Products > Vintage Computing

IDing a mystery vintage digital card

(1/1)

D Straney:
Just wondering if anyone recognized the card from the attached images.  Some points:

* UNIVAC and CDC systems seem to have connectors rather than using gold fingers on the board; I've seen photos of DEC and IBM cards with gold fingers but nothing yet that looks quite like this
* No search results for "CM305" written on the front
* The flatpack ICs are quad NAND gates, from Fairchild's µDTL family, with what looks like 1969 date codes.It's definitely nothing high-performance, as there was faster stuff than DTL by then, so I'd guess it's not from a mainframe or even the actual processing logic.  It also doesn't look like anything aerospace/military (like my CP-901 card), which would justify the older tech but would have more considerations for conduction cooling.  So, I haven't traced out the schematic yet, but from the way it combines discrete transistors/passives and ICs, I'm guessing it's meant to drive some kind of load, either core memory or as part of a peripheral device, or something like that.

PA0PBZ:
Looks a lot like this: https://rcsri.org/collection/honeywell-316/

D Straney:
Awesome, good find, yes that's almost definitely it!  Industrial control makes sense as an application: doesn't need raw processing power, so non-bleeding-edge is fine, but just needs to be proven and reliable but without all the aerospace-level environmental robustness.

Pretty sure this is a relay or other load driver.  The circuitry is arranged as 8 Darlingtons, with a 100Ω resistor across the lower base-emitter, with the collectors broken out separately and the emitters tied to 2 common power grounds (one for each side).  Each collector (/output) has a catch diode to another common point, which is probably the load power supply.  The NAND gates are essentially used as inverters, with one input left floating (which is equivalent to a high input - see gate schematic attached).  Darlington bases are directly driven by the gate outputs.  Each output is driven by 2 NAND-gates-as-inverters with their outputs connected together (specifically recommended by the datasheet, and works because of the open-collector-and-pull-up config. of the outputs): one gate input goes to a unique pin (the control input), and the other gate input goes to a common pin.

At first I thought this common input was a lamp-test input that would unconditionally enable all the outputs, but realized I was being thrown off by the datasheet's "wired-OR" description of connecting gate outputs together, when it's really more like a "wired-AND" - either gate can pull the shared output low and keep the Darlington off.  So instead, this common input looks like a global enable that allows the individual outputs to be enabled separately.

The yellow cap in the lower-left is for filtering on the logic power supply.  The 4 blue caps are weird though: they connect to 4 individual pins on one side (and nothing else), and to the power grounds on the other side.

PA0PBZ:
You can see a few CM305 boards in the pictures and according to the description they are part of the magnetic core control. Looks like the complete thing is what was later called a PLC.

Navigation

[0] Message Index

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod