(last updated: 22-Sep-2011)
This machine was rescued without any boards in it, except the System Industries 9700. I have acquired somewhat more than the minimum boardset (particularly the maximum of 8MB memory).
A suitable SMD Unibus controllers is still needed, to support my two Fujitsu M2361A "Super Eagle" drives.
| DD11-DK | A | B | C | D | E | F | |
|---|---|---|---|---|---|---|---|
| CMI | 1 | (empty) | |||||
| 2 | L0002 : DPM : Datapaths module | ||||||
| 3 | L0003 : MIC : Memory interconnect cache, TLB, Prefetch buffer, CMI | ||||||
| 4 | L0004 : UBI : Primary Unibus Controller + TOY clock, Console, TU58 UARTS, interrupt controller | ||||||
| 5 | L0008 : PCS : Patchable control store (1k x 80) | ||||||
| 6 | (empty) | ||||||
| 7 | Systems Industries 9700-6301 | ||||||
| 8 | (empty) | ||||||
| 9 | (empty) | ||||||
| 10 | L0016 : CMC2 : Memory controller, maximum 8 x 1MB | ||||||
| Extended Hex | 11 | M8750 : 1MB ECC MOS Memory | |||||
| 12 | M8750 : 1MB ECC MOS Memory | ||||||
| 13 | M8750 : 1MB ECC MOS Memory | ||||||
| 14 | M8750 : 1MB ECC MOS Memory | ||||||
| 15 | M8750 : 1MB ECC MOS Memory | ||||||
| 16 | M8750 : 1MB ECC MOS Memory | ||||||
| 17 | M8750 : 1MB ECC MOS Memory | ||||||
| 18 | M8750 : 1MB ECC MOS Memory | ||||||
| Unibus | 19 | M9202 : Unibus backplane interconnect | |||||
| 20 | M9202 : Unibus backplane interconnect | ||||||
| 21 | M7792 DEUNA : Ethernet controller (port) | ||||||
| 22 | M7793 DEUNA : Ethernet controller (link) | ||||||
| 23 | (empty) | ||||||
| 24 | (empty) | ||||||
| 25 | M7819 : DZ11-A : 8-line Async | ||||||
| 26 | M7486 : UDA50 : SDI disk controller (disk) | ||||||
| 27 | M7485 : UDA50 : SDI disk controller (bus) | ||||||
| 28 | M9313 : UNIBUS exercisor/terminator | ||||||
04-Oct-2007
The donor had contacted Richard, a member of the cctalk@classiccmp.org mailing list, who posted her message that a VAX 11/750 was available, free for pickup. I was lucky enough, I guess, to be first in line and was given the opportunity to take the system. During the pickup event, Lori mentioned that she had experienced a flood of interest, from as far away as Europe, begging for the machine to be saved for 'somebody'. She was somewhat in awe with this level of concern for the system.
04-Oct-2007
I got a chance to peek inside the cabinet, hoping to grab a quick inventory of the boards. To my horror, I found only a single board in the backplane! In (CMI) slot 7 is a System Industries 9700-6301, which Ethan (cctalk:) suggests might be a SI-9900 interface. Well, I don't even know what SI-9900 is!
A second board was found found loose in the base of the cabinet, and appears to be for a Qbus. It is an MDB MLSI PC-11. Ethan suggest that it might be a paper tape reader/punch interface.
Other than those two boards, the system is gutted. There are a number of ribbon cables coiled up, and the end of what looks like a SDI cable set (for RA81-type drives) but I haven't gone through all of that. I'll do more inventorying when I tear into the system for cleaning.
08-Oct-2007
After my post to the cctalk: group, I've had a contact from three other list memebers with offers of some boards, maybe. I've asked them to examine what they have in their spares box, and if they feel that they can let something go, I'll be interested.
I've sent my list of requirements and we'll see what turns up!
