Could you, Bill McD., or any other folks who've joined the list in the past several years, shed more light on the HP 9825 CPU and others around that time? Al Kossow's statement wondering whether info is extant on HP's microprocessors is right on the money for me. Bill's message, copied below, is a good bit of information that I had not known to now. Thanks! (A BPC manual was available??!!! Scan that sucker if found!!) Additionally, I understood the CPU in the HP250-series business computers different microprogramming. A source seems to confirm something like this, except mention is made of using a CPU similar to the 9845. Check this interesting URL which is that source I mention: http://www.hp-eloquence.com/history/history.html My system consists exactly of the "Delila" pictured lastly on the page. CPU box is that to the left of the very similar looking HP7908 16 mb drive on the right, both on wheels. Both boxes are equipped with DC600 tape drives and typical gooey tape rollers which need to be repaired :-/ Firstly, are the 9825 and 9845 processors the same part? Secondly, can the HP250/9825(45) uP connection be verified by any other source(s)? Thirdly, could anybody add more to the limited amount of technical info available on the HP250 machines? I have a 250-30 upon which I've done some study back in the early 90's ? la Tony Duell (e.g., self-generated hand drawn schematics, some address decoding, bits and pieces of other tech notes scrounged from here and there). Up until a few years ago there was a severe lack of available info on the machine. I haven't focused on searching since then, mainly because of other stuff getting in the way (such as life, etc.) Also, anyone know whether the boot ROMS were devices electrically compatible with anything known commercially? I want to know so that I can possibly dump them for preservation and also learn a bit by observing human readable system messages and from whence they are called plus guessing at some machine code and hand disassembling a few areas. That's sometimes interesting and fun when discoveries come about. Right Tony? In short, is there anything substantial available out there on the HP250 machines other than the URL shown above? Google caught the URL I gave but picking through many hundreds of others yielded very little of additional substance. My machine still ran last I checked about two years ago and came with an HP7908 hard disk and HP Business BASIC operating system (with the very last version update released which was implemented by our HP CE in 1987 or '88), two HP2631 printers and two or three 2622 terminals. BTW, I have an HP7912 65 mb disk drive that I got with the system which always showed a "Servo Error" error code on the two-character LED display at the back of the machine. Anybody have a set of boards real cheep for this machine which I could swap-out and see if it can be brought up? If swapping boards doesn't work then for sure the disk unit is shot and I'll have to scrap the darned thing and pass the boards back to the original seller if need be or others. No room around here for such unusable 130+ pound boatanchors :-( Regards, Chris F. NNNN Upon the date 12:40 AM 4/30/04 -0600, Bill McDermith said something like: Christian Fandt, Electronic/Electrical Historian Jamestown, NY USA cfandt(a)netsync.net Member of Antique Wireless Association URL: http://www.antiquewireless.org/

At 06:16 AM 5/3/04 +0100, you wrote: There are unless you got the optional Hi-Performace bit-slice CPU. One was called the LPU (Language Processing Unit) and the other called the PPU (Peripheral Processing Unit). The LPU handles the BASIC interpreter and the PPU handles all of the system I/O. Go read my HP 9845 webpage at <http://www.classiccmp.org/hp/hp9845.htm>. Interestingly HP found that the dual CPUs didn't improve perfromace very much so they dropped one of them on the 9835 and found the performace was almost as good. You can read about that on the 9835 page. For more information about the HP 9825 go look at my HP 9825 page at <http://www.classiccmp.org/hp/hp9825.htm>. See <http://www.classiccmp.org/hp/hpdcalc.htm> for links to all the desktops models. Joe

At 06:45 PM 5/3/04 +0100, you wrote: That would be cool! I have four or five 9845s but I don't think any of mine have that option. I believe it was a hybrid. hard to find? Sheesh I find them all over. nearly every 9825 that I find comes with one. I just passed one up in a scrap yard a couple of weeks ago. Also passed up a couple of the RTC modules. The RTC modules with the IO cable are cool. You can interface them to external devices and trigger the devices or use the device to trigger events in the 9845. Joe

