Patrick Finnegan wrote: if these are what I think they are, they are PA architecture and there is an HPUX 11.11 release that HP will send you that is pretty good. You can use GNU tools for compiling since the HP compiler is expensive and not free. Main drawback is that the main disk storage is HVDiff SCSI . I plan to use a pacific converter to single ended and use regular drives, but the usual system will have a big old HVD scsi drive in it which isn't very big in capacity. Jim

The first trap for the unwary is that this machine also goes under the number 'HP9816'. Maybe a search for that would be more helpful. For the record : HP9000/226 == HP9826 HP9000/236 == HP9836 HP9000/217 == HP9817 too. Go to http://www.hpmuseum.net/ . I don't think there's much machine-spcific infopr fo the 9816 on that site, but there should be things like HP BASIC disks for the 9000/200 family (of which this is a member). And if you can stick my handwritinmg, there should be a set of schematics. It's a beautiful little machine IMHO, built very well. It's possible to remove all the PCBs apart form the PSU only having to remove 1 screw, for example. There are 6 PCBs (and 2 expansion slots) packed into that little case.Flat in the bottom there's the processor board, which also contains RAM, boot ROMs, RS232 and HPIB ports, keyboard interface, etc. That plugs into a little backplane with 2 DIO slots. DIO is the HP expansion bus for the 9000/200 series. physically it's the same connectors as S100 bus, but the PCBs are smaller, and the signals totally different. It's basically the 68000 bus, with lites to talk to an optional DMA controller. The power lines are regulated +5V and +/-12V. On top of that assembly is another larage PCB mounted horizontally that plugs into the same backplane. It cotnains the monitor circuitry and the +/-12V regulators (the PSU outptus +/-14V). Plugged into that are 2 small PCbs on the right of the CRT (lookinf from the back of the machine), the larger one on the left is the text video board (6845 + 4K RAM + character generator + TTL), the smalelr one is the graphics board (16K DRAM + TTL). Finally on the left of the PCBs is a metal box containing an SMPSU board. You are lucky if you have a machine with 1M on the CPU board. Most have 256K (32 off 4164 DRAMs). I figured out when looking at said board that it had been designed to take 41256 chips too, so I desoldered all the RAMs in mine, replaed them with 41256s (alas you can't use sockets, the PCB won't fit under the backplane if you do), fitted a 74F158 mux chip and a resistor for the extra address input, and cut a couple of jumpers. Only much later did I hear that HP had shipped machines like that. The 2 DIO slots are wired in prallel. The lower one can take a PCB with an external connector (like a GPIO parallel card), the upper one can only take 'internal' cards (RAM, DMA controller, ROM, bubble memory, etc). Without an HP disk drive, it's not going to do a lot (well, if you have a ROM BASIC or ROM HPL board in the machine, it'll boot to that, but you can't save your programs). You want to track down an HP9122 (dual floppy) or HP9133 (floppy/hard disk) or something like that. Thos plug into the HPIB interface. The serial port is a strange connector (50 pin microribbon), but standard RS232 signals (once you've worked out where they all are). Finding the wirelist of the cable was an entertainment. I finally found it in the Pascal standard procedures manaul, it's also in one of the Basic manuals. It's not in any specidfically-hardware manual that I could find. Mine came with the samll keyboard. It also has the 'dial', actually a rotary encoder. That's used for scrolling. On its own it moves the cursor horizontally, if you hold down the shift key it moves it vertically. It feels very nice to me. The keyboard interface is power, ground, clock (processor board to keyboard) and a bidirectional line. From the procesor board to keyboard it's a reset to clear the scan counter and start a new scan. From keyboard to processor, it carries the state of the current key. Strangely, for all the keys are scanned by the hardware in the keyboard, they are not wired in a matrix. Instead one side of each key is grounded, the other goes to an input on a 4051 mux chip (the keyboard is full of those!) When mine came, it was missing the left shift key and a couple fo the key contacts were mangled. I soon realised the latter were the same as used in the VT52, VT100, TRS-80 Model 1, etc. Digging in my junk box, I found a 'spare' VT52 keyboard PCB from whih I raided some contacts, and parts like key plungers to replaced cracked ones. I found the shift key on that was the right size, even if the wrong colour and design to fit the 9816 keybard, so I put it on. If I ever find the right one, I can swap it. [1] No I didn't raid a repariable VT52. I was given a load of DEC bits, and this keyboard was amongst them. I assume somebody else raided a VT52. I've enver seen the larger keyboard 'in the flesh'. I assume it's electrically compatible, although it might well be different internally. HP9000/200 machines don't seem to have much collector interest, which is a pity. I find them all to be very well made, repairable, interesting designs. Tracking down expansion boards is a bit of a problem, but when you find them, they tend to sell for low prices, thankfully. -tony