• Please review our updated Terms and Rules here

pdp-8/a build from parts

The 8/a gets some love again (but ain't giving back yet)

The 8/a gets some love again (but ain't giving back yet)

So with Reinhard's RL02 emulator working good, I want to go back to my work of making bootable RL02 OS/8 pack images. But, it's cold in the garage where the 8/e lives. The 8/a is small, light, and inside the house. So I hauled inside one of those Decmate II RX02 teacart floppy cabinets and put the 8/a on top with an RX8E inside.

I expected too much I guess that I could boot right into OS/8 off of an RX01. I have never really put the 8/a through its paces after assembling it earlier. It ran tests punched in on the keypad just fine, but I never tried any of the proper MAINDECs.

It is sick, with an interrupt related problem from the way it behaves with the RX01 bootstrap. So I looked for and found off the internet the DJKKA processor diagnostic. I can't find the diagnostic write-up for this diagnostic anywhere though. The 8/a users manual explains how to start it, and how it should behave if all is well, but all is not well. I pulled the M8317 and RX8E, swapped a spare M8316, checked all the switches, cleaned the backplane, everything usual. I can load the bin loader fine, and even ran D0AA which is an 8/e instruction set diagnostic, but DJKKA bombs.

The DJKKA does start to run when started per the user's manual from 0200, and stops at location 0222 as it's supposed to. The manual says that the link should be set and that's all that should be set in the status register. However, I also see ION (interrupts enabled). However, when I look at the disassembled code (http://dustyoldcomputers.com/pdp-common/reference/papertapes/maindec/maindec-08-djkka-c-pb1.pal ) It does look like ION should be set since there was only a 6001 (ION) so far and no IOF (6002).

The code halts at 1317 (1320 in the display) after restarting (the manual says to do so) after the halt at 0222. Restarting again, the machine runs for a moment and halts at the same spot. I need to study the code and figure out why. If all was well, the machine would not halt, the tests would loop.

Does anyone out there have the diagnostic write-up for DJKKA? Adrian, you have an 8/a - do you have it?

Lou
 
Lou,

I assume you found this in your searching - https://groups.google.com/forum/#!msg/alt.sys.pdp8/d75mHszBt7s/TgUFyls_ZjYJ - a note from Bob Vines saying he has the docs for "MAINDEC-08-DJKKA-B-D, PDP-8/A CPU TEST, dated Feb. 15, 1975". Bob is an active member of the classiccmp/cctalk - try him at bobvines00 at gmail.com if this info is useful. I'll dig around here a little bit as well.

Jack
 
Jack,

Yes, I did find that post also and wrote to Bob. So far he's only found the cover sheet to the listing, but he said he would keep looking.

I have started to, and will need to continue, look carefully through the disassembled code (unless the write-up turns up.)

Lou
 
Lou,
I have the listing, 08-DJKKA-B in the original DEC paper form, I can scan it if you dont find it else were, let me know.
I also have some other listings for 8a maindecs that may be of use, DJEXB, DJMSA, DJEKA and DJKMA
Dave
 
DaveH said:
Lou,
I have the listing, 08-DJKKA-B in the original DEC paper form, I can scan it if you dont find it else were, let me know.
I also have some other listings for 8a maindecs that may be of use, DJEXB, DJMSA, DJEKA and DJKMA
Dave
Click to expand...

Dave,

You are the hero of the day! Yes, please do scan the DJKKA. I'm not sure that Bob V. will find the listing given how he found the cover only. The others would also be useful as I will want to exercise the entire machine. I'll send you my proper e-mail address in a private message.

I will owe you a few pints if you come to New York (or if I visit Scotland again!)

Lou
 
Seconded

Seconded

jackrubin said:
Amen to that! Please scan ALL your 8/A MAINDECs - I'm happy to share the cost in pounds, dollars or pints. These definitely belong on bitsavers.

Jack
Click to expand...


Lou,

I'll chip in too. I need to get my 8/a resurrected also ( again a former Tim T machine).

Oh, and Lou I have the electronics of an 8/a console, of unknown functionality that you're welcome to for spare parts etc.

? ; - )
 
Crawford said:


I'll chip in too. I need to get my 8/a resurrected also ( again a former Tim T machine).

Oh, and Lou I have the electronics of an 8/a console, of unknown functionality that you're welcome to for spare parts etc.

? ; - )
Click to expand...
Good to know my former machines have good homes and are either getting used
or will be in the future. I do miss my pdp-8e and of course the 8i I had.
 
Tim,

Yes, they all have good homes. The 8/I may well have the best home of all though!

