The Saga of the Sting Reserve Power Pack…

[[kby]]

…aka the HP-10.

I've finally managed to put together a working HP-10 with decent cosmetics. Third time was the charm, so I thought it was time to collect the learned experiences here so others don't have to.

I actually have 1.8 working HP-10s. The 0.8 is one that needs a prettier back label (more on that here and in another thread I'm going to eventually start on labels) and a new printhead: the top row usually doesn't show up, although it occasionally has. Measuring resistance also is intermittent. I don't see an obvioius cable problem either.
Cost-wise, I did overpay on the last one, and the previous ones weren't that cheap, either. However, I think I still come out cheaper than the one currently on TAS where he quotes from this meseum site, but seems to ignore the price range completely. We'll have to see what happens. I probably contributed to opening the flow of machines, but not that much. I've also gained what is at least to me a fair amount of valuable experience.
But, back to the beginning. HP-10 #1 is the one I previously wrote about. It came as part of a bundle of two non-working HP-19c and the HP-10 which was also not sold as working. The claim was that I could probably put together one HP-19c from the two parts—one suppsedly was working, but someone decided to fix what wasn't broke (at least not completely) but trying to superglue the famous gold contacts together, and then tried to take it apart with solvent. That had about the effect you'd expect. However, I was really only interested in the HP-10 anyway. Unfortunately, that added to the cost.
The cosmetics on this HP-10 were OK, but had the usual suspects: a small amount of worn silver trim and a few marks on the back label. It also had a crack near the connector for the power supply connector. Internally it had a pretty bad "green corrosion" case, so, after the acetic acid bath, many traces were destroyed.However, after a bunch of work, I did rebuild the traces. However, the trace rebuild kit I used was the older epoxy type so things were sealed in epoxy…which is fine, but unfortunately if something breaks again and needs to be redone, it is a whole pain in the…which is where it now stands as I re-damaged some of the repaired traces in flexing the board while using it to try to get HP-10 #2 to work. I did get that calculator to working as well as it could, but encountered the bad CPU chip that was detailed elsewhere (the 5 and 8 keys which are tied to the same CPU pin did not work; the keyboard was fine.
This brings me to my admittedly amateur assessment of working on the HP-10:
It's a real pain, and here are the reasons:

1. The construction is a three-board construction (plus the display board, in a three-layer sandwich configuration. It shares all of the issues of the HP-19c and then adds one more. Like the HP-19c, the bottom board is the power supply board. To take this board out of the case requires flexing the board a fair amount due to the way it's installed, so that is one problem. Another effect of this is that there are two guide posts that help to hold the board in place, but you have to flex the power supply board over these. Inevitably the post on the power connector side gets snapped either in long term use of the calculator (such as a drop) or upon removing the power supply board. Both HP-10s #1 and #3 have this issue. It's not a deal-breaker, but it is an annoyance. The power supply board connects to the CPU board through two sets of gold fingers, similar to the HP-19c, so you have to get both sets of fingers lined up when putting the boards together. This isn't impossible, however. Both calculators have the display board connected to the CPU board through yet another set of gold fingers, but this is not a real issue for either calculator as they are readily accessible and have no dependencies. The difference with the HP-19c is that the 19c connects the CPU board to the keyboard through a short flexible ribbon cable. It's possible, although requiring some manual dexterity, to connect the two sets of fingers connecting the CPU board and the power supply board while the keyboard has already been connected to the CPU board. You can then just snap the case halves together after everything is already connected.
The HP-10, in contrast connects the CPU board to the keyboard (at a point which is hidden from view by the case sides) with another completely independent set of gold fingers, which basically have to be lined up by feel and with no visual feedback, as the case halves are snapped together. One pin out of sync and you've just damaged that pin. Of course, some amount of force is necessary to snap the case halves together, so this is a difficult task. The way I've found is the best is to line up the top edges of the two cases, then begin to swint the top case down until the fingers make contact. If they are not lined up, you hear the sound of them scraping—it's kind of like if you plucked the musical part of an old wind-up music box. Slide the top case around (usually it's forward and back) until this sound stops and you feel that every pin is in place. Then, and only then, snap the bottom halves together and everything should be connected properly. Note that when you took the calculator apart, you would have done it like the 19c so the top edge would have been opened first and then the bottom edge separated. For the HP-10, this will usually mean the CPU remains attached to the keyboard and detached from the power supply.

2. While the HP-19c is the obvious more glamours sibling of the HP-10, it turns out that its first cousin the HP-27: It is (afaik) the only other NMOS HP calculator. This explains why its power supply is different (5-tap t ransformer), and, if one is astute, why the signs on Vgg and other voltages are reversed. This presumably explains why it also has a chip which is not listed in Eric Smith's chips pages: the 1858-0056 quad PNP transistor array that is used for the printer motor and head drivers instead of the 1858-0044 quad NPN array that is used in the HP-19c, HP-91, and HP-97. This chip doesn't seem particularly failure prone, but it IS mde of unobtanium, at least as an OEM part: there are no ther machine to scavange it out of of, similar to the unique CPU chip.

