You're in luck. The 11/45 was the first '11 that I obtained, and I have the manuals for it (printset, maintenace manual, etc).

Hello Classiccmp. I recently acquired a nice, rust-free, mostly-original 11/45 from an old lisp hacker in Los Altos Hills, CA. He told me that it had been running the last time he'd powered it up, which was in 1984 IIRC, but sadly, something has gone awry and it won't run now, though it seems to be close. Here's a brief summary of its status: It was originally a five-rack system with each rack containing the controllers for the housed peripherals. The bus was chained through to each rack on a BC05 and terminated with an M930. I only took the processor and the TU-10 racks, so to make troubleshooting easier, I reduced the system to only the CPU and memory boxes (memory is a 220KB box from Monolithic Systems that has unibus in and out). The processor has 12 KW of core as well, so I guess I could cut the bus down a bit more for the purpose of debugging. Mode of failure is such: I powered up the shortened system and the front panel lights lit [excitement]. Then I tried to examine some memory locations via the panel, but when I hit HALT, the RUN light stayed lit. I tried every combination of switches I thought was appropriate, but the RUN light wouldn't go off. (Sounds like a simple bus error, really.) Upon realizing that something was amiss, I decabled all the power connectors from the logic and tested the voltages, which, surprisingly, were all about 6%-9% above spec (and I'm not too sure that my VOM is accurate even though it's analog). I didn't readjust the voltages, but reconnected the power cables to the logic boards and retested the voltage levels. All stayed about the same except for the -15VDC. It dropped to about 2V. Same reading on both -15V regulators, by the way, and they seem to be connected to different areas of the backplane, but I'm not sure. There seems to be a high-pitch squeal (that sound of a laboring regulator) coming from somewhere in the power supply cluster, but there's no smoke or crackling sounds. I don't remember noticing the squeal with the power cables disconnected, but that may be a subconscious hope and not a fact. So that's my predicament. I don't have much time to mess with this beautiful old girl, so I hope someone here can increase my repairing efficiency. Help me save this great piece of computing history. In return, I'll at least give you an account on it if I get it to the point where it'll run 2.9BSD or V7. I have SMD interfaces and disks, etc and everything else I need (except a DELQA) to get it hooked up to the net. Thank you in advance. jake

You're in luck. The 11/45 was the first '11 that I obtained, and I have the manuals for it (printset, maintenace manual, etc). It can be got to run again. I am certain of that. I am going to assume that you don't have the KM11 maintenance board, useful though it would be. You don't need _any_ memory to start with. As this is an 11/45, the core can't be on the processor backplane (which has special slots for MOS/bipolar boards only). So stick the M930 in the top of the rearmost slot (slot 28 IIRC) of the CPU backplane for the moment. You'll not be able to do a lot, but the CPU will at least talk to the frontpanel in that mode if everything is working right. That is a classic indication on the 11/45. If one of ACLO or DCLO (I forget which) is asserted, the CPU microcode goes into a tight loop and won't listen to the panel. This is either something on the bus pulling one of those lines, or a fault in the PSU/wiring. Yes, the old power modules will rise a bit off-load. That is nothing to worry about. Which version power cableform do you have? Old machines have a distribution PCB inside the CPU box with 8 pin mate-n-lock (flat) connectors on it. Later machines have a distribution panel on the frame at the top right of the CPU with 6 and 15 pin connectors on it. I have manuals for both types, BTW. This you aare going to _love_... There are 3 -15V sources in an 11/45 PSU. There's an H745 regulator in both power supplies (and they do, indeed feed different parts of the machine). There's also a -15V feed from the power control board in the lower PSU (this board generates what is normally considered to be a +15V rail. On the lower PSU, the +ve side goes to ground, and the -ve side to the rest of the machine). There is also a +15V supply from the top PSU power control board. This is needed by the H745 regulators. Is this pressent and correct? Looking at the schematic, it appears that if DCLO is asserted (low), then Q3 on this board will turn on. This will turn on Q2 and shut down the +15V regulator. Aha.... But only if DCLO is asserted by the PSU, not by anything else in the machine. Start on J4/P14 (12 pin mate-n-lock) on the top PSU. On the transformer housing at the front of the PSU, not on the regualtor blocks. Pins 2 and 3 are the +15V output (They're linked together inside the PSU). Pin 12 is DCLO (this is pulled up by resistors on the bus terminator, so you must have the rest of the machine connected). Pin 8 and 10 (NOT commoned inside the PSU) are the ACLO pins. Let me know how you get on. As I said, I have all the schematics of this machine (including the PSU), and I've repaired mine on the odd occasion... -tony

OK, it _is_ an M930 terminator and not an M9302, then. I was going to ask you about that. The reason is that the later M9302 board can 'lock the unibus' if there are grant problems. An M930 won't.

Sounds perfect to me. And yet _both_ H745s are giving no output?

There should be an H745 in the rear slot of the top PSU. It has an 8 pin connector on top. Pin 1 is the -15V output. Pins 4 and 5 are a +15V (control-ish) input.

I think _all_ 11/45 systems have the 8 pin connectors on the CPU backplane. The difference is how the extra System Unit backplanes get power On old machined (pre S/N 2000?) there was a PCB on the inside back of the CPU box. It had a number of 8-pin flat connectors on it. The system units plugged into that. On later machines there was a board on top of the CPU fram on the RHS (over the 'hole' that goes over the PSUs). System units had the conventional 15 pin and 6 pin connectors.

You mentioned that 'both -15V outputs are low' Which pins of which connectors are you measuring? At the PSU there are 3 -15V lines on my diagrams - pin one of P21 (top PSU rear slot H745), pin one of P25 (bottom PSU, front slot, H745) and pin 4 of P22 (bottom PSU, 12 pin connector on transformer housing). Which (any/all?) of those are missing?

Sure, but the thing is, the PS seems to be fine when there's no logic plugged in. It puts out -15VDC right on the money. When I connect all the mate-n'-locs to the logic cage, the voltage dies down severely. Doesn't that indicate a 'logic' problem? Or at least a short somewhere on the logic boards?

At 03:31 PM 31/08/99 -0400, you wrote: These are old switching supplies with outputs that vary by load.