Difference between revisions of "PDP-11/40"

From Computer History Wiki
Jump to: navigation, search
(Note config changes)
m (Fix formatting of Infobox (I broke it last edit))
 
(15 intermediate revisions by 2 users not shown)
Line 1: Line 1:
 
{{Infobox Machine
 
{{Infobox Machine
 
| name=PDP-11/40
 
| name=PDP-11/40
| image=Pdp-11-40.jpg  
+
| image=Pdp-11-40.jpg
| caption=PDP-11/40 with [[TU56]] dual [[DECtape]] drive.
+
| imgwidth = 200px
| manufacturer=[[Digital Equipment Corporation]]  
+
| caption=PDP-11/40 with [[TU56 DECtape Transport|TU56]] dual [[DECtape]] drive.
| word size= 16 bit
+
| manufacturer=[[Digital Equipment Corporation|DEC]]
| virtual address= 16 bit
+
| year introduced = January, 1973
| physical address= 18 bit
+
| word size= 16 bits
 +
| virtual address= 16 bits
 +
| physical address= 18 bits
 
| bus arch= [[UNIBUS]]|
 
| bus arch= [[UNIBUS]]|
 +
| cpu = [[KD11-A CPU|KD11-A]]
 
|}}
 
|}}
[[Image:PDP11-40-geerol.jpg|150px|thumb|right|A PDP-11/40 setup.]]
+
The '''PDP-11/40''', introduced in 1973, was the second machine to support 18-bit [[main memory]] [[address]]ing on the [[PDP-11]] platform, after the [[PDP-11/45]] (albeit with a subset of the [[PDP-11 Memory Management]] [[architecture]]). The [[Original Equipment Manufacturer|OEM]] version of the -11/40 is the '''PDP-11/35'''; the machines are completely identical, except for the printing on the [[front panel]] inlay.
  
The '''PDP-11/40''', introduced in 1973, using the [[KD11-A CPU]], was the second machine to support 18-bit memory addressing on the [[PDP-11]] platform, after the [[PDP-11/45]] (albeit with a subset of the [[PDP-11 Memory Management]] [[architecture]]).
+
It used the [[KD11-A CPU]], which connected to the [[KY11-D Console|KY11-D]] front panel via a pair of BC08-R [[flat cable]]s. Optional additions to the [[Central Processing Unit|CPU]] included:
 +
 
 +
* [[KT11-D Memory Management]]
 +
* [[KE11-E Extended Instruction Set]] ([[PDP-11 Extended Instruction Set|EIS]])
 +
* [[KE11-F Floating Instruction Set]] ([[FIS floating point]]
 +
* [[KJ11-A Stack Limit Register]]
 +
* [[KW11-L Line Time Clock]]
 +
 
 +
It came in two types of [[mounting box]]es: the 21" [[BA11-F Mounting Box]] (usual for the -11/40); and the 10-1/2" [[BA11-D Mounting Box]] (early units) and [[BA11-K Mounting Box]]es (later units) (usual for the -11/35).
 +
 
 +
An -11/40 with a special blue-themed front panel was used as the [[front end]] for the [[KL10]] CPU in [[DECsystem-10]]s and [[DECSYSTEM-20]]s - though the -11/40 panel was hidden behind a cabinet door on the -20.
  
 
The 11/40 was also a primary platform for early [[UNIX]].
 
The 11/40 was also a primary platform for early [[UNIX]].
  
==Configuration changes==
+
==Configuration Changes==
  
Early versions of the /40 omitted the 4 fans at the bottom of the pull-out [[BA11-F mounting box]] used to house the [[Central Processing Unit|CPU]], and had only the 5 fans at the top.
+
[[Image:PDP11-40-geerol.jpg|150px|thumb|right|A PDP-11/40 setup]]
  
The [[harness]] from the [[H740 power supply]] was also different in later models, allowing use of the [[H754 power supply module]] which provided +20V, for use with the [[MM11-U core memory]].
+
Early versions of the BA11-F in the -11/40 omitted the 4 fans at the bottom of the pull-out used to house the CPU, and had only the 5 fans at the top.
  
The [[OEM]] version of the -11/40 is the '''PDP-11/35'''; the machines are completely identical, except for the printing on the front panel.
+
The BA11-D used 9-pin [[DEC power distribution connectors]], whereas the BA11-K used the 6-pin/15-pin pair, so presumably the [[backplane]] for the BA11-D had a different [[power harness]] than the one for the BA11-K (which it is confirmed had the 6/15-pin connectors).
 +
 
 +
The power harness from the [[H742 Power Supply]] was also different in later models, allowing use of an [[H754 +20V, -5V Regulator]] which provided +20V, for use with the [[MM11-U core memory]].
 +
 
 +
The front panel had round LED cutouts and the data ribbon cables were on the right. Later revisions had square LED cutouts and the ribbon cables were on the left.
 +
 
 +
==Diagnostic Software==
 +
 
 +
Listed below is a non-exhaustive list of PDP-11/40 diagnostics. In brackets is the name of the paper tape binary, most of these test can be loaded and run from XXDP directly.
 +
Much of the PDP-11 diagnostic software has been aggregated into one place, the "[https://www.retrocmp.com/images/stories/joerg/pdp11_diagnostic_database/index.html PDP11 Diagnostic Database]", listings and binary files can be found there.
 +
 
 +
===Modules===
 +
*'''KD11-A - CPU'''
 +
* [https://www.retrocmp.com/images/stories/joerg/pdp11_diagnostic_database/modules/BQEA.html BQEA] (BQEAC1) - KD11-A CPU DIAGNOSTIC
 +
* [https://www.retrocmp.com/images/stories/joerg/pdp11_diagnostic_database/modules/ZKAQ.html ZKAQ] (ZKAQG0) - PDP-11 POWER FAIL DIAGNOSTIC
 +
** This module is best exercise by having a UniBone assert DC LO via the pwr command.
 +
* [https://www.retrocmp.com/images/stories/joerg/pdp11_diagnostic_database/modules/CKBA.html CKBA] (CKBAB0) - PDP11 45-11/40 BASIC CP TESTS
 +
 
 +
'''KT11-D - Memory Management'''
 +
* [https://www.retrocmp.com/images/stories/joerg/pdp11_diagnostic_database/modules/BKTA.html BKTA] (BKTAD1) - KT11-D BASIC LOGIC TEST
 +
* [https://www.retrocmp.com/images/stories/joerg/pdp11_diagnostic_database/modules/BKTB.html BKTB] (BKTBB0) - KT11-D ACCESS KEYS TEST
 +
* [https://www.retrocmp.com/images/stories/joerg/pdp11_diagnostic_database/modules/BKTD.html BKTD] (BKTDC0) - KT11-D PROCESSORS STATES TEST
 +
* [https://www.retrocmp.com/images/stories/joerg/pdp11_diagnostic_database/modules/BKTF.html BKTF] (BKTFD0) - MEMORY MANAGEMENT ABORT TESTS
 +
** Across my two pdp-11 40 CPUs neither of them will pass these tests, I am suspicious there is an ECO or other factors at play.
 +
* [https://www.retrocmp.com/images/stories/joerg/pdp11_diagnostic_database/modules/BKTG.html BKTG] (BKTGD1) - KT11-D EXERCISER
 +
 
 +
'''KE11-E - Extended Instruction Set'''
 +
* [https://www.retrocmp.com/images/stories/joerg/pdp11_diagnostic_database/modules/CKBI.html CKBI] (CKBIB0) - ASH INSTRUCTION TEST
 +
** This module tests for ECO M7238-0004, older boards may fail this test. More info can be found in the module listing.
 +
* [https://www.retrocmp.com/images/stories/joerg/pdp11_diagnostic_database/modules/CKBJ.html CKBJ] (CKBJA0) - ASHC INSTRUCTION TEST
 +
* [https://www.retrocmp.com/images/stories/joerg/pdp11_diagnostic_database/modules/CKBK.html CKBK] (CKBKA0) - MUL INSTRUCTION TEST
 +
** Tests can fail because of faults in the KE11-F. KE11-F tests may not catch all failures.
 +
* [https://www.retrocmp.com/images/stories/joerg/pdp11_diagnostic_database/modules/CKBL.html CKBL] (CKBLA0) - DIV INSTRUCTION TEST
 +
** Tests can fail because of faults in the KE11-F. KE11-F tests may not catch all failures
 +
 
 +
'''KE11-F - Floating Instruction Set'''
 +
* [https://www.retrocmp.com/images/stories/joerg/pdp11_diagnostic_database/modules/BKEA.html BKEA] (BKEAB0) - KE11F (PDP-11 FIS) INSTRUCTION TESTS
 +
** The KE11-F is dependant on the KE11-E, test it first.
 +
** Tests 70-72 exercise the SACK timeout logic on the M7234 Timing board, this is the only module I am aware of that does so.
 +
** The test will print three lines of random length lengths of "@". This is misleading, the test checks the floating point abort, it interleaves floating point operations and prints. When it successfully aborts it moves onto the next test. The number of "@"s written to the screen is an indication of the latency of the abort and should be low, typically less then 6. Seeing 10s of "@" per test is a should be considered as a failure.
 +
** EIS MUL or DIV instructions are not tested, this is important because the FIS board can have failures that will cause the EIS to malfunction, specifically Auxiliary Alu Control, see KE-5 & KF-4.
 +
** Tests 70-72 exercise BR4 and BG4, to accelerate this testing enable SW11 (skips sub-test, subsequently iterates tests quicker)
 +
* [https://www.retrocmp.com/images/stories/joerg/pdp11_diagnostic_database/modules/BKEB.html BKEB] (BKEBA0) - KE11F (PDP-11 FIS) EXERCISER
 +
 
 +
'''KJ11-A - Stack Limit Register'''
 +
* [https://www.retrocmp.com/images/stories/joerg/pdp11_diagnostic_database/modules/CKBF.html CKBF] (CKBFD0) - STACK LIMIT TEST
 +
** Documentation indicates this test should take approximately 1 minute, in my experience it takes closer to 2 (this might be anomalous).
 +
** Internally this test does 1500 passes before indicating success, this can be manually altered if desired.
 +
 
 +
'''KW11-L - Line Time Clock'''
 +
* [https://www.retrocmp.com/images/stories/joerg/pdp11_diagnostic_database/modules/ZKAQ.html ZKWA] (ZKWAG0) - LINE FREQUENCY CLOCK TEST
 +
** Exercises BR6 and BG6
  
A special -11/40 with a blue front panel was used as the front-end for the [[KL-10]] in DECSystem-10s and DECSystem-20s - though the 11/40 panel was hidden behind a cabinet door on the -20.
 
  
 
== hampage.hu ==
 
== hampage.hu ==
Line 29: Line 92:
 
<i>
 
<i>
 
[[Image:pdp11-40.jpg|200px|left]]
 
[[Image:pdp11-40.jpg|200px|left]]
The /40 was introduced in January, 1973. It was a higher performance version of the /20, with a microcoded multi-board CPU. 18-bit MMU (-> max. 128 kw memory), FPU and EIS/FIS (extended/floating point instruction set) were available options.  
+
The /40 was introduced in January, 1973. It was a higher performance version of the /20, with a [[microcode]]d multi-board CPU. 18-bit MMU (-> max. 128 kw memory), FPU and EIS/FIS (extended/floating point instruction set) were available options.  
  
There was a 'graphical' workstation model of the /40, marketed as the GT44. Many /40s were used in DECsystem-10's (KL10's, to be exact), as the frontend/IO-processor. These were painted differently, the panel looks awsome in blue too. (no, it's not just the frontpanel, that I like about the -11...)
+
There was a 'graphical' workstation model of the /40, using a [[VT11 Graphic Display Processor]], marketed as the GT44. Many /40s were used in DECsystem-10's (KL10's, to be exact), as the frontend/IO-processor. These were painted differently, the panel looks awsome in blue too. (no, it's not just the frontpanel, that I like about the -11...)
  
On the picture to the left we can see two RK05's (disk pack drives) and two TU60 (DECcasette) units. On this picture we can see a close-up at the frontpanel (with bootdevice-selector panel) and the paper-tape puncher/reader.
+
On the picture to the left we can see two RK05's (disk pack drives) and two TU60 (DECcasette) units. <!-- On this picture we can see a close-up at the frontpanel (with bootdevice-selector panel) and the paper-tape puncher/reader. -->
  
 
OS's at this point: DOS/BATCH, RSX11, RT-11, RSTS.
 
