Been trying to find out more on this subject, but there seems to be frustratingly little. I'm interested in finding out data on 1970s-1980s soviet and american radar processors to form an idea about the oft-mentioned gap between the two in this domain. i'm not a computer man so anything detailed will fly over my head unfortunately, just the raw numbers are sufficient, such as RAM, ROM, operations per second, maybe weight etc. Having apple to apple comparisons would be really helpful.
So from what i gather the N-019 Ts100 prcessor has 8k RAM, 136k ROM and 170,000 operations per second, and weighs 32kg?
But how about the relevant figures for the N-001 and SBI-16s Argon-15A? For N-001 is it same as N-019 as from what i read they use the same processor?
I found some useful info here to start with:
As to the US radars, APG-63 processor (what is it called, is it CP-1075 or something else?) seems to operate at 340,000 or 400,000 operations/sec. But what about RAM/ROM etc? I've seen figures like 16k, 24k, 96k but i'm confused which is which (seems 1000k refers to APG-70?). I've also seen 1,4 million operation per second listed, but not clear if for APG-63PSP or APG-70 (in connection with which 30 million operations per second is mentioned too).
For AWG-9, is the processor called CDC-5400B? From extrapolation based on some numbers for APG-71 (3,2 million operations/sec which is said to be 6 times more than AWG-9), it would be about 500,000 operations/second?
For APG-66, 800-900,000 operations/sec?
APG-65 at 7,2 million operations/sec?
Thanks for those, though the APG-63 stuff is way over my head (unless burried somewhere in the 400 pages is info on the processor speed/RAM etc?). And i can't see the Scribd file, am i mistaken in having seen that file freely available somewhere?
Been trying to find out more on this subject, but there seems to be frustratingly little. I'm interested in finding out data on 1970s-1980s soviet and american radar processors to form an idea about the oft-mentioned gap between the two in this domain. i'm not a computer man so anything detailed will fly over my head unfortunately, just the raw numbers are sufficient, such as RAM, ROM, operations per second, maybe weight etc. Having apple to apple comparisons would be really helpful.
To make things more complicated, some Russian technical book claims 800 000 ops per second on Ts100.
(See topic about Mig-29 avionics on this forum)
Not clear... Anyway... as you can see in Russian museum, the predecessor of C-100 (Ts100) Orbita-20 has performance of 200 ops/s in (register - register) and 100 000 ops : register - memory
Orbita-20 was used in late 70 tees design like Su-17M4 , and later. It was probably also used as subcomponent in display system for Su-27 and maybe for display system of Mig-29...
It predecessor Orbita -10 had performance 125 000 ops R-R and half of that in register-memory..
Thus C-100... well it is hardly believed that it had the same performance as Orbita-20...
For C-100 there is mentioned only "200.000" or "170000 but with no discrimination between R-R or R-M...
And for C-100 there was introduced some complex , application dependent instruction set (with 40 bit command length?) called POISK...
So ... it is possible that its performance was "170 000" but complex one and with one operand in memory... not the same as the previous one ...
BTW : in this museum there is section in English and Russian. In Russian there are more articles.
And this refers also to corresponding Russian versions of articles that have version in English ... so check both versions...
But how about the relevant figures for the N-001 and SBI-16s Argon-15A? For N-001 is it same as N-019 as from what i read they use the same processor?
I found some useful info here to start with:
I think - the above link about will explain a lot...
According to N-001 it uses C-100 series computer, (the same as Mig-29, ... there are some subtle differences that Su use version with slightly different version like Mig C100.02 and Su C100.06 - changes may lies in memory?)
SBI-16 use Argon-15K with performance quoted as 500.000 ops per second ( ... Argon-15A had 200.000)
As to the US radars, APG-63 processor (what is it called, is it CP-1075 or something else?) seems to operate at 340,000 or 400,000 operations/sec. But what about RAM/ROM etc? I've seen figures like 16k, 24k, 96k but i'm confused which is which (seems 1000k refers to APG-70?). I've also seen 1,4 million operation per second listed, but not clear if for APG-63PSP or APG-70 (in connection with which 30 million operations per second is mentioned too).
For AWG-9, is the processor called CDC-5400B? From extrapolation based on some numbers for APG-71 (3,2 million operations/sec which is said to be 6 times more than AWG-9), it would be about 500,000 operations/second?
According to APG-65... well this refers to its DSP ...
7.2 is not much for DSP ... so maybe this refers to operation on complex numbers (that is for multiplication: 4 multiplications and 2 addtition).. BTW I think APG-65 and -63PSP are very similar in this manner...
and those are there is a lot said about APG-65 in another topic of this forum (technical documentation about its soft/ hardware..)
Many thanks for your input LukaszK. I think i linked to the computer museum page earlier, but probably missed the page you showed above, very interesting details, would need to translate and save all that.
