On Feb 1, 17:31, Gunther Schadow wrote: There are two possibilities here. One is that there are board sets such as the RLV11 controller which use the CD-interconnect to talk to each other, and those won't like being in a serpentine backplane, where both A-B and C-D carry Q-bus signals. They're designed to be used a a straight backplane, where the Q-bus is only on A-B, and the lower side of each C-D is connected to the upper side of the next C-D row down. The other possibility is that some backplanes (notably some Plessey ones) have battery voltages or occasionally even AC voltages on the "spare" pins and they may fry boards that expect signals there rather than a power supply. Well, a better way of looking at it is that there's a design feature which is advantageous in providing prioritisation of interrupts according to position relative to the front (processor end) of the bus, but the penalty is that you need grant cards as a result. It's the interrupt chain. In -11 busses it's designed so the chain passes through each device, and therefore any device can be built with the capability of locking out devices behind it while it's being serviced. Anything further away will wait until the locking device releases the interrupt, then it's own request will automagically be seen by the processor. There's no such scheme in an Omnibus, as the signal goes "past" each card rather than "through" it. The grant cards only need to carry the interrupt request and NPR lines; these are the only ones that have an IN and an OUT on each slot. Other signals just daisy-chain the slots together like a busbar in a power distribution panel. The other way is to use a wirwrap on the backplane between the IN and the OUT. The NPR line is often dealt with in that way, and G727 grant cards actually only jumper the four interrupt lines (G7272 jumpers NPR as well). There are a few reasons you may only have grant cards in some slots. It's quite possible that something has been removed and nobody put a grant card in even though they should. It may well work like that, if nothing further back actually needs the signals. Another possibility is that you're looking at some custom backplane which doesn't have standard SPC slots -- not all Unibus backplanes are wired the same way -- though I can't think of one that has SPC in some slots and custom wiring in others. No DEC one, anyway. The Unibus signals fit onto two rows, A and B. On a normal expansion backplane, some of those signals are redistributed onto rows CDEF as SPC signals, so in effect you have two busses side by side -- Unibus (or MUD : Modified Unibus Device) in slots A and B, and SPC (Small Peripheral Controller, same signals more or less, but different arrangement) in CDEF. You can fit a dual-height Unibus card in rows AB, and/or a quad-height SPC card in CDEF, or a hex card in ABCDEF. If you really mean single-height (only one slot) or dual-height when you write 1x or 2x, those are usually cards which are parts of some device built from simple(ish) logic modules; they're not Unibus or SPC cards, and generally fit into specific places in a specially-wired backplane where only A+B of the first and last slot are Unibus. The rest are some custom point-to-point wiring for a particular device, built from simple modules. -- Pete Peter Turnbull Network Manager University of York