Other fans of the 8/a,

While it is on my mind, I have found at least 5 different bootstrap rom sets for the M8317 KM8-A:

87 and 88 : The original roms, described in the 8/a users handbook, hi/lo papertape / RK8E / TC08 / RF08 / TA8E
156 and 157 : The roms for the DS310 (8/a in a desk configuration), RX8E bootstrap only, but some nice self test diagnostics too
158 and 159 : The later roms also described in the 8/a user handbook, hi/lo papertape / RK8E / RX8E / RF08 / TA8E
465 and 469 : The latest roms, RL8A / RX8E
Unreadably marked : When dumped appear to download the secondary bootstrap from device 40/41 (secondary SLU)

I have 156 and 157 on one of the boards from Tim, because the systems that Crawford and I have are (Crawford) / were (me) from the DS310 desk configurations. When the bootstrap runs, it displays 0123 in sequence. First 0, then 1, then 2, then 3. The DS310 installation manual (on the internet) says

0 - VT50H check passed
1 - 8/a test passed
2 - Memory test passed
3 - Floppy test passed

Currently, I get all four of those (when RX8E is connected), but have this interrupt problem that I need to fix.

Anyone know of any other boot roms?

Lou


Someday I will burn a set of 465 and 469.
 
Hi Lou,
Ok will do, glad to be of help, soon as I can get a few mins, its a cheap and cheerful scanner so will have to experiment a bit.
Pints will be fine though I cant see me being in New York unfortunately but if you ever visit Scotland you will be more than welcome.
Will put them in Dropbox and post a link.
Dave
 
Hi All;
I have Downloaded these files, just in case they might need to have another place to reside, not that I have an 8.. Which I don't..
THANK YOU Marty
 
Hi All;
Have looked at the code, and I am NO expert, but according to the Comments, it looks like Your problem at this point is that the PC (I assume Program Counter) is not working or storing what it was given.. I have not looked at any of the previous code, so I can't say that this has anything to do with ION on or off..
THANK YOU Marty
 
Marty,

Yes, indeed this halt has nothing to do with interrupts. The listing says that the jump to subroutine (JMS) isn't storing the program counter properly in location 1311.

After the halt though, I do see a 1310 stored in location 1311. Something else must be broken about JMS.

From single stepping the machine, it appears that JMS is not continuing execution from one location beyond where the PC was stored, but that very location where the PC was just stored, like a JMP (except JMP doesn't write out the value of the PC). So in this bizarre case, a 1310 is executed, which is TAD with location 1310, and sure enough I see a 7402 in the AC and it's downhill from there.

How on earth the BIN loader works properly is beyond me. It uses JMS. It just must be dumb luck that the instruction that would be represented by the PC address value that was stored does not screw anything up. But, the RX bootstrap does use JMS also, and stuff there does get screwed up.

This has been an enormous help. Now I need to fix the JMS instruction. For the hell of it, I may load D0IB, which is the 8/E JMP/JMS test to make sure it shows the same behavior.

Lou
 
Last edited:
Lou,
While looking out the listings I found that I have more than I remembered in documentation for the a, including schematics, and a very beaten up copy of the users manual. I guess you have all you need but if not I may be able to help, let me know.
Dave
 
Lou,
the "users" manual I have is more in the line of the PDP maintenance manual, the binding is gone and its now in 3 pieces so I cant tell its actual title, however as per the PDP8/E and M maintenance manuals its got some very detailed descriptions of the cards ect. Attached is a scan of the first 2 pages of the CPU chapter. Its a bit of a pain to scan as if I try to flatten it to much its going to disintegrate even further however I should be able to scan this chapter if it will be of use to you.
Arr, cant attach, ok its in the dropbox ,same link, 8a service man test.pdf
Dave
 
jackrubin said:
Dave,

The 8A User's Manual is already scanned and available on bitsavers - http://bitsavers.trailing-edge.com/...2_PDP-8A_Miniprocessor_Users_Manual_Dec76.pdf , along with Al's usual collection of helpful docs. It's the 8A MAINDECs that seem to have vanished. As noted earlier, we'll cheerfully reimburse you for the MAINDEC scans if you want to take them to a commercial copy center.

Jack
Click to expand...

Arr good, I must admit I dident check. I have scanned 2 using my fairly basic scanner but will do the rest at a friends who has a good ADF scanner.
Dave
 
Self inflicted wounds hurt the most....

Self inflicted wounds hurt the most....

Well, I did it to myself: http://www.vintage-computer.com/vcforum/album.php?albumid=182&attachmentid=17005

So the diagnosic hung on one and only one instruction - JMS. And, the problem was that the PC would not increment to one instruction past where the PC was stored to resume execution after jumping. The pdp8/a operators manual explains at the top of page 4-63 exactly what is required for the increment to occur. CPMA DIS L (on CU1) and ROM ADDR L (on AU2) both have to be negated. I looked at the 7408 at E47 where these two are anded before going to the carry-in mux. ROM ADDR L was stuck low (true) and there was no good reason why. When I shut everything down, E47 pin 1 was hard grounded. I went back to the backplane and AU2 was hard grounded. I pulled every damn card out of the backplane and AU2 was grounded.

I pulled the backplane out and saw AU2 pin bent over at one of the slots near a standoff - it must have been touching that standoff. I obviously screwed this up when I put the machine back together after cleaning. I am an unlucky bastard.


Lou (and you thought you did something stupid today.)
 
Last edited:
You must log in or register to reply here.
Back
Top