I only discovered the NMOS connection when I looked at Tony Duell's schematics where it is noted. This is the only place I've found this documented in a publicly accessible way.

HP-10 #2 was also sold as a non-working machine. Its cosmetics were better and it did not have as much corrosion damage (although it still required the acetic acid bath). However, at one point, it seemed I was going to be able to get the repaird power supply board from HP-10 #1 to work better in it. In the meantime, another issue with the HP-10 that is similar to the HP-19c is the brittleness of the wires that connect from the power supply board to the printer (two power wires and two printer carriage home position switch wires) and the two that connect to the external power brick socket. With any amount of the maniuplation that is required to remove the power supply (such as for soaking), one or more of these is likely to break either at the board end or at the printer end (they don't seem to break at the power socket).
I first got HP-10 #2 to display the 0.00 display without paper. This can be seen with jumper wires across the power pins of the CPU-keyboard connector so the keyboard doesn't have to be reassembled: pins 1 & 2 are Vbatt and the power switch connects them to pin 3, which is Vcc. With no keyboard connected, the "DISP"/Both/Print mode switch is effectively in the "Both" position, so it is required that the printer at least think it printed 0.00 and the appropriate number of blank lines has been printed. If the printer doesn't do this, then you'll never get to 0.00. I am not sure at this point whether the this needs to be the case if the mode swith is in the DSP mode. I thought at one point that the printer always printed the 0.00 on power-on even when the calculator is in Display Only mode, but that was not the case with HP-10 #3, and if it was true with HP-10 #2 as I thought I'd seen evidence for, at the current point in time #2 and #3 act identical in this respect (no printer movement on power-on in DISP mode unless the printhead is not in the home position).
At this point, unfortunately, when I actually loaded paper, I found that the printer printed very wide characters and in fact printed the end of the line on the return pass (overprinting the earlier part of the line). Ugh.
While the printer is physically the same as (and thus interchangeable with) the HP-19c's printer, it is run in a dumber way. I believe I read elsewhere on this site that the HP-97 printer prints bidirectionally whereas t he HP-19c printer is unidirectional. While this is true, the HP-19c seems to be able to at least approximate when the printhead is on the return vs. the forward (printing) stroke and accordingly drives the return stroke much faster during the return pass. The HP-10 has no such logic. The scheme is a that there is a MOTR transmitted when the printhead is supposed to move. There is also a BRK signal which instructs a transistor wired across the motor supply wires to short part of the current to GND, thus affecting the speed of the printer. I was able to get this printer to print properly by slowing down the printer, but at this point it is at the minimum speed, which is very slow indeed (maybe about 1.5–2x as slow as HP-10#3 which seems about normal and is slightly slower than the demo movie the guy who wants mucho bucks on TAS posted. HP-10 #2 came with the original box.
The CPU board has a potentiometer which can adjust the speed within a limited range. There appear two be at least two layouts for the board: one has adjustment pots scattered across the CPU board that have normal pot-sized screw adjustments; the other style has all five pots along the bottom edge with large gear-like adjustment knobs. HP-10s #2 and #3 have the latter style. The former style was pictered elsehwere on this site when I talked about the bad CPU chip
HP-10 #3 had almost perfect cosmetics, and came with the original box and a manual. The label was pristine, but, unfortunately, the paint on labels post HP-67 seem to age to a much more fragile state. They tend to nick much easier and also, if you touch them up and have to remove what you hadded (with isopropanol), you tend to have a problem with I term "paint cancer" in that larger flakes of paint come off rather than just the wet paint you're trying to remove. This seems relatively rare on Classic machine labels, but seems much more common from the HP-67 label on in my experience. I plan on a labels thread elsewhere to discuss this. Both the HP-19c and HP-10 have very thick labels, so they are not prone to the wrinkling issues, but the paint does overall seem more fragile, at least when its old. The machines often come with them in fairly good shape so I assume they held up reasonably well when new, but each time you put it down, you risk making another tiny nick in the paint, and as detailed above, you can't make multiple pssses at repairing as safely as I ahve with other labels. Anyway, my working HP-10 actually has the label from (I think) #1 which does have a slightly rounded lower right corner, although this is not normally noticable as its a black-on-black issue.
HP-10 #3 did require acetic acid, but the corrosion was minimal and I managed to not break any wires. The lower right had (from the upside down position) post is/got broken when romoving the CPU board, but was satisfactorily (if not perfectly) reglued.It had the printhead that seemed to work early on, except it was a bit faint on the bottom. Then the bottom worked find and the top seemed to constantly fail—except for a few occasional times. Adjusting the height did not help. I finally ended up swapping with the printhead that originally was in HP-10 #2 (hence #2's less than perfect printhead currently.
I will edit or add to this post later as I accumulate the photo or remember more issues encountered while rebuilding.-kby