OEM 11/73 Issues

[antiquekid3]

Hi all,

I recently saved a Marquette MUSE system from Memphis, TN. It came with an 11/73 CPU, a CDC SMD drive, a Fujitsu Super Eagle hard drive, and various other goodies.

http://i.imgur.com/3hUB3Nd.jpg

http://i.imgur.com/wDvFzWS.jpg

http://i.imgur.com/js8FW28.jpg

It was supposedly powered up in the early 2000s, so I figured it'd be safe to power it up now without doing much. Just to be on the safe side, I only powered up the CPU box and Super Eagle (assuming that it boots from that drive).

The 12V rail indicator didn't light up, but the 5V did. Then there was crackling and smoke. Turns out one of the varistors went out. I replaced that today and fired it back up. Both 12V and 5V measure fine with no cards in it. I put the cards back in and found that my 12V rail was out, again. I removed all the non-essential cards (all but RAM, CPU, serial interface, and drive interface) and sure enough, that fixed the issue. So one of the other cards probably has a shorted tantalum cap on the 12V rail.

I hit the boot button on the front panel after connecting a VT-220 up to it. I get a backwards '?' sometimes. Maybe my baud rate is off? It says "2400" on the card itself, but perhaps it was changed. I tried 2400, 9600, and 19200. In any event, I don't think there's much activity, as it doesn't seem to respond to keystrokes. Who knows. Maybe my terminal is bad too. Haven't checked this one out thoroughly yet.

Maybe it doesn't even boot from the Super Eagle first. It's an Emulex SMD interface. There are two Emulex SMD interfaces, one for the big CDC drive, and one for the smaller, but still quite large, Fujitsu. There's also a dual 8" drive, but it seems to have some sort of parallel interface card instead of what I would normally expect to see for such a drive. I believe there's much more to this system than just a bare OEM 11/73. It looks like an S-100 chassis with a bunch more cards in it.

It seems to have a TU58, a serial-based 3.5" floppy drive, the two 8" floppy drives, the two big hard drives, and seemingly lots of telephone connectivity. I don't have any tapes or 3.5" floppies; just two 8" disks and three CDC disk packs.

If anyone has tips on bringing this system up again, I'd love to hear them. I've never done -11 stuff before, just lots of -8s.

Manuals for the Super Eagle would be awesome. It occasionally spews an error code "18" which I assume means the operator doesn't know what he's doing.

Thanks!

Kyle

 

[gslick]

Are those Emulex QD33 controllers?

http://www.bitsavers.org/pdf/emulex/QD3351002-M_QD33_Dec90.pdf

The M8192 KDJ11-A has no on board boot ROMs so the system is probably configured to autoboot from one of the Emulex QD33 controllers, if that is what they are. If you pull the CPU and the two Emulex controllers and take hi res photos of the jumpers and DIP switches that will help figure out how things are currently configured and what device it should be trying to boot from.

 

[antiquekid3]

Are those Emulex QD33 controllers?

http://www.bitsavers.org/pdf/emulex/QD3351002-M_QD33_Dec90.pdf

The M8192 KDJ11-A has no on board boot ROMs so the system is probably configured to autoboot from one of the Emulex QD33 controllers, if that is what they are. If you pull the CPU and the two Emulex controllers and take hi res photos of the jumpers and DIP switches that will help figure out how things are currently configured and what device it should be trying to boot from.

I took some more pictures today of the system in various stages of being on.

http://imgur.com/a/2tSnU

After powering on the system, the only lights on are three of the four on the M8192. They are the three right ones in the picture (a little hard to tell thanks to the flash). The hard drive seems to spin up and lights up the "Ready" indicator. I guess that's good. I then press the boot button on the front panel, and all three lights on the Emulex board turn on, then just the left two, then just the right two, then they all go off. After a while, the M8192 lights up just the right one, the LED on the RAM board lights up, and all three lights on the Emulex come on again, if I recall correctly.

I briefly looked at the manual, and from best I can tell, this is supposed to boot from the Super Eagle. Maybe the drive has been wiped; who knows!

