On Sun, Oct 4, 2020 at 10:26 AM Brent Hilpert via cctalk < cctalk@classiccmp.org> wrote:
> On 2020-Oct-04, at 1:22 AM, Josh Dersch via cctalk wrote: > > > More mysteries while poking at the MDP-1000. Spent some time this > evening > > working out the rest of the signals on the power harness (I suspect > inputs > > for an LTC circuit and a "power good" signal, as well as something > > connected to a relay on the backplane, probably related to power > control). > > > > There are a lot of unidentifiable ICs on the main CPU logic board and on > > the backplane, mixed in with bog-standard 7400-series TTL. Curious if > > anyone has any ideas, as my searches and perusal of datasheets/databooks > on > > Bitsavers have turned up nothing. These are all TI-manufactured ICs, > 1969 > > manufacturing dates, with "SN48xx" and "SN63xx" part numbers (a few omit > > the "SN" prefix.) I'm wondering if these are just standard 7400 ICs with > > special codes; for example there are several SN4816's near the edge > > connector for the I/O bus, where a 7416 might (?) make sense, and from > some > > basic probing and following traces I think the pinouts make sense. > > (Everything's conformal coated so it's a real bear to beep things out...) > > > > Any ideas? > > I have run across TI ICs from that era with odd/unknown series numbering, > in particular the SN3900 and SN4500 DTL ICs. > Notably: > - by pinout they match up with standard DTL series ICs, > - I have only found these in equipment from one manufacturer: > calculators built by Canon. > > I received a solitary page of datasheet for some of them (by way of a > Canon service center many years ago), but I have never seen them mentioned > in TI databooks from the era, even in those sections where they list e.g. > "other products from TI" and proceed to list little known series and part > numbers. > > So an obvious guess is these were house numbering systems of standard > parts done for the purchaser/equipment manufacturer but with TI's format > scheme rather than a format specified by the manufacturer. Another guess > would be standard parts tested and selected for purchaser-specified > parameters, although that seems a little excessive for these cases. > Yeah, either of those options seems a likely possibility. It definitely seems like they were building this thing for bulletproof operation, so maybe it really is the latter. > > The 54/7400 series originated with TI in 65, I'm not aware of them > producing any other TTL series, other than perhaps second-sourcing some > other manufacturer's. > I guess that's another possibility - another manuf's TTL series, labeled > differently. > > Odd that this Motorola CPU is filled with ICs manufactured by TI. > Yeah, the irony of this was not lost on me. Other than the aforementioned linear ICs in the core modules, every IC in this has a TI logo on it. > > It's conceivable, although it seems less probable, that they're DTL rather > than TTL. > > On the whole, best guess would seem to be 7400-series inside. > Yeah, I'm guessing (hoping) 7400 as well, especially since there are actual 74xx-labeled ICs in here casually mixed with the 48xx and 63xx. A thought occurred, that if I get desperate I could remove a few of the ICs and attempt identification using the IC tester I have. Not going to do that unless I can't figure it out any other way. The good news is there aren't too many varieties. I haven't done a full inventory but there's not more than a dozen or so different types in here. > > -- > > On another issue, did you trace the +/-15V lines to the core > address/inhibit drivers? > Could some of the remaining wires from the PS be other core supplies - 15V > was a little low compared to most core systems I've seen. > > I did manage to trace the +/-15V lines to the drivers on the core memory. I don't believe any of the extra signals are other voltages -- three of them go to the CPU board, one goes to the aforementioned mystery relay, one goes to a tiny bit of logic on the mainboard (I suspect some manner of LTC), and the last goes to a trace that just dead-ends and goes nowhere at all. The very little documentation I have on the CPU itself suggests that the supply provided only +5 and +/-15. There are two sets of power rails for each; the first goes to the CPU, front panel, and two memory slots, the second set goes to the other two memory slots. I suspect that upgrading past 8K of memory required a power supply upgrade. - Josh