) and maybe even a few blank prototype boards. These things were cheap, because you could buy the bare boards through Tektronix Engineering Stock for <$20 or so, and buy the parts to stuff them at Tek's cost plus 5%, and put together some pretty capable systems for very little (compared to "commercial" minicomputers like Altair/IMSAI/Processor Technology/Tandy-Radio Shack/Digital Group/Polymorphic Systems/Smoke Signal Broadcasting, and SWTPC [among many others]) money. I even could get the folks on one of the flow-solder lines to run my stuff boards through the flow-solder machine, which really sped things up as opposed to hand-soldering everything, which I also did a lot of. I even built a "ROM" board using Dallas Semiconductor 32Kx8 NVRAMs for loading up dedicated applications into, and used a number of these machines for a very early home automation system that could control lighting (X-10), security system, weather monitoring (through a hacked Heathkit weather station), a voice-response (using a Votrax Chip, and a DTMF decoder IC hacked into one of the modem boards) system that one could use to remotely dial-in and check status of the system and command things to happen) as well as an IR interface that could command devices like the TV, VCR, Cable TV box, etc. I have a complete 6908-based system that I use from time to time for various little projects. It works great to this day. I sure had a lot of cheap fun (and great learning experiences) from those machines back in those days (early 1980's as I recall). I also have some old Tektronix code that I believe works under Utek that provides assemblers for the 6800 and 6809. The ROM monitor has a means by which the 2nd serial port on the I/O board could be hooked up to a "host" computer, and assembled absolute code from the assembler could be downloaded in to RAM pretty easily. Sorry to wander off-topic, but figured that there might be some interest in these old internal Tektronix "toys". Rick Bensene The Old Calculator Web Museum http://oldcalculatormuseum.com

Tony wrote: host parallel (so surplus). Yes, the drives that I got my hands on (I also had an 8560 at one point running TNIX) all had the Micropolis "controller" board that was positioned over the top of the "bare drive" logic board. I probably had three or four of these things at one time, but over the years, they either died, or I gave them away to friends who had interested once I was able to afford another Micropolis drive, the 5 1/4" ST-506 interface 20MB drive (can't remember the model number). The 8" Micropolis drive was also used in the Tektronix Magnolia...a very early prototype workstation-class machine that used a bitmap display, and was based on Smalltalk as the native operating system. Tek (to my knowledge) never sold any of these machines, but quite a few of them were built as prototypes, and used in various areas of Tek as engineering workstations. The Magnolia was way ahead of its time. Had Tek had the marketing and sales skills to sell this thing as a computer, they could have grabbed the engineering computing workstation market before it even really existed. This was long before Sun or PERQ. It was more in the timeframe of the early Xerox PARC machines. In fact, some of the software engineers that developed the working environment for the Magnolia came from Xerox PARC. The 5 1/4" full-height Micropolis drives were the "standard" drive in the early Tektronix 6130 UTek machines, and I could get them at Tek cost plus 10%, making them reasonably affordable, as Tek purchased them in large volumes. These were interesting drives, because they had two platters in them. One surface had pre-recorded information used by the electronics for accurate head positioning. In the drive as it came from the factory, only two surfaces were used for data (for a total of three surfaces, two data, one servo). However, there was a head, and amplifier for the remaining surface. Each surface held 10 Megabytes, so the standard drive had a capacity of 20MB. There was a little flex circuit board that came out of the sealed HDA that terminated on a 2x4 square pin header that plugged into a socket on the drive logic board. On these drives, the two pins that went to the head/amplifier for the "unused" surface were simply cut off. We found that if we carefully desoldered the disfigured header, and soldered in a full 2x4 header, we could format these drives as 30MB drives, and they worked just fine. Sometimes when low-level formatting, there would be a few flaws on the third surface, but nothing major...resulting in only a slight reduction in capacity. I'm thinking that originally all of the drives were manufactured as 30MB drives, and that those that had some number of flaws in a certain surface would have that surface "disabled" by simply cutting off its pins to the main logic board. Apparently the drive electronics could tell which were "good" surfaces by those which had a signal riding on them, and make it all transparent to the ST-506 interface. A cheap way to make sellable (albeit lower-capacity) drives out of drives that failed QA testing as 30MB drives. These drives were pretty bulletproof, too. I still have one (modified as described above) in my Tektronix 6130, and it ran literally for something like 8 years straight. Then, I got my hands on some Sun machines, and the 6130 was put aside. However, about a year and a half ago, I had occasion to run into the 6130 while unpacking stuff from a storage unit, and I brought it home, hooked it up to a trusty Heathkit H19 terminal, and powered it up. It booted up right away. Not even a fsck required. I even remembered the root password! Did a full "dd" of the disk to /dev/null, and no read errors at all. I think I still have a couple of these drives stored away. I should drag them out and see if they still spin up. Probably could get them to talk to an old Adaptec ACB-4000 SCSI to ST-506 converter, and hook 'em up to a PeeCee with an Adaptec Single-Ended SCSI controller (I have a slew of ISA and PCI-bus HBAs) to see if they still work. I'd bet dimes to donuts that they will work just as good today as they did in the mid-'80's. -Rick

On 7 May 2007 at 21:02, John Kourafas wrote: I believe they're talking about *bus* speeds. In any case, I've used the 1006WAH and 1006SR2 sucessfully on a 100Mhz 486 DX4 system. Aside from a few oddball "times things in a CPU loop" cards, most ISA cards should work in modern (e.g. PIII, P4 CPU) systems. ISA bus timings haven't changed drastically from the old PC-AT days. Heck, I've even found PII ISA buses that are *slower* than the original PC- AT Cheers, Chuck