 

[gslick]

That board is a QD32, not a QD33

http://www.bitsavers.org/pdf/emulex/QD3251001-B_QD32_Nov85.pdf

Looks like the SW2 DIP switches are SW2-2 and SW2-6 ON and the rest of SW2 switches OFF. So that would be configured as the first MSCP controller at 17772150 with 22-bit addressing enabled and automatic bootstrap enabled.

To work with the QD32 automatic bootstrap the M8192 KDJ11-A needs to be configured in Power-Up Option 0 with jumpers W3 and W7 installed. You don't have a clear photo showing the jumpers on the M8192 KDJ11-A. It looks like there could be two jumpers installed, possibly W3 and W7 so that might be correct.

Unfortunately it looks like the QD32 does not have an on board configuration utility as the QD33 does so it might be hard to tell what the functional state of the controller and drive is.

Might be good to start with a minimal system, CPU, memory, console SLU only plus some way of getting XXDP diagnostics loaded and verify that those base components are working OK before moving on to checking out the SMD controllers and drives.

 

[antiquekid3]

That board is a QD32, not a QD33

http://www.bitsavers.org/pdf/emulex/QD3251001-B_QD32_Nov85.pdf

Looks like the SW2 DIP switches are SW2-2 and SW2-6 ON and the rest of SW2 switches OFF. So that would be configured as the first MSCP controller at 17772150 with 22-bit addressing enabled and automatic bootstrap enabled.

To work with the QD32 automatic bootstrap the M8192 KDJ11-A needs to be configured in Power-Up Option 0 with jumpers W3 and W7 installed. You don't have a clear photo showing the jumpers on the M8192 KDJ11-A. It looks like there could be two jumpers installed, possibly W3 and W7 so that might be correct.

Unfortunately it looks like the QD32 does not have an on board configuration utility as the QD33 does so it might be hard to tell what the functional state of the controller and drive is.

Might be good to start with a minimal system, CPU, memory, console SLU only plus some way of getting XXDP diagnostics loaded and verify that those base components are working OK before moving on to checking out the SMD controllers and drives.

Great, thanks for the tips! Looks like W3 and W7 are in fact installed.

Having never done PDP-11s before, are there some things I need to know about the Qbus? For instance, I notice that the card rails are alternating in color. Are these significant? Does it matter which side the CPU (or any other cards, for that matter) go? Right now, I'm just going on exactly how the system was delivered to me. I'm used to the Omnibus...things were so easy then!

I've got an RX-02 and interface card, in unknown condition, that I can try. But I don't know what I'd need to do to try to get it to boot from that. I have some RX-02 floppies too; I'll see if any of them have some obvious labels on them. These came from another system belonging to a friend of mine, who is unfortunately quite busy and can't spend a lot of time helping me with this.

Being that the KDJ11-A lacks a ROM, I assume the only means for ODT are through another board of some sort?

I'll photograph some of the spare cards my friend has and see if any of them could be useful in testing the base components.

 

[RSX11M+]

...Having never done PDP-11s before, are there some things I need to know about the Qbus? For instance, I notice that the card rails are alternating in color. Are these significant? Does it matter which side the CPU (or any other cards, for that matter) go? Right now, I'm just going on exactly how the system was delivered to me. I'm used to the Omnibus...things were so easy then!

I've got an RX-02 and interface card, in unknown condition, that I can try. But I don't know what I'd need to do to try to get it to boot from that. I have some RX-02 floppies too; I'll see if any of them have some obvious labels on them. These came from another system belonging to a friend of mine, who is unfortunately quite busy and can't spend a lot of time helping me with this...

I'm going to address myself to two issues:

• Card Slot types
• Board Types

It's a cautionary tale that's often needed around here. We should really develop a set of articles rather than post it anew each time.

Looking at your photos (very helpful BTW) one can see that you have a 9 slot "S" BUS. In this type of backplane, all slots are "AB"-"AB". They must be filled in an "S" pattern. When the modules being installed are DUAL wide, this is implicitly ok - as they must also be AB type.

There are also QUAD modules. (your photo shows one in Slot 2) I suspect this is a memory. It must be AB-AB compatible in order to live in that slot. (indeed, in that system)

Here's where we talk about physical danger. Some boards are ABCD. Your backplane doesn't have any ABCD slots. Installing an ABCD module in the wrong kind of slot can, and often does, damage them severely. The RLV11 is one such example often destroyed by such mishaps.