Last night I spent some time searching ebay for the boards, and lo!, I was surprised to see so many actually. Most of the L-series boards, are going for about $200 each, but the 1MB memory boards are offered currently at $30 each, which isn't so bad.
20-Oct-2007
Big day, today! Time for cleanup and possible relocation to my basement work area.
A few general comments are in order before the details. This unit was kept in a rented storage unit so it has suffered certain extremes in temperature and possible infestation by insects and rodents. As it turns out, mice have at one time taken up residence in the system. I could see the tell-tale signs when I first opened the front cover. So, I was kind of expecting a messy time of it, and I was right!
The system had a lot of mice droppings, and significant corrosion of the bottom tray and other parts where the mice had made a home. In fact, one of the power supplies was completely stuffed with dry grass and other bits used to make a cozy nest. This, unfortunately, is not a good thing. I'm probably going to completely disassemble both power supplies and the power controller, and run the PCBs through the dishwasher.
As usual, I used my normal process of vacuuming what I could, first, then wiping, and scrubbing as required, with Windex. I have cleaned up the main rack after removing the side panels and top panel. There are many spots of corrosion that I cleaned, but they still look dirty.
I wiped the backplane box inside and out, and all other surfaces of the bare-naked rack. I cleaned the various storage and IO cables that are just hanging there in the left side of the cabinet. I hung, for now anyway, the SDI cables on a bracket just to take tension off of the single pre-existing wire-tied hanging point on the welded wire above the backplane.
31-Oct-2007
After three or four weeks, I have not had any responses from any other classiccmp'rs that offered to help with the boards.
In the meantime, I have noticed that these boards *are* available from sources on ebay. Last night, I used Buy-It-Now for (x8) M8750 1MB Memory boards at $30 each from 'emisolutions'. Today, I used more Buy-It-Now auctions to get the L0002, L0004, L0006, L0008, L0016 boards; and the M7819, M9313 boards.
So, as it stands now, here is the list of desired boards (x indicates ones ones that I've acquired so far):
L0001 FPA (opt) x L0002 DPM (60+12 emisolutions; guruboards auctioned but didn't have it) x L0003 MIC (40+12 guruboards) x L0004 UBI (60+? emisolutions) x L0006 RDA (60+? emisolutions) x L0008 PCS (60+? emisolutions) x L0016 MMC (20+8 guruboards)x M8750 1MB x8 (30+? emisolutions)
x M7819 DZ11-A (30+? emisolutions) x M9313 Uterm (30+? emisolutions)
[one or more of:] x M7485+M7486 UDA50 SDI controller Dilog DU256 SMD controller (emulates MSCP controller) Emulex UD33 SMD-E controller (emulates UDA50, supports Fujitsu 2361A) Emulex UC12/13/14 SCSI controller (emulates MSCP controller)
[one of:] x M7792+M7793 DEUNA Ethernet controller M7521 DELUA Ethernet controller
Power key: XX2247 "Code Cut 7-pin round"
10-Nov-2007
One of the first things that I want to do is copy the four TU58 cartridges that were found inside this system. I'm hoping that I can cable the TU58 drive to a PC or a PDP-11 and just dump the cart contents without too much effort. I've downloaded four documents from bitsavers.org relating to the TU58 and am reading up on the drive and controller.
I physically removed the TU58 drive. To do so I removed the grounding spade from the drive's lug, then disconnected the individual brown wires that connted to the front-panel LED. They are keyed (one being male and the other female) so reversing polarity on reassembly won't be a problem. I then disconnected the cable from the controller. With the connections out of the way, I removed the (4) slotted hex-head screws from underneath the drive to release it from the molded bracket that it hangs from.
I then cleaned the drive in my usual way of the accumulated dust and mouse residue. Checking the puck roller, it is firm and solid, not gooey as other classiccmp.org members have seen. This is fortunate, since I won't have to trouble with replacing the rubber roller.
I removed the controller's power and serial data cables, then unscrewed two small screws that hold it to the mounting bracket. I lifted it straight up through the top opening of the rack. It is quite clean, having been protected in it's upper corner in the rack.
