PDP-11s behind the Iron Curtain

The PDP-11, as the PDP-8 before it, was cloned and copied extensively behind the so-called Iron Curtain. A number of plants produced PDP-11 compatible systems in the Soviet Union, including Elektronika-79 (11/70) and several machines without direct DEC analogues (DWK-4 with proprietary video controller).

Nobody knows how much of these clones were issued by many of the plants in the countries of the Warsaw Pact, but I believe that the total amounts of units should counted by hundreds of thousands. AFAIK the following countries issued a DEC clones of computers and peripherals:

• Soviet Union (SM-4, SM-1420/xxx, SM-1600/xxx, Electronika-xxx, DWK-n, UKNC,...)
• Bulgaria (SM-4, SM-1420/xxx)
• Eastern Germany "DDR" (SM-1420/xxx)
• Poland (Mera-xxx)
• Hungary (SM-4)
• Note that `SM' appears as `CM' in Cyrillic (see discussion below regarding Cyrillic letter usage on this page)

Not all clones had analogues at DEC product line, but most of them were very close to some model (not in quality though). All DEC software and Unixes runs on these computers without problems (in fact, Soviet users used adopted versions of software because the KGB stole ALL source codes for RT11 and RSX11 and made it freely distributed; after a while it was modified to have support for Russian language and some nonstandard devices/architectures, RT11 became RAFOS, FOBOS, and FODOS, and RSX11 became OSRV).

Computers

We can divide the clones into four categories:

• - UNIBUS
• - QBus with LSI-11
• - QBus with Single chip CPU
• - Proprietary bus

Below is the list of clones we know of so far:

• MINICOMPUTERS
  • * SM-3 UNIBUS. 11/05 clone. There's some argument that better be described as a microcoded 11/20. Eventually, someone will have to sit down with the English and Russian spec sheets and cross-match to determine the best description...
  • * SM-4 UNIBUS. Produced in SU and most other Warsaw Pact countries. Analog of the 11/40. most had core memory.
  • * SM-1420/xxx (SM-1420/xxx, SM-1600/xxx, Elektronika 100/xx, Mera-xxx) UNIBUS. Most popular models, analogues of 11/34-/55 depending on the "xxx" (even 11/74 extensions) and had lots of options. Based on bit-slice processors.
  • * TPA-11/40 UNIBUS. From Hungary. Direct PDP-11/40 clone.
  • * TPA-11/48 UNIBUS. An enhanced TPA-11/40. The enhancements included 22 bit maps, split I/D, cache, etc, so that it looked enough like an 11/70 to fool DEC OSes. Ran much more slowly than a real 11/70 (actually, the 11/48 was slower than a TAP-11/40). No MASSBUS (on any TPA model?).
  • * TPA-11/110 aka TPA-Janus. From Hungary. A multi-processor machine, with a Russian Elektronika-MC1201.01 CPU (LSI-11 clone) and a Z80 as secondary processor. There could be 2 or 3 Z80s on the QBUS.
• - MICROCOMPUTERS
  • * Electronica-60 QBus/LSI. Very popular microcomputer in the beginning of 80's VERY similar to the original LSI-11.
  • * DWK-n line Single chip CPU (K1801BMx), QBus. Most popular computer of the 80's ranged from DWK-1 to 4.
    • DWK-1 - PDT 11/130.
    • DWK-2 - PDT 11/150.
    • DWK-3 had an 1801BM2 CPU with the same features as DWK-1/2 (technically, a DWK-2M) plus a built-in MMU and was equipped with a proprietary black&white graphical display and two floppies (each 200KB, 40 tracks, single-sided). No direct DEC analogue.
    • DWK-3M was a DWK-3 with 400KB floppies (80 tracks, 10 sectors, double-sided), a 10MB CM5508 hard disk (made in Bulgaria). Later models have a color graphics controller.
    • DWK-4 had a EIS/FIS enhanced CPU (1801BM3), 1MB of memory, a 20MB hard disk (MC5405/MC5410, made in Rostov-on-Don, Russia) and vt100-compatible programmable colr display. Some models were equipped by 5MB RD50-like HDD (sometime these HDDs worked ;). No direct DEC analogue.
    • DWK-5 is based on the 1801BM4, with the same peripherals as the DWK-4. No one has yet admitted to having seen one, though.
  • * Elektronika-85, models MC0585 and MC0585.1. Used the 1811B1 processor, and had 512KB of memory (on the motherboard, rather than the DEC Pro's CTI-slot memory). There were three hard disk options: 5MB MC5401 (made in Rostov-on-Don), 10MB MC5402 (Rostov-on-Don) and CM5508 (Bulgaria), and 20MB MC5405 and MC5410 (Rostov-on-Don). The system is still in production in Voronezh, Russia, not far from Rostov-on-Don. There's a rumour that some later E-85s were produced with the 1831BM1 processor, but no one admits to having seen them.
  • * TPA-11/170. From Hungary. A J11-based microPDP-clone, designed around 1986-1987.
  • * TPA-EMU-11. A microprogrammable universal processor built on a UBUS board. It could be reprogrammed, but the basic microcode made it behave like a -11. I don't know if this beast was ever used, or if it was only planned.
  • * MICRO 11/23 SYSTEM 2500 TEAM COMPUTER. Produced by MIKI, did several (metric) QBUS systems based on Elektronika's CPU products (the M2 and M6, LSI-11 and KDF-11A clones, respectively). The PSU, CPU, and SLUs are Elektronika boards, everything else (parallel port, disk controllers, floppy controller, enclosure) was made by MIKI. "Kind of" OEM.
• - HOME/SCHOOL COMPUTERS
  • * UKNC - designed to be basic School computer of Gorbachev's period of '80s. Proprietary design and peripherals. The MC0511 has two 1801WM2 CPUs. The "system" CPU has 56KB of regular memory, 8K I/O page, and 8K of hidden memory (see WM2 discussion below). The "peripheral" processor has 32KB RAM and 32KB PROM with built-in peripheral service routines. Also, this machine has a 3*32KB graphic video memory. Originally, it came with two floppies and no hard disk. Several MFM and IDE controllers have become available since then, however.
  • * BK-001x - first Soviet home PC, based on 11/2 CPU with proprietary chipset to serve RAM/Peripherals. 32Kb RAM, no OS - ROM BASIC or FOCAL interpreters. In the SU this computer had a success similar to TRS-80. Late models had more memory and CPU from DWK-3. Many enthusiasts created peripherals to this beast as it has QBus-compatible socket on the rear side. Note that it shipped with no operating system because it shipped with no peripherals. Hence the provisioning of such by enthusiasts. Several OS options now exist: many variants of RT-11,  MKDOS, CSIDOS, et cetera.
• PROPRIETARY BUS
  • TPA-11/440. From Hungary. 32-bit internal bus called the X-bus, but since there weren't many devices for it (only the ones the KFKI (where the TPA's were built) produced), it used the UBUS (KFKI-version of the UNIBUS, electronically the same with mechanical differences). Its designers said it had 0.7-0.8 of the performance of the MicroVAX II (which was introduced by Digital around the same time), but users say it was less (around 0.6-0.7).
  • TPA-11/420. Based on the J-11 chip, with the X-bus of the TPA-11/440.

Chips

There were several chip-sets made, corresponding (usually) to what DEC produced.

• 581. The first chip-set was the 581 which involves many chips marked as 581. The 581IK1,IK2 is the main manager (controller, "heart") of the computer. 581RUx are memory chips which involve processor commands.
• 1801. There are four models: 1801WM1, WM2, WM3, WM4. They are very different between each other, as described below. There's been some indication that the 1806 and 1836 prefixes have also been used in conjunction with the WMn numbers to mean the same basic chip. The significance of the prefix value has not yet been reported.
  • WM1. Exactly an 11/03 without FIS instructions. This chip is in a 42 pin planar package. Up to four processors could be connected into a single SMP system. No such system was put into production, but a two-processor proof-of-concept system was made. Apparently this was not originally intended as a PDP-11-compatible processor.
    