Like i said i'm not terribly versed into this computer stuff, but regarding the Ts100, from my understanding the Ts100 was lighter and cheaper than the Orbita series and at roughly comparable performance, hence it was seen as an advance and adopted widely.
The Soviet Union fell behind on the technology of shrinking componentry. Smaller components use less power, generate less heat and can run faster.
For a whole, they were able to keep up acceptable performance with acceptance of heavier devices made with older processes and clever hardware and software tweaks, but they fell off the performance curve even when accepting heavier components.
WIth regard to 'radar processors', you have to draw a firm distinction between data processors and signal processors and analogue versus digital. Also be careful of comparing raw numbers - you may be comparing oranges to bananas.
Many thanks for your input LukaszK. I think i linked to the computer museum page earlier, but probably missed the page you showed above, very interesting details, would need to translate and save all that.
Like i said i'm not terribly versed into this computer stuff, but regarding the Ts100, from my understanding the Ts100 was lighter and cheaper than the Orbita series and at roughly comparable performance, hence it was seen as an advance and adopted widely.
As to the US radars, APG-63 processor (what is it called, is it CP-1075 or something else?) seems to operate at 340,000 or 400,000 operations/sec. But what about RAM/ROM etc? I've seen figures like 16k, 24k, 96k but i'm confused which is which (seems 1000k refers to APG-70?). I've also seen 1,4 million operation per second listed, but not clear if for APG-63PSP or APG-70 (in connection with which 30 million operations per second is mentioned too).
IBM CP-1075 Central Computer
24K on initial F-15A, 96K in later upgrades. Used core memory. Replaced by CP-1075C in 1992 (using VHSIC). Does it actually do radar data processing?
AN/APG-63 initially used a hardware digital signal processor but the PSP replaced it with a programmable digital signal processor.
The F-15 Central Computer (CC), designated the CP-1075/AYK. uses the IBM AP-1 computer. It is a high speed, stored program, binary, fractional, two's complement, general purpose digital computer specifically designed ta meet the real time requirements of the high performance F-15 Air Superiority aircraft.
The CC has successfully completed formal Qualification Testing to Mil Standard 810B and Reliability Testing to Mil Standard 7818. It has proven itself flightworthy in over four thousand hours of successful flight testing at Edwards and Luke Air Force Bases.
The computer consists of three major functional sections:
o Central Processing Unit
o Memory
o Input/output
These functional sections are implemented by means of 12 plug-in modules and a single plug-in modular power supply contained in one Line Replaceable Unit(LRU) weighing about 40 lbs and occupying less than0.9 cu ft.
Table 11 summarizes the salient functional and physical features of the computer. The Central Processing Unit (CPU) is a high executing between 300,000 and 400,000 operations per second depending on the instruction mix. It executes and sequences instructions, processes interrupts, initiates input / output operations and performs the arithmetic and logical operations. The CPU has many functional similarities to the IBM System/360 computer. The Similarity extends even to the point of using the Same mnemonics for those assembly language instructions in the AP-1 which are the same or similar to instructions in the System/360.
The CPU is organized around eight 32-bit general registers, three 32-bit working registers, and a 32-bit carry-look-ahead adder. The fight general registers can be used as accumulators and index registers. Four of the general registers can also be used as base registers for addressing of up to 65,536 halfwords of memory. Instructions can be either 32 bits or 16 bits long and can operate on data that is either 32 bits or 16 bits long. The Memory contains In excess of a half a million bits of storage. It is organized as 16,384 full words (32 bits) and is addressable down to the half word (16 hits), The memory uses 1 microsecond cycle time, destructive read-out, random access, 7/13 mil ferrite cores organized into a 2 1/2 D array.
[lots more highly technical stuff snipped]
In summary, the F-15 Computational Subsystem is a Consolidated configuration of computers, in which the mission-oriented computations are performed in a single Central Computer and the sensor oriented computations arc performed in special processors in various sensor and display equipment. The Central Computer has adequate memory and execution time available for future growth. with further additional growth provided by the internal expansion capability of another 8K full words of memory. This reserve memory and speed along with the flexibility of the CC multiplex input/output system and the functional modularity of the sensor and display computations makes the F-15 Computational Subsystem easily adaptable to changes and expansions in F-mission requirements and ready to share a long and successful future with the F-15 aircraft.
When IBM began its System/4 Pi development program early in 1965, the following guidelines were established:
• Only proven technologies and proven manufacturing techniques would be considered.
• Maximum commonality with IBM's commercial technology would be maintained to insure that the resources of the entire IBM Corporation would be available.
• An off-the-shelf production capability would be developed so that reliable systems could be produced in large quantities with minimum lead time.
• The design of each computer would be general-purpose, so that each computer model could address a wide range of military and space applications.
