On 6/7/05, John Allain <allain at panix.com> wrote: The most definite way is by looking up the handle number in a DEC handbook or (trustworthy) field guide on the 'net. By deduction, if it's a dual-height card, extended length (i.e., not approximately square), there's a good chance it's Qbus. All PDP-8 OMNIBUS cards are quad height, extended length. Unibus SPC (Small Peripheral Controllers) are either quad-height or hex height, extended length. Some older Unibus controller parts (from the backplane-per-peripheral era) are the exact same size as dual-height Qbus cards, so mere silhouette is _not_ definitive. If you see bus interface chips like DC005 and DC010, _that_ is Qbus. If you see DC013s, that's Unibus. If you see 74LS240s, it's probably Qbus. Beyond that, things start to get muddy (ISTR the National 8641 can sit on either bus, but we used to use it on our Unibus boards). There's a whole stable of bus interface chips that DEC used, some proprietary, some available over the counter (the 7438, for example). Determining which ones are in use is somewhat helpful, but not an absolute guarantee. The area around the interrups grant lines is different between Qbus and Unibus (and nothing even close to similar for OMNIBUS), so if you happen to have memorized what a Qbus and Unibus grant card look like, you can map that picture over the card you are looking at and differentiate one from another. One thing one learns from experience is how they are supposed to look. A quad OMNIBUS card has a very different component density than, say, a quad Qbus card. A hex OMNIBUS card might resemble, superficially, a hex Unibus card, but the OMNIBUS card will lack the BR-level selector (small board or dip-resistor-pack-looking-part) of the Unibus card. Also, the positions of the I/O connectors can be different (there's room to exit signals at the fingers on one end of an OMNIBUS card, and that's never possible on a PDP-11; plus even hex-height OMNIBUS cards don't use the sixth card edge finger - there's never any backplane socket there, so there's no connections - that's a dead giveaway). Keep in mind, though, that these visual cues may or may not positively identify a board as to its bus type, but certain characteristics are more telling than others. It was never meant to be particularly easy or hard to tell one set of boards from another - the DEC guys always had the right docs to tell them apart, plus they were familiar with the numbering scheme. If someone handed me an M7xxx board, I would think Unibus long before I thought OMNIBUS. Conversely, an M83xx board would make me look for OMNIBUS characteristics, but, again, that's only a guideline. As for Q22 vs Q18, besides the dates on the chips, one can whip out the Qbus spec and look for connections to the extra 2 address lines at the edge fingers. A no-connect pretty much guarantees that it's Q18. Qbus vs Unibus is trickier. Besides the location of the grant pins, and the BR selector, I can't think of any positively defining characteristic except the handle number or perhaps the presence/absense of a 'V' in the name of the peripheral, should it happen to be legible in copper at one edge of the board (DPV11, RLV12...) For that matter PDP-8 boards will have an 8 rather than an 11, but I doubt all options are clearly labelled. Hope this helps rather than confuses, -ethan

On 6/7/05, Paul Koning <pkoning at equallogic.com> wrote: That's the only way to be sure. Except that it _might_ be from a PDP-11 memory bus, _not_ Unibus (like the DRAM cards fo r the 11/70, 11/725, 11/730, 11/750...) or part of a peripheral (RK611) or OMNIBUS (compare the visual appearance of, say, a DZ-11 and an RL8A - they _are_ different, in distinguishable ways, but a novice might throw them on the same pile). About the only thing that is true is that "single or hex must not be Qbus". Everything else is up for grabs. That is a good technique for a first-order approximation. Sometimes, that's as far as you have to go. -ethan

Eric> Paul wrote: Eric> DEC made *many* hex modules that are not Unibus. There are hex Eric> Omnibus modules for the 8/a, and plenty of hex non-Unibus Eric> modules in the KL10 (PDP-10) and some early VAXen. There were Eric> also "extended hex" modules used in the KS10 and other early Eric> VAXen. Sure. I should have been more precise. The original question was: "how can I tell by inspection whether a board is Unibus, Q18, or Q22?" Given just those three choices, what I said is accurate -- hex can't be Qnn so it must be Unibus by process of elimination. paul

May not even be DEC, or for a DEC machine... Somewhere I have a thing which claims to be a 370/E. I am not sure who made it, but I am told it emulates the user-mode instruction set of an IBM370, but it's built from F TTL on boards of the same form factor as DEC Hex boards. -tony