  • WM2. Similar to the 11/03, complete with EIS (but no FIS) instructions. The packaging is a 40 pin ceramic DIP, and the pinout is noticeably different from the WM1. When an FIS instruction is encountered, a synchronous trap to "a hidden memory area" is taken, with vector 160XXX (exact vector not reported). When this trap is taken, the I/O page is "disconnected", and the hidden memory bank is switched in. The FPU handler is in this "hidden" memory.
  • WM3. No apparent DEC equivalent, although similar to the 11/70 or 11/73, but 11/70 is large, while WM3 is smaller then F-11! It has a 32-bit internal bus. It is not a J-11.
  • WM4. Another internally-32-bit processor. It is relatively fast, in the range of 16-25 MHz. Very limited quantities were produced due to low yields in the Soviet chip fabs (i.e., it was uneconomical to produce).

  All of these processors use standard controllers, special support chips are marked as 1801 series (1801WP1, etc). The base 1801WP1 chip is a clear universal programmed matrix with Q-Bus compatible interface logic. This allows one to build a number of custom ASICs using EPROM-like programming of non-permanent "matrix" logic. So you have a possibility to create single-chip device controllers based on that uniform base chip. In fact, most of 1801WP1-xxx chips are controllers or interface chips. All matrices MUST be programmed at the plants, there is NO WAY to reprogram/program them like EPROM chips.

  For example, some WP1 chips are:

  • 1801WP1-027 - Bubble Memory Controller (together with -032)
  • 1801WP1-030 - DRAM interface
  • 1801WP1-032 - Bubble Memory Controller (together with -027)
  • 1801WP1-033 - 8" Floppy disk controller (DX:/DY:), can also act as a parallel port controller
  • 1801WP1-035 - ASYNC port controller
  • 1801WP1-097 - 5.25" Floppy disk controller (MY:)

  Editor's note: I believe this is equivalent to a Programmable Gate Array (PGA), as opposed to an FPGA (Field PGA).

  
  

• 1811. A complete analogue of the F-11. It uses chips marked 1811. The 1811-series processor is marked 1811WM1, but there is no WM2, WM3, etc. All chips in this family are in 40 pin DIP packages. Some of the support chips are:
  • 1811WU2, WU3. Provide floating-point instructions, 46 extended instructions.
  • 1811WM1 is the main 11/23 processor.
  • 1811WU1 is the Microm containing the 92 standard instructions.
  • 1811WT is the memory extender to access 4096KB (22bit support).
• 1831. A clone of either the J-11 or the T-11. Reports differ.
• 581. The chips in the base family were all 42 pin planar packages.
  • 581IK1. Micro-programmed ALU.
  • 581IK2. Interface and maintenance chip for IK1.
  • 581RU1,2,3. Microms. RU1 and RU2 held the standard instruction set. RU3 had the EIS/FIS.
  • 581VE1. Relatively rare, this chip combined all of the above 581-family chips into a single 48 pin package.

Cyrillic and General Nomenclature

Information provided by Alexey Chupahin:

I've been told that the Cyrillic V is best represented as W in the Roman alphabet. Thus, you're referring to the DWK-*, right?

Yes, DWK and DVK is the same name.

I've gotten information on 1801BM*, 1801VM*, and 1801WM*. Are these all the same thing, or are the letters really different? I believe the WM and VM to be the same. What is BM?

Yes, VM (WM) is BM in Cyrillic. Chips, exported to German or other countrieswas marked as 'BM', if I remember, so I write 'BM' instead 'VM' or 'WM'. Unfortunately, I don't know (and my friends) what is 'BM' meaning. All russian processors (DEC or Intel compatible) are marked as 'BM'. The letter 'K' before processor model(number) means 'ceramic box'. K1811BM1 for example.

Can "MC" be interpreted as "machine"? As in, an MC prefix means this is a model/system designator, as opposed to being a chip designator?

Hmm this is very interesting question.The first line of computers ( IBM/360 compatible ) and peripheral devices designed for it was named EC in cyrillic.(ES in English). EC means 'United Series' in Russian. DEC minicomputers PDP-11 series and peripheral devices was named 'CM'.(SM) 'CM' - 'Small Series'. CM-3, CM-4 are first models., CM1425 - one of the latest, with 2MB-4MB of memory, processor 1831BM1 (J-11 analogue). CM5508 - is 10MB hard disk, CM5509 - 30MB hard disk, CM6329 - Epson compatible matrix printer.

Series 'MC' is another. Usually MC is "smaller" than 'CM'. 'MC' means personal computer equipment. MC0585 - Elektronika-85, MC0502 - DWK's.