If you stay with only your original set of boards, you're pretty safe on that backplane. But when you introduce new modules, especially QUAD ones, be sure to research them specifically for compatibility with an AB-AB slot before installing them in your backplane.

~

Your RX02 controller module needs to be researched to see if requires an ABCD slot. If so, it can not be installed in this system.
​

Being a novice, you should go through this research exercise with any new boards you introduce into your collection. I would mark ABCD modules on the top edge in a way you'll recognize, for extra safety. (an inch of red nail polish?)

Alternating color guides are intended to help you not skew boards while installing.

 

[gslick]

Depending on the how that system is packaged it might take some work to verify which backplane you have. Most DEC brand Q-Bus backplanes have a label on the top of the frame around the card edge connectors. You would have to be able to look down on the backplane from the top to see that instead of the usual view from the rear of the backplane.

There are two main varieties of dual wide Q-Bus backplanes, those that have the Q18/Q22 bus in both halves of the backplane, and those that have the Q18/Q22 bus in one half and the CD interconnect in the other half. There are also some with a mix, such as those in the BA23 and BA123 which have Q22/CD in the first 3 or 4 slots and then Q22/Q22 in the rest.

In your backplane photo there are dual wide card in both halves of the backplane so I assume it must be the Q22/Q22 type. If it's a 9 slot DEC brand backplane I believe that should make it an H9275-A.

Download the EB-23144-18_QbusIntrfs_1983.pdf manual from Bitsavers and it has a section on the H9275-A starting on page 376. On page 382 Figure 6 shows the serpentine bus structure. The CPU should go in the upper left position and you don't want any gaps in the serpentine bus structure. If you remove a card you should either shuffle up the cards below it or replace the card with an M9047 grant continuity card. In your system it might have been OK to have that TCU card in the last position. I think that might be battery backed time of day clock the doesn't need bus grant.

The KDJ11-A implements ODT in it's microcode. With just the CPU, memory, and console card installed you should be able to get into ODT. Either by halting the CPU, or removing jumper W7 on the CPU to configure it in Power-Up Option 1 to unconditionally enter ODT.

 

[antiquekid3]

Alright, I've now arranged the cards in the serpentine pattern. It's definitely doing some different stuff now. The lights are constantly flickering on the Emulex card, but still nothing on the terminal.

I then removed the Emulex card, removed the jumper so that ODT would auto-start, and tried it out. The ODT light comes on, but still nothing is printed to the terminal. Both D1 and D4 are lit, which means ODT is supposedly running, and that the memory system is not responding. Taking out the memory board lit up D1, D3, and D4. The manual says that means that reading the location of the SLU timed out. I guess since the SLU card is still in place, it may have the wrong address.

I can't confirm that the address of the SLU is set correctly. The manual on bitsavers doesn't match my revision, which is an MDB DLV11-ED. I can't make sense of the switches. One must be for the interrupt and the other for the address. The third DIP switch is for the serial configuration, which is the same from best I can tell as the older revision.

And yes, the backplane is indeed an H9275-A.

I went through the spare cards around here. Here's the list:
M3106
M7504
M7546
M7555
M7606
M7609 (x2)
M7941
M8029
M8047 (x2)
M9058
Andromeda DFDC-11 (paper label says DMDC11) (x2)
Camintonn CMV-504
Clearpoint QRAM-2 SAB-1 (x4)
Clearpoint QRAM-11B GSB-2X2 (x2)
Dilog DQ-215 (x2)
Plessey 703340
Sigma DZV11 (x2)

I have no idea if any of these will help me in my ventures. I suppose it would help to have one of these RAM boards in a known configuration to start debugging the possible memory issue.

There's also a Micro PDP-11 in a tall tower, but it's in unknown condition, as well as some Unibus machines (PDP-11[/20] and PDP-11/34), and two MicroVAX IIs. I don't think the latter ones will help me though. I'll be tackling them next, once I get a little accustomed to the -11 architecture.

 

[RSX11M+]

Alright, I've now arranged the cards in the serpentine pattern. It's definitely doing some different stuff now.

Umm.. The module installation pattern in this photo was just fine

...how is it now?

...There's also a Micro PDP-11 in a tall tower, but it's in unknown condition, as well as some Unibus machines (PDP-11[/20] and PDP-11/34), and two MicroVAX IIs. I don't think the latter ones will help me though. I'll be tackling them next, once I get a little accustomed to the -11 architecture.

Ok - you've awakened a fear in me. Please be absolutely certain you do not intermix UNIBUS and QBUS modules or systems. Doing so is the fastest way to make hopeless junk of both.

I think I'll stand aside and let gslick work with you.

 

[antiquekid3]

Umm.. The module installation pattern in this photo was just fine...how is it now?

Ok - you've awakened a fear in me. Please be absolutely certain you do not intermix UNIBUS and QBUS modules or systems. Doing so is the fastest way to make hopeless junk of both.

I think I'll stand aside and let gslick work with you.

The module installation is now the CPU, SLU, memory taking up the top slots, from left to right, top to bottom. I have added and removed the Emulex below-left of the memory card when testing it, as to keep the serpentine pattern. One of those cards that came with the system has a 12V-to-ground short, so I am only powering up the bare minimum to test basic functionality. I have not yet proceeded with finding which card has the short, as it's not pertinent to my current tests.

I can assure you I know better than to mix Unibus (or Omnibus, for that matter!) and Qbus. I was only mentioning them because they are future projects in the -11 family, not because I think they would provide useful parts for my Qbus system currently (though I don't know the internal bus of the Micro PDP-11). Despite having never done Qbus stuff before, I did recall there being different bus patterns, so I wanted to make sure I got that straight before proceeding. You cleared that up very quickly, so thank you.

 

[RSX11M+]

Phew!

Thank you.

Until gslick gets back... I would pull everything but the CPU, 4-SLU Sigma, the QUAD Memory and the card in slot 9. Then work on getting TT0: (console SLU) to show a MicroODT prompt.

Once that works, then add back cards to a "Working configuration".

You'll have three simultaneous things to sort out...

• which port on the 4SLU card is set as TT0: 177560
• resolving the RS232 cable to you're connecting DTE to DTE (usually needs a null modem cable)
• Baud rate and word format

It will be all downhill from there.

Short?

 

[gslick]

You have some useful boards there if they are good working boards.

M3106 - DZQ11 - Q-Bus 4-line async MUX. Cannot be used as a console. Really only useful if you're running RSTS/E or 2.11BSD, or VMS and you want multiple terminal connections.
M7504 - DEQNA - Q-Bus Ethernet.
M7546 - TQK50 - Q-Bus TK50 tape controller.
M7555 - RQDX3 - Q-Bus MFM floppy / hard drive controller.
M7606 - KA630 - Q-Bus MicroVAX II CPU. MUST GO IN A Q22/CD SLOT (BA23 / BA123)
M7609 - MS630 - Q-Bus 8MB for KA630 CPU. MUST GO IN A Q22/CD SLOT (BA23 / BA123)
M7941 - DRV11 - Q-Bus 16-bit parallel line unit. No idea what you use this for.
M8029 - RXV21 - Q-Bus RX02 floppy controller.
M8047 - MXV11 - Q-Bus Multifunction RAM, 2 SLU, ROM sockets. -AA (4Kw) or -AC (16Kw) version?
M9058 - Distribution board for M7555.

The first thing you need to do is to sort out the minimal CPU, memory, console configuration.

If it was me, after that I might try adding in the M7555 RQDX3 and the M9058 cabled up to an RX33 equivalent 5.25-inch floppy (e.g. TEAC FD-55GFR) and get XXDP diagnostics booting and running off of the floppy.

For the absolute bare minimum you could try just the CPU and one of the M8047 MXV11-A cards, although there are enough fiddly jumpers on those that might not be the best bet unless you can't get one of the other async cards sorted out as the console.

 

[antiquekid3]

I would pull everything but the CPU, 4-SLU Sigma, the QUAD Memory and the card in slot 9. Then work on getting TT0: (console SLU) to show a MicroODT prompt.

I got it! I put the Sigma card back in. I figured the MDB SLU card was the console TTY, but it is in fact the Sigma. I've got it connected to my USB/RS-232 converter. The baud rate is now set to a more usable 19.2k baud from 300 baud.

ODT works just fine, but I still can't get it to do anything with the Emulex card. It just loops over and over, lighting up the three lights on the board in the same pattern. I'd at least expect it to spell something out to the console if it were trying to boot. Maybe I need to look at connecting the rest of the peripherals.

I'll investigate next which board has the dead short to ground on the 12V rail and look at replacing the likely faulty tantalum cap.

 

[Marty]

Hi All;
AntiqueKid3, "" I'll investigate next which board has the dead short to ground on the 12V rail and look at replacing the likely faulty tantalum cap. ""
It will probably take You longer to find the Cap than finding the Faulty Board..
Just take a Meter between the Ground Finger and the +12 finger on each Board, and it should Show up pretty quickly..
But, Lifting one end of each Cap and re-putting it back in, will take longer..
THANK YOU Marty

 

[gslick]

You could try manually issuing the bootstrap command to the QD32 using ODT and see if it can load a boot block from the drive.

In section 6.6 BOOTSTRAP COMMAND of the QD32 manual it lists a sequence of steps to load a boot block from the drive and then execute that block.

Steps 1-5 load the block, then steps 6-10 execute that block. You could use ODT to example memory starting at address 0, then manually follow steps 1-5, then example memory starting at address 0 again to see if it has been changed and if it looks anything like it might be a boot block from the drive. If that doesn't seem to work as expected it won't really give you much more information, but I don't know what else to try with that controller and drive for now.

It those steps IP would be 17772150 and SA would be 17772152 for your controller configured as the first MSCP controller. (See Table 6-1 QD32 IP and SA Registers in the manual).

 

[antiquekid3]

You could try manually issuing the bootstrap command to the QD32 using ODT and see if it can load a boot block from the drive.

In section 6.6 BOOTSTRAP COMMAND of the QD32 manual it lists a sequence of steps to load a boot block from the drive and then execute that block.

Steps 1-5 load the block, then steps 6-10 execute that block. You could use ODT to example memory starting at address 0, then manually follow steps 1-5, then example memory starting at address 0 again to see if it has been changed and if it looks anything like it might be a boot block from the drive. If that doesn't seem to work as expected it won't really give you much more information, but I don't know what else to try with that controller and drive for now.

It those steps IP would be 17772150 and SA would be 17772152 for your controller configured as the first MSCP controller. (See Table 6-1 QD32 IP and SA Registers in the manual).

The manual says to update the PSW with 340, but then it says 304. Any idea which is the correct value? Otherwise, everything is going swimmingly well so far. Thanks for the help, too.

 

[antiquekid3]

Seems to be 340, not 304. It does the same thing, unfortunately, issuing the bootstrap sequence as it does when it's jumpered for auto-boot. It does successfully transfer data to address 0 and beyond.

00000000/000240      
00000002/000406
00000004/000000
00000006/016522
00000010/005007
00000012/005007
00000014/005007
00000016/005007
00000020/112737
00000022/000340
00000024/177776
00000026/012704
00000030/001000
00000032/005002
00000034/012703
00000036/000400
00000040/012224
00000042/005303
00000044/001375
00000046/012706
00000050/002000
00000052/012707
00000054/001056
00000056/013767
00000060/000004
00000062/000014
00000064/013767
00000066/000006
00000070/000010
00000072/012705
00000074/172150
00000076/000413
00000100/000000
00000102/000000
00000104/100000
00000106/000744
00000110/000000
00000112/000001
00000114/000744
00000116/000011
00000120/000744
00000122/001041
00000124/000000
00000126/012702
00000130/004000
00000132/010225
00000134/010703
00000136/062703
00000140/177746
00000142/010304
00000144/005715
00000146/100463
00000150/031502
00000152/001774
00000154/012315
00000156/006302
00000160/100371
00000162/012301
00000164/001456
00000166/010102
00000170/005041
00000172/005701
00000174/001375
00000176/010201
00000200/012761
00000202/000060
00000204/177624
00000206/012761
00000210/000001
00000212/177626
00000214/012761
00000216/000001
00000220/177630
00000222/110061
00000224/177634
00000226/112361
00000230/177640
00000232/112361
00000234/177645
00000236/005713
00000240/001006
00000242/016761
00000244/177634
00000246/177664
00000250/016761
00000252/177624
00000254/177666
00000256/012722
00000260/000660
00000262/011422
00000264/012722
00000266/000574
00000270/011412
00000272/005765
00000274/177776
00000276/005715
00000300/001006
00000302/005762
00000304/177774
00000306/100773
00000310/105761
00000312/177726
00000314/001722
00000316/005045
00000320/000702
00000322/005045
00000324/010004
00000326/005007
00000330/005007
00000332/005007
00000334/005007
00000336/005007
00000340/005007
00000342/005007
00000344/005007
00000346/005007
00000350/005007
00000352/005007
00000354/005007
00000356/005007
00000360/005007
00000362/005007
00000364/005007
00000366/005007
00000370/005007
00000372/005007
00000374/005007
00000376/005007
00000400/005007
00000402/005007
00000404/005007
00000406/005007
00000410/005007
00000412/005007
00000414/005007
00000416/005007
00000420/005007
00000422/005007
00000424/005007
00000426/005007
00000430/005007
00000432/005007
00000434/005007
00000436/005007
00000440/005007
00000442/005007
00000444/005007
00000446/005007
00000450/005007
00000452/005007
00000454/005007
00000456/005007
00000460/005007
00000462/005007
00000464/005007
00000466/005007
00000470/005007
00000472/005007
00000474/005007
00000476/005007
00000500/005007
00000502/005007
00000504/005007
00000506/005007
00000510/005007
00000512/005007
00000514/005007
00000516/005007
00000520/005007
00000522/005007
00000524/005007
00000526/005007
00000530/005007
00000532/005007
00000534/005007
00000536/005007
00000540/005007
00000542/005007
00000544/005007
00000546/005007
00000550/005007
00000552/005007
00000554/005007
00000556/005007
00000560/005007
00000562/005007
00000564/005007
00000566/005007
00000570/005007
00000572/005007
00000574/005007
00000576/005007
00000600/005007
00000602/005007
00000604/005007
00000606/005007
00000610/005007
00000612/005007
00000614/005007
00000616/005007
00000620/005007
00000622/005007
00000624/005007
00000626/005007
00000630/005007
00000632/005007
00000634/005007
00000636/005007
00000640/005007
00000642/005007
00000644/005007
00000646/005007
00000650/005007
00000652/005007
00000654/005007
00000656/005007
00000660/005007
00000662/005007
00000664/005007
00000666/005007
00000670/005007
00000672/005007
00000674/005007
00000676/005007
00000700/005007
00000702/005007
00000704/005007
00000706/005007
00000710/005007
00000712/005007
00000714/005007
00000716/005007
00000720/005007
00000722/005007
00000724/005007
00000726/005007
00000730/005007
00000732/005007
00000734/005007
00000736/005007
00000740/005007
00000742/005007
00000744/005007
00000746/005007
00000750/005007
00000752/005007
00000754/005007
00000756/005007
00000760/005007
00000762/005007
00000764/005007
00000766/005007
00000770/005007
00000772/005007
00000774/005007
00000776/005007

It's a good start, anyways. Looks like I should start disassembling the boot sector!

 

[gslick]

Did you actually see the memory starting a location zero get written to by something that looks like it might be a boot block before you got as far as initializing the PSW?

A PSW of 340 octal would be the usual value there. Priority Level set to 7 and all condition codes clear.

The PSW bits are described in section 1.3.1 Processor Status Word (Address: 177 777 776) in the KDJ11-A_UsersManual.pdf if you're not already familiar with that.

 

[antiquekid3]

Yes, I zeroed out a good portion of the first page starting at 0, then watched as they all magically updated. The dump pasted above is prior to executing it.

 

[gslick]

Oh, OK so you're getting some progress if the QD32 is actually able to read a boot block from the drive into system memory. That is a good start. Unless someone else has better ideas now I might start by disassembling it and then maybe stepping through it and see what happens when it is executing. Maybe it manages to load additional blocks from the drive and then jumps to that, but then whatever OS it is trying to load is having problems.

I've never used PDP11GUI with real hardware. Maybe you could get setup to use that and see if you can dump and upload more of the disk to a PC for further analysis of what is currently on the disk.

http://retrocmp.com/tools/pdp11gui/disk-images-readwrite