It's a 1968 HP9100A computing calculator. It was working when I put it in my loft 15 or 20 years ago but now has an ever increasing number of faults. Unfortunately the schematics for the computing engine appear to be unobtainium although there are some available for the model that followed this one but with a lot of significant differences. It has four memories. A "rope" memory holding the microinstructions that control operation of a ROM in the form of a linear inductive array. There are no ICs and all the logic functions are created on the motherboard in diode resistor logic. Conveniently all the diodes face the same way and all the connectivity is contained internally in the multilayer PCB.
Without schematics it was recommended that I trouble-shoot by shotgun tactics individually testing the 300 odd diodes and about the same number of transistors on the plug-in cards that write to and read from the core memory and an even greater number that access the linear inductive array. All this testing did reveal one faulty transistor in one of 40 discrete flip-flops and that did restore numeric entry but the computing and logarithmic and trig calculations still eluded me. Unfortunately the more I disassembled and reassembled the thing for test access the more problems it seemed to gain and I came to the conclusion that the edge connectors were the culprit. There are rather a lot of them and they are mounted on the multilayer motherboard so I don't think I'll be attempting to replace them and I think my only option is the subtly bend the contacts to ensure that they are actually sitting higher than the plastic walls that separate adjacent pairs of contacts. Connectors are similar to those below.
I did try giving the motherboard a cycle in the dishwasher (with dishwasher detergent) but it made no difference to the faults. I also tried subtly pushing each contact out and that made no difference either.
I was planning to repeat the contact tweaking but then decided to look at them under a microscope and discovered that my impression of the contacts hiding behind the plastic walls was erroneous.
Next, it was suggested that Deoxit has miracle effects of failing connectors so I placed an order for some last week but am still awaiting delivery. ?
I did try giving the motherboard a cycle in the dishwasher (with dishwasher detergent) but it made no difference to the faults. I also tried subtly pushing each contact out and that made no difference either.
I was planning to repeat the contact tweaking but then decided to look at them under a microscope and discovered that my impression of the contacts hiding behind the plastic walls was erroneous.
Next, it was suggested that Deoxit has miracle effects of failing connectors so I placed an order for some last week but am still awaiting delivery. ?
Peter
Doesn't your dishwasher have a salt reservoir in it? Mine does....
Well the Deoxit arrived so I sprayed every edge connector before inserting the mating board but, alas the faults remain unchanged. That said, I also sprayed the volume control of my bathroom radio that was decidedly intermittent and it did work magically on that.
I've an uncomfortable feeling that poor edge-connector contacts might just be "a feature, sir" with something of this vintage! We had to contend with a lot of this earlier-generation PCB and edge-connector equipped gear at work in an environment where it got a lot of temperature changes and being clattered around the country in lorries and it became an accepted thing to have to lift the lid on things and do a thorough, firm but careful re-seat before some devices would play properly (till next time!). I can't help thinking that edge-connectors were a bit of a bean-counters' delight in the sense that one side of the connector was virtually free (beyond a bit of tinning or gold-flashing) but that the very existence of the likes of DIN 41612 connectors is as a result of dissatisfaction with the edge-connection scheme. (Admittedly, 41612 allows high-density and multi-layer interfacing, but it still cropped up in simple, single-sided, low-density usage). It's notable that Racal's RA1772 receiver was archetypally something from the edge connector era but chose to use a hard-wired loom between its multiple signal modules, fault-finding being eased by allowing each module to swing upwards through right-angles for probing. My take is that Racal had already realised that edge-connectors were problematic and sacrificed a bit of maintainability for the sake of reliability.
Good luck with getting it to function consistently, though, I can't help thinking that if the edge connectors can be cleaned/treated to give more dependable contact, that will be most of the way to full functionality. I'm thinking that this is a Nixie-era device?- if so, at least it hasn't been vultured by the clock gang. Things like this need to be preserved so that you can put a current-production £10 Casio scientific calculator next to it and say, look kids, it wasn't so darned easy and cheap back then, a whole load of sweat and shekels for someone.