OS's at this point: DOS/BATCH, RSX11, RT-11, RSTS.
Line 40: Line 103:
 
</i>
 
</i>
  
{{stub}}
+
{{semi-stub}}
 +
 
 +
==See also==
 +
 
 +
* [[PDP-11/40E]]
 +
 
 +
==External links==
 +
 
 +
* [https://iamvirtual.ca/collection/systems/minis/PDP11-10/PDP11-10--PriceList.pdf PDP-11/10-40 Price List] (January 1, 1973)
 +
* [http://www.bitsavers.org/pdf/dec/pdp11/1140/DEC-11-H40SA-A-D_PDP-11_40_System_manual.pdf PDP-11/40 system manual] (DEC-11-H40SA-A-D)
 +
* [http://www.bitsavers.org/pdf/dec/pdp11/1140/1140_SystemManual.pdf PDP-11/40, -11/35 (21 inch chassis) system manual] (EK-11040-TM-002)
 +
* [http://www.bitsavers.org/pdf/dec/pdp11/1140/PDP-1140_System_Engr_Drawings_Rev_P_Jun74.pdf PDP-11/40 system engineering drawings]
 +
* [http://www.bitsavers.org/pdf/dec/pdp11/memos/700925_PDP-1140_Technical_Memo_Index.pdf PDP-11/40 Technical Memo Index & Abstracts] - interesting overview of a series of design memos (most also available) on the -11/40
  
 
{{PDP-11}}
 
{{PDP-11}}
  
 
[[Category: UNIBUS PDP-11s]]
 
[[Category: UNIBUS PDP-11s]]

Latest revision as of 09:11, 24 December 2025


PDP-11/40
Pdp-11-40.jpg
PDP-11/40 with TU56 dual DECtape drive.
Manufacturer: DEC
Year Introduced: January, 1973
Word Size: 16 bits
CPU: KD11-A
Physical Address Size: 18 bits
Virtual Address Size: 16 bits
Bus Architecture: UNIBUS

The PDP-11/40, introduced in 1973, was the second machine to support 18-bit main memory addressing on the PDP-11 platform, after the PDP-11/45 (albeit with a subset of the PDP-11 Memory Management architecture). The OEM version of the -11/40 is the PDP-11/35; the machines are completely identical, except for the printing on the front panel inlay.

It used the KD11-A CPU, which connected to the KY11-D front panel via a pair of BC08-R flat cables. Optional additions to the CPU included:

It came in two types of mounting boxes: the 21" BA11-F Mounting Box (usual for the -11/40); and the 10-1/2" BA11-D Mounting Box (early units) and BA11-K Mounting Boxes (later units) (usual for the -11/35).

An -11/40 with a special blue-themed front panel was used as the front end for the KL10 CPU in DECsystem-10s and DECSYSTEM-20s - though the -11/40 panel was hidden behind a cabinet door on the -20.

The 11/40 was also a primary platform for early UNIX.

Configuration Changes

A PDP-11/40 setup

Early versions of the BA11-F in the -11/40 omitted the 4 fans at the bottom of the pull-out used to house the CPU, and had only the 5 fans at the top.

The BA11-D used 9-pin DEC power distribution connectors, whereas the BA11-K used the 6-pin/15-pin pair, so presumably the backplane for the BA11-D had a different power harness than the one for the BA11-K (which it is confirmed had the 6/15-pin connectors).

The power harness from the H742 Power Supply was also different in later models, allowing use of an H754 +20V, -5V Regulator which provided +20V, for use with the MM11-U core memory.

The front panel had round LED cutouts and the data ribbon cables were on the right. Later revisions had square LED cutouts and the ribbon cables were on the left.

Diagnostic Software

Listed below is a non-exhaustive list of PDP-11/40 diagnostics. In brackets is the name of the paper tape binary, most of these test can be loaded and run from XXDP directly. Much of the PDP-11 diagnostic software has been aggregated into one place, the "PDP11 Diagnostic Database", listings and binary files can be found there.

Modules

  • KD11-A - CPU
  • BQEA (BQEAC1) - KD11-A CPU DIAGNOSTIC
  • ZKAQ (ZKAQG0) - PDP-11 POWER FAIL DIAGNOSTIC
    • This module is best exercise by having a UniBone assert DC LO via the pwr command.
  • CKBA (CKBAB0) - PDP11 45-11/40 BASIC CP TESTS

KT11-D - Memory Management

  • BKTA (BKTAD1) - KT11-D BASIC LOGIC TEST
  • BKTB (BKTBB0) - KT11-D ACCESS KEYS TEST
  • BKTD (BKTDC0) - KT11-D PROCESSORS STATES TEST
  • BKTF (BKTFD0) - MEMORY MANAGEMENT ABORT TESTS
    • Across my two pdp-11 40 CPUs neither of them will pass these tests, I am suspicious there is an ECO or other factors at play.
  • BKTG (BKTGD1) - KT11-D EXERCISER

KE11-E - Extended Instruction Set

  • CKBI (CKBIB0) - ASH INSTRUCTION TEST
    • This module tests for ECO M7238-0004, older boards may fail this test. More info can be found in the module listing.
  • CKBJ (CKBJA0) - ASHC INSTRUCTION TEST
  • CKBK (CKBKA0) - MUL INSTRUCTION TEST
    • Tests can fail because of faults in the KE11-F. KE11-F tests may not catch all failures.
  • CKBL (CKBLA0) - DIV INSTRUCTION TEST
    • Tests can fail because of faults in the KE11-F. KE11-F tests may not catch all failures

KE11-F - Floating Instruction Set

  • BKEA (BKEAB0) - KE11F (PDP-11 FIS) INSTRUCTION TESTS
    • The KE11-F is dependant on the KE11-E, test it first.
    • Tests 70-72 exercise the SACK timeout logic on the M7234 Timing board, this is the only module I am aware of that does so.
    • The test will print three lines of random length lengths of "@". This is misleading, the test checks the floating point abort, it interleaves floating point operations and prints. When it successfully aborts it moves onto the next test. The number of "@"s written to the screen is an indication of the latency of the abort and should be low, typically less then 6. Seeing 10s of "@" per test is a should be considered as a failure.
    • EIS MUL or DIV instructions are not tested, this is important because the FIS board can have failures that will cause the EIS to malfunction, specifically Auxiliary Alu Control, see KE-5 & KF-4.
    • Tests 70-72 exercise BR4 and BG4, to accelerate this testing enable SW11 (skips sub-test, subsequently iterates tests quicker)
  • BKEB (BKEBA0) - KE11F (PDP-11 FIS) EXERCISER

KJ11-A - Stack Limit Register

  • CKBF (CKBFD0) - STACK LIMIT TEST
    • Documentation indicates this test should take approximately 1 minute, in my experience it takes closer to 2 (this might be anomalous).
    • Internally this test does 1500 passes before indicating success, this can be manually altered if desired.

KW11-L - Line Time Clock

  • ZKWA (ZKWAG0) - LINE FREQUENCY CLOCK TEST
    • Exercises BR6 and BG6


hampage.hu

Quoting...

Pdp11-40.jpg

The /40 was introduced in January, 1973. It was a higher performance version of the /20, with a microcoded multi-board CPU. 18-bit MMU (-> max. 128 kw memory), FPU and EIS/FIS (extended/floating point instruction set) were available options.

There was a 'graphical' workstation model of the /40, using a VT11 Graphic Display Processor, marketed as the GT44. Many /40s were used in DECsystem-10's (KL10's, to be exact), as the frontend/IO-processor. These were painted differently, the panel looks awsome in blue too. (no, it's not just the frontpanel, that I like about the -11...)

On the picture to the left we can see two RK05's (disk pack drives) and two TU60 (DECcasette) units.

OS's at this point: DOS/BATCH, RSX11, RT-11, RSTS.

Trivia: The 11/40 was one of the most often "cloned" machine. the soviet/tscech/polish SM-4 (CM-4), the hungarian TPA-1140 are all copies of this computer.

See also

External links