Adhering to those guidelines, a computer engineering team designed and developed a family of computers that consists of three basic models:
• Model TC (Tactical Computer) — A briefcase-size computer for applications such as missile guidance, helicopters, satellites and submarines
• Model CP (Customized Processor) — An intermediate -range processor for applications such as aircraft navigation, weapons delivery, radar correlation and mobile battlefield systems.
• Model EP (Extended Performance) - A large-scale data processor for applications requiring real-time processing of large volumes of data, such as manned spacecraft, airborne warning and control systems and command and control systems.
In 1966, System/4 Pi computers were selected for four major military programs with a current contract value to IBM of more than $50 million. Deliveries began in March 1967.
Typical computer configurations are shown in the illustration at the beginning of this section. The characteristics of each configuration are given in Table 1-1.
The F-15 central computer controls pilot displays, weapon launch systems, and the aircraft g-load warning system. The VHSIC central computer will have improved memory/throughput capabilities and a greater mean time between failures. The operational software will be programmed in Ada in order to improve maintainability. An IBM VHSIC1750A processor forms the basis for this program. IBM is a subcontractor of McDonnell Douglas in this effort for the Air Force. The program began in September 1988. The first limited production unit is planned for 1990. The operational software is planned for early 1992.
This graph from the IDA/OSD Reliability & Maintenance study you can see the gradual development of the Signal processor and Data processor of the APG-63.
Here - article and pictures of predecessor of Oribita: TsVM-264 https://fanfics.me/message237810
Many interesting details inside:
It was used on IL-38 and Tu-142 - aircrafts for detecting and destroying submarines.
Production started in 1964.
Just 2 pictures (more in article):
ROM:
Computer:
Weight: 330kg.
Based on that computer, using micro-devices, there was created computer Orbita-1 (late 60tees), Oribita-10 , 1970) Orbita - Wiki
(See footnotes in that article - interesting about components base)
Finally, (in 1974?) there was created Orbita-20. It was based also on "micro-devices" but generally on real integrated circuits, low scale of Integration (several transistors).. from series 133.
133 was widely used in 70 tees, for example in computer of system S-300.
It was also used (along with other IC series, medium scale of integration) in C100 computer.
Generally, comparing to US counterparts, we see some delay 5 ... 7 years. While they introduced computer based on "micro-devices" in 1970 , two lates later US introduced avionics based on IC (like in F-14), even (someone claims), with first microprocessor...
So the gap was clearly visible...
6) not exactly in this topic, but: https://kaf401.rloc.ru/files/BRLSChars.pdf - table with russian radars https://kaf401.rloc.ru/files/BRLS_RUS.pdf - pictures of russian radars in one place https://kaf401.rloc.ru/files/BRLS_Foreign.pdf - pictures of western radars in one place https://kaf401.rloc.ru/PRBRLS.htm - generally on that page - there can be found references, especially for "Kopyo" radar used for example in modernized Mig-21I - for India. Diagrams, work etc. Generally this was representing the same level of technology as Zuk (late 80 / early 90)- first generation, never put in production. This can insight how original Zuk radar migth work. (Do not confuse with later Zuk-M, currently used on some planes like Mig-29K)
7) Baget processor - pdf https://kaf401.rloc.ru/ASORLD/Baget.pdf - probably this family of computers is currently (from mid 2000), used on Russian planes including Su-35 -34 and so on) To be checked....
Ok, I just released all this .. links... next post will be sth about Orbita family and C100.
I wanted to write about Oribita familly...
But.. I founded artilcle - with history of computer development company "«Электроавтоматика»
There is... much ... starting from TsVM-264 till Orbita-1, -10 -20 , C-80-300, C-80-400 C-90- etc etc..
According to Orbita - there is said much ... nothing to add, I think. The whole article is relevant and interesting.
Maybe later I will to translate /copy some fragments...
Link is to Ukrainian, known here "study files". It contains book, about general informatics. What is interesting is "attachment1".
It could be download as PDF, or view by studfile page.
It starts from that page https://studfile.net/preview/3571881/page:31/ page 308
Enjoy..
One more post with links.. Sorry for putting all of them.. but ... I hope someone can find this interesting..
Something about computers used in S300 familly 5E26 / 40U6 ; Argon 11 https://habr.com/ru/companies/ua-hosting/articles/390035/
Ferrite memories, detailed pictures (on page there are pages, with additional details, after removing cover). For example include also elements for A-50 AWACS. Radars used in S-300 system (like 64N6)
Nice to see... https://habr.com/ru/companies/ruvds/articles/648483/
Buran - computer - microprocessor Link Facebook
Article about russian computers used in space systems. https://red-army-1917.livejournal.com/58794.html
microprocessors KOMDIV - they are used widly in post 80 russian computers... somehow "obsolete" but there are versions with "radiation hardening"... https://www.niisi.ru/devel.htm
Some picture of Orbita-10 (along with some stories of people involved in development and its introduction)
technik_9 - the new blog in LiveJournal. There should be new interesting records soon.
technik-9.livejournal.com
Here page dedicated to Su T-4 "100" project... itself interesting... slightly offtopic ... but also contains in context of soviet computers and avionics https://coollib.net/b/373652/read
TO THE HISTORY OF AVIATION ON-BOARD COMPUTING SYSTEMS
R. A. Shek-Iovsepyants, Yu. I. Sabo, B. V. Utkin
(....) The appearance of nuclear submarines armed with Polaris-type ballistic missiles in the US nuclear strike forces required the Soviet anti-submarine defense to extend the detection and destruction lines of submarine missile carriers to a distance exceeding the launch range of their missiles.
At the same time, the analysis of the tasks to be solved on board the anti-submarine aircraft showed the practical lack of prospects for using analog computing technology for this purpose. The only solution was to design an onboard information and control complex with an onboard digital computer as a central computing and integration tool in a short time, almost from scratch. The Il-38 and Tu-142 aircraft of General Designers S. V. Ilyushin and A. N. Tupolev were designated as carrier aircraft. The developer of the complex was the Leningrad Research Institute of Radio Electronics of the Ministry of Radio Industry (hereinafter referred to as "Leninets"), the digital computer was assigned to the Leningrad OKB-857 of the Ministry of Aviation Industry (modern name - FSUE "St. Petersburg OKB "Electroavtomatika" named after P. A. Efimov", hereinafter referred to as OKB "Electroavtomatika").
The Plamya VT digital computer was chosen as a prototype; its development was carried out at Research Institute No. 17 of the Ministry of Radio Industry in the department of Chief Designer Karmanov.
Based on and around this work, a team was formed in OKB-857 by 1960, which carried out the design and production in 1964 of the first prototypes of the on-board digital computer, with the help of which the integration of on-board equipment could begin, and laboratory and flight tests could be carried out.
Therefore, we consider this year – 1964 – to be the year of birth of the first domestic aviation digital computer.
The distinctive properties of the first generation of digital computers were the use of discrete electronic radio components as an element base - there were simply no others at that time - and double-sided foil-clad getinax connecting boards.
The first specific applications of the above-mentioned works were the creation of the first generation of on-board digital computers TsVM-263 and TsVM-264 for two versions of aviation search radio-hydroacoustic systems installed on aircraft of the General Designers Ilyushin and Tupolev (Fig. 1 – 3). These on-board digital computers had
315
Anti-submarine aircraft Tu-142
very similar electrical circuits, but differed significantly in the layout of their devices due to the sharply different conditions of their placement on aircraft.
All these digital computers were first-generation machines, since discrete semiconductor devices were used as the main logical element base: diodes and transistors and single-sided printed circuit boards made of copper-clad dielectric. As we can see, it was possible to bypass the stage of using vacuum tubes and immediately switch to semiconductors. This had a beneficial effect on a number of digital computer characteristics. It should be noted right away that work on improving the element base was ongoing, so even within one type of digital computer, samples could differ from each other. Naturally, later samples were more advanced.
The main characteristics of these digital computers: – speed 62 thousand op/s, – RAM 256 16-bit words, – ROM 8K 16-bit words, – T = 200, (MTBF)
– Mass = 330, – Powet = 2000.
Fig. 1. General view of the on-board computer of the Il-38 aircraft anti-submarine defense system
Fig. 2. General view of the on-board computer of the anti-submarine defense system of the Tu-142 aircraft. Computing unit
Fig. 3. General view of the on-board computer of the anti-submarine defense system of the Tu-142 aircraft. Communication unit
Already in 1964, the first batch of design documentation for serial production of the TsVM-263 was transferred to UPZ and
CVM-264.
Soon, in order to expand the production front of the digital computer in Cheboksary, construction of the Cheboksary Instrument-Making Plant began on the banks of the Volga (1965–1970).
Over the past years, these factories have produced many thousands of sets of on-board digital computers, the factories have grown many times over, have significantly expanded their subject matter, and are still operating today.
In total, more than 500 TsVM-263 and TsVM-264 sets were produced by the experimental production of OKB-857 and the serial plant (UPZ).
During these years (1964 and later), the demand for onboard digital computers was so great that, despite the initial period of their development and implementation in onboard equipment, the lack of experience among developers, a perfect element base, and a proven technology for design-production-operation, OKB-857 had to accept orders for several onboard digital computers at once:
- the above-mentioned on-board digital computers for the anti-aircraft defense systems of the Il-38, Tu-142 aircraft and the air defense system,
– On-board computer of the precision navigation system for the MiG-25 operational reconnaissance aircraft,
– On-board digital computer for the Su-24 aircraft system.
The on-