My 98036A serial interface is in a plastic case but the 98034B HPIB interface is in a diecast case. DaveB Christchurch, NZ ----- Original Message ----- From: "Tony Duell" &lt;ard(a)p850ug1.demon.co.uk&gt; To: &lt;cctalk(a)classiccmp.org&gt; Sent: Tuesday, May 04, 2004 3:09 PM Subject: Re: hp proprietary uP history (9825 uP?) snip a lot

[...] They shouldn't be any worse than any other HP machine unless they've been stored badly and the connectors have not been properly cleaned. They use much the same edge conenctors as everything else. Yes. You get the pin header on the PCB (with a loopback test connector normally). You have to remove that loopback conenctor and plug in the cable. It's a standard 0.1" pitch header, so the mating connector is trivial to find. None of mine came with cables, but that was soon cured :-) [The watch chip ans the microcontroller] Prsumably they needed the functionallity of the microcontroller when the machine was powered up for things like triggering external events, responding to the inputs, etc. But the micorcontroller, ROM, etc took too much power to be battery-backed so they had to add a lower-power clock for when the machine was turend off. A watch chip, suitably repackaged (it's in a 24 pin DIP IIRC) was the obvious choice at the time. If you have anything apart from the user-terminated one and the 085 floppy one, can you buzz out the connections and note down which links are fitted, please. I assume you have the 98032 user manual which explains the connections and links. -tony

Christian Fandt wrote: Two quick hits looking on google for "HP BPC Processor" -- clears up at least a couple of the questions relation to where this processor was originally used... http://www.cpu-museum.com/5061_e.htm -- this link specifically mentions this as the processor for the 9825/9835/9845 processors -- see also Joe R's reply to your message... http://www.hpmuseum.org/tech9825.htm A different HP 250 hit: http://www.ylem.co.uk/hp250.asp The eloquence page mentions specifically the HP250 as a follow-on to the HP 9845, so I would suspect the same processor set -- but so many things were re-invented over and over again at HP it's always hard to say for sure... Regarding the 64000: As I probe my memory, it seems to me that I remember that the BPC chips we used in the 64000 were specially constructed for us and left the BCD arithmetic chip off the hybrid, so there were only the BPC chip and the IOP present in the hybrid package... Just a vague memory... Bill

On 3 May 2004 at 16:19, Joe R. wrote: It's described in the August 1983 HP Journal article, "An 18-MHz, 32-Bit VLSI Microprocessor." Excerpt: "The heart of HP's new 32-bit VLSI computer system is the Memory/Processor Module. The central processing unit in this module is an NMOS circuit containing 450,000 transistors on a single chip operating at a clock frequency of 18 MHz. This compact CPU chip, which implements a 32-bit version of the HP 3000 Computer's stack-based architecture, is designed and microprogrammed to support multiple-CPU operations within a single Memory/Processor Module." That issue was dedicated to the hardware, and the March 1984 issue covered the 9000/520 software. -- Dave

Jay West wrote: Correct, but if my memory isn't completely gone, it had some of the 21mx instructions, some of the additional registers, like the amusing but less useful index registers, but not all of the 21MX instructions, and some of the ones that were included weren't implemented in the same way (or with the same bit pattern? Can't remember...) -- It seems to me that you couldn't just use a 21MX compiler to target a BPC, for example. It's just hard to remember, programmed it every day for a year in 1980, then moved on to the displays division in Colorado Springs, and never messed with it again... Bill McDermith

The 98x0 proxesor also does that -- you can indirect as many times as you like (and, indeed, tie the processor in a tight-ish microcode loop if you get an indirect address that tries to indirect through itself). AFAIK the 98x0 porcessor has a different instruction set to the 2100 series, but it's similar in concept. -tony