HP 9820A

VERY nice finds. Those are examples of the second series of HP desktop caclulators, known as HP9800s. The HP9810 is a 3-level stack machine. It's not strictly RPN in my opinion becuase there is no automatic stack lift/drop on most operations. It's simialr to the HP9100 in this respect. The HP9820 is HP's first algebraic notation calculator. There are 2 other machines in the family. The (rare) HP9821 is an HP9820 with a built-in digital cassette drive in place of the card reader. The HP9830 is a full BASIC language computer. Both are programmable, but have only a few built-in functions. Are there any ROM modules in yours (look on top behind the display)? The maths ROM, for example, adds logs, SIN/COS/TAN, etc. The machines can work without any extra ROMs, so not having them is not a disaster. The 9820 will have a built-in thermal printer (I think it's standard on that model), it's an option on the 9810 (Opt 004 IIRC). It is the same unit, so if you have no pritner in the 9810 you could move one from one of the 9820s into it. I'll answer that in a moment, since there's quite a bit to say. YEs, it's magnetic, yes it writes too. It puts 3 data tracks nad oen timing track on the card, you put one end fo the card in, then turn it round and put the other end in to get twcie the capacity (in other words it only uised half the width of the card at a time. It's a pity you didn't get any cards, those are not easy to find. You can get user manauls, the HP service manaul and some unofficial scheamtics from http://www.hpmuseum.net/ OI know a little abotu the insides of the machines. The first thing is that the 9810 and 9820 are vey closely related. About 60" of the internals are the same. The CPU is bit-serial, 16 bit (it's one of the few bit-serial machines you are likely to come across). Actually, that's a simplification. It's bit-serial for binary operations, but does BCD adition one nybble at a time. The CPU is 4 boards (I/O inteface, clock( and I/O state machine), control logic (microcode), data path (ALU and registers)). The CPU is the same on both machines. So is the PSU. THis is 3 boards, One is a 5V swithing regulator. One is +/-12V fopr the analogue electroncis and the ROMs. The last is +16V and +19V for the (PMOS) 11/3 DRAMs and +124V for the printer. These boards are under a cover at the right side of the machine, there are labelled testpoints for all rails apart from the +5V (which you can find on any backplane conenctor). The metal chasis (and mains earth) is conencted to the logic 0V rail and proivides a convenient place to clip the black lead off your meter :-) The printer and card reader are the same too, as is the main backplane. So what's different? The display and keybord obviously, although the keyboard operates on the smae peinciple in both machines. And it's not normal. In fact there's a littke pulse transsformer formed by spiral PCB tracks under each key. The transformers for 2 keys are conencted in inverse series. When yuo press a key you damen one transformer, causing the puse fro mthe other one to be no longer cancelled out. The main diffence thoug, is nthe memroy box. This is the metal box at the left side of the machine, behind the display, It contains the firmware (HP custom masked ROMs), one board for a 9810 (2K words) and 2 boards (all boards are electrically identical, just with differnet ROMs fitted) for 4K words on the 9820. Thern there are 3 boards of control logic, including the memroy addres (M) and memoory data (T) shift registers. Finally, boards of RAM. Most machines use 1103 1K bit DRAMs. I am told that there is a version of the 9820 using 2102 static RAMs, I have never seen it (and I want to). Basically what I do when I get one of thse machines on the bench is : Dismantle it. Remove top cover, keyboard assembly, all the boards, memory box, printer, card reader, etc. I often take out the rear panel and backplane too, cna clean everyting up Refit the backplane and rear panel, reconenct the pwer swtich (on the keyboard assembly) and apply mains. The fan should run. Check the AC voltages on the PSU board connectors. Fit the PSU bvoards and power up again. The PSU cover is part of the heatsinking, but you an run the machine for testing without it. Check all PSU output votlages. Fit the CPU boards, the display and the mnemoty box (do not fit the printer, card reder/controlelr or the keyboard data cable at this point). Power up. You will not get a display. You should do, but there will be logic faults. Unfotruantely as the display is scnaned by the CPU (under control of the firmware) there could be a fault just about anywheere. And now for the fun part. Assuming you want to repari it rather than boardswap it, you now have to work through the logic. Ithink you need several thigns to do this, others re desirable Need : The scheamtics, a good logic probe, the CPU microcode sources [1], and a brain that knows how the darn thing should work. Desirable : A logic anaylser [1] Ask me nicely if you want this Once you've got it working, you'll need to sort out the mechanical side. The roller in the card reader will have died. The pritner platten probably has as well. The fromer turns to crumbly dust which si nto a problem the latter turns to a sticky black goo that gets everywhere. nyway, I can tell you how I repair them, but you need a good workshop. Others do it with hand tools, but I am not convinced it's as reliable. This is a blatant advert in 2 senses, but you might consider joining hpcc. The point being that there's an ongoing series of articles in the magazine entitled 'Repairing HP9800 Desktop Calcualtors' that might be useful. I suspect you can get last year's magazines to get the first 4 parts if you join and ask the membership secreatary. Please feel free to ask for any information on these great machines -tony