After becoming familiar with the controller layout, and having the drive cleaned, I returned it all to the rack. We're now ready to go.
After some deeper study of the TU58 Technical Manual, particularly page 4-8, I believe the controller is configured this way:
RS-232 Selected (WW17-WW19) RS-423 Driver (WW18-WW20) 19.2k baud (WW8-WW10-WW11) 8 data bits 1 stop bit no parity
The J2 pinouts are:
pin 1 - AUX A (not used) pin 2 - GND pin 3 - XMIT+ pin 4 - XMIT- pin 5 - GND pin 6 - nc pin 7 - RCV- pin 8 - RCV+ pin 9 - GND pin 10 - AUX B (not used)
The connector, according to the TU58 Technical Manual, from the cable's view, looks like this:
+------------+ | 9 7 5 3 1 | | 10 8 x 4 2 | ================
I believe a 'null-modem' cable for use with a PC would be:
TU58 J2 IDC10 PC DB9 PC DB25 --------------- --------------- --------------- pin 3 (XMIT+) pin 2 (RCV) pin 3 (RCV) pin 8 (RCV+) pin 3 (XMIT) pin 2 (XMIT) pin 5 (GND) pin 5 (GND) pin 7 (GND)
I will build this cable and try it out! In the meantime, I'll also need to work on getting a power connection to the drive, separate from the whole VAX power supply. Perhaps I can use a normal PC power supply, since all I need are +12V, GND and +5v:
pin 1 (+12v) pin 3 (GND) pin 5 (+5v) pin 6 (GND)
12-Nov-2007
After further review of all the materials that I have, I'm ready to try out the TU58 drive, and if possible, get image copies of the 4 Tu58 DECtape II carts.
I made a cable with a DB25 connector (rather than a DB9) so that I could directly use it with a PDP-11 host. The cable is built with signals crossed (ala null-modem) so all I needed was a gender changer to connect up to the Darien PDP-11/23+. Fortunately, the baud rate on the PDP is already set for 19.2k so it matches that of the TU58 drive, and that should do it.
I connected a spare regular PC-style power supply to the controller, after pulling one of the ground wires. That wire (in position 2) would have contacted Pin 2 on the TU58 controller and I wasn't sure if that pin was also a ground pin or not. So to be safe, I pulled the pin/wire from the molex connecter to prevent contact. Later, I reviewed the docs, and it is a good thing I pulled that wire, since pin 2 on the TU58 controller has some flavor of +12V according to the print set! I took photos of this setup.
I powered up and saw the LED on the TU58 controller flash on, off, then stay on as expected, according to the TU58 Technical Manual. I inserted the TU58 cart labeled "SI4400B" and entered this command on the PDP-11 console:
.COPY/DEVICE DD: DL2:SI400B.DD
and it seemed to work! The TU58 drive spun the tape back to the beginning of the tape and then started grinding away. It copied to the RL02K pack in DL2:. While it was working, I cleaned the other three TU58 carts, getting rid of the mouse-iness!
Stupidly, I forgot to include the /FILES switch in the copy command, so I just overwrote my RL02K cart in DL2:! Argh! The command should have been:
.COPY/DEVICE/FILES DD: DL2:SI400B.DD
That was my boot RL02 cart, and had a specially SYSGEN'd RT11 that recognized the third RL02 drive. Ah well, I have images of everything, and I can get it rebuilt.
The one good thing that comes from this day's experience is proof that the TU58 puck roller is in good shape.
13-Nov-2007
After last nights successful connection and operation of the TU58 drive connected directly to the PDP-11/23+, tonight I rebuilt the 'destroyed' RL02K cartridge, and copied all four TU58K carts.
First, I had to re-SYSGEN the RT11 V4.0 on the RL02 disk, configured for 3 RL drives, instead of the defualt of 2 drives. I just used all of the SYSGEN responses taken from the SYSGEN documentation, except for the number of RL drives.
With that out of the way, then I started my copying of the tapes. I had no trouble with any of them, but I did copy each tape twice. After all had been copied (I now had 8 image files), I used BINCOM to compare the two images for each tape. They matched so I feel that I have good images. I transfered them to my /dimage directory on beelzebub. They can also be downloaded from here.
I posted my results to the classiccmp.org list, and had John Dundas respond with a request for the images. He has a very nice site with many resources, and he wants to put these images up there too. I posted them to rogerwilco.org and sent him email with the link.
15-Nov-2007
This system isn't ancient, and so the capacitors are of more modern construction. They probably don't really need to be reformed, but like always I will do it anyway.
I reformed the capacitors in three groups according to the voltages. I ganged all four (or five) capacitors together in series for each reformation sequence.
16-Nov-2007
With the capacitors reformed, and the power supplies cleaned up as best as I could make them, I re-installed the capacitors and reassembled both power supplies. One PCB I held out from the 2.5V supply. It has significant mouse residue on it, even after my vacuuming and surface cleaning, and I want to send it through the dishwasher to really rinse it clean.
28-Nov-2007
Today, I completed adding a 30A circuit breaker and L5 socket to my basement workspace, and now am ready to provide power to this system.
I made sure that nothing was plugged into the backplane, and made sure that the TU58 controller's power cable was diconnected, just to be safe while I checked the power supplies.
I plugged in the power cord and saw a good AC power indication. I then pushed on the circuit breaker handle, and rotated the power select switch from OFF to LOCAL. The power kicked on and the blower spun up. It makes a racket!
However, I'm seeing fault indications on the power controller. The '+2.5 FAIL' and 'REG FAIL' LEDs both light up, then several seconds elapse and the 'OVER VOLTAGE' LED will flick on and the REG FAIL flicks off for about 1/2 second, then the whole sequence repeats. Each cycle takes about 8 seconds.
After reading in the 'VAX 11/750 H7104 Power System Technical Description' I conclude that the +2.5V-at-the-backplane monitoring circuit is open, probably because there are no boards installed. After posting this to the classiccmp cctalk list, others suggest that it might just be the lack of load and suggest attaching one or more automobile headlamps to the supply to give it a load to work with. I need to find some headlamps!
28-Nov-2007
With the help of my friend, Bill Gillman, in a couple of sessions we isolated the fault in the 5V supply. It turned out that two of the rectifying output diodes were shorted, causing the crowbar to clamp down continuously. This explains the buzzing that was heard from the PS.
I web-searched suppliers and asked for quotes from a half-dozen of them. Two responded, one offering a minimum quatity of 15 pieces, and the other, Parts XXXXXXXX in Texas, was willing to sell 6 @ $30 each. I asked for a quote of six, even though there are only four diodes installed, with the idea that if we blow one or two of the new ones while we finalize our troubleshooting, I'd still have four total that could replace the originals. As it turns out, we did replace all four diodes, and the PS works perfectly now. So there appears to be no upstream faults that caused the diodes to fail; the faults were diode specific. This is the good news.
The bad news is that somewhere throughout the 5V PS efforts, the 2.5V PS failed. Bill and I shook our heads at that, but over the next several months, with a few delays for the holidays, etc., have still not been able to isolate the problem. We have the schematic, and Bill finally took the 2.5V PS home to work on it, but still no joy. The fault seems to be in the 2.5V CONTROLLER board, which is the board that has the most damage from the mouse nest. There is clearly noticable corrosion on a number of traces, though they have good continuity, but other components might have some internal damage. Bill has been trying to make sense of the complex and unusual power supply. It's rather convoluted, so we're still not having success.
04-Aug-2009
After a very long time, I put my hands on this system's power supplies again. I checked all the internal connections, closed them up, and cabled everything together on my bench (photos). No change, of course. When I power up, I get good 5V, but nothing from the 2.5V supply. I just emits a faint 'tick'-ing sound about once a second. No change from where Bill and I left this a year ago.
I have, in the meantime, read up on the Theory of Operation section of the PS documents, and I still want to try to track down the problem. I've been reviewing my NRI Electronics Engineering Technology course from the mid-80s to rebuild my little bank of electronics knowledge, hoping that I can at least check currents, voltages and whatnot with my VOM and oscilloscope. Better get ready, because clearly there's something that needs finding!
28-Aug-2009
In the past week I have read portions of the H7104A Technical Description, and it plainly indicates that the 2.5V supply requires the 5V supply in order to operate. Also, further web searching suggests that not only do the two supplies require a load to operate, the 2.5V unit has a backplane sense connection, and I suspect that the 2.5V supply might behave the way it does (on the bench) due to this missing connection. I am now determined to re-install the power supplies and controller into the cabinet chassis, make the connections to the backplane, load the main outputs with automobile headlamps, and try it.
28-Aug-2009
I've done a little research to see what other module I need to hook up the two Fujitsu M2361A "Super Eagle" drives to this system. There a quite a number of options for SMD drives (may not support the Super Eagles though):
Unibus:Dilog DU236 Dilog DU256 Emulex UD33 (SMD-E, supports Fuji Super Eagle) Emulex SCxx
CMI:
Emulex SC700x
30-Aug-2009
The power controller and power supplies were re-installed in the cabinet. I connected the main power bus cables for the 2.5V and 5V supplies, plus I connected all auxillary output and input cables. For testing, I also temporarily connected my automobile headlamp bulbs to the 2.5V and 5V outputs, four filiments each. Beforehand, I measured the resistance of each set and found it to be just 1 ohm. So, this should draw 2.5A and 5A respectively. This may not quite be a high enough load, but I try it.
With everything cabled up I powered up. I saw the 2.5V lamps light! But, the 5V lamps did not, with the controller showing the "+5 Fail" and "Reg Fail" lamps. After about 5 seconds, the power controller circuit breaker tripped and shut everything off. Hmm, interesting. But, at least the 2.5V seemed to work!
So, I started to think about what the problem might be. I'm guessing a short across the 5V leads could do it, but on second thought, that would just result in the crowbar shunting the output and I should hear the high-pitched sound that was emitted so long ago when I originally tried the system which had bad output diodes. I'm now thinking it's something else. Perhaps a short on the auxillary lines?
I disconnected the +/-15V plug (P4) from the connector (J3) on top of the 5V H7104D box and tried again. This time I get 5V output! But again, after a few seconds the power controller circuit breaker trips and shuts off the system. I wonder if I have a short on the +/-15V lines. (Later I learned that the airflow sensor uses the 15V line, thus when the 15V line was disconnected it allowed the system to start. See next paragraph.)
Before doing that, it occurred to me that since the power controller is tripping the circuit breaker, it might be some other issue that the controller cares about. Considering that I don't have the blower plugged in, it might be the airflow sensor that is triggering the shutdown, due to lack of airflow. I plugged in the fan and powered up (still with the +/-15V connector off). Sure enough, I have good lights on the 2.5V and 5V headlamps and the power controller shows no faults!
Measured voltages at the supply outputs are:
2.53V 5.21V
and at the +/-15V connector on the 5V supply:
-15.19 +15.32
Checking the P3 connector at the backplane (which has etched expected values, conveniently) show all right on. I'm seeing:
+5.08V (+5VB) -4.99V (-5VB) +15.32v -15.19V +12.10V
On the power controller itself, I see no lights other than the main AC "Power" lamp. I DO NOT see the "Power OK" light. Maybe it's burned out, or maybe that only lights when the various PCB circuit boards are installed.
On the front panel of the chassis, I see "Run" dimly lit and "Error" fully lit, but the "CPU Power" lamp is out.
To further check the general operation of the power system, I switched the power controller into "Remote" and used the front panel key to power the system. It works fine.
Copyright © 2006-2025 Jared Blaser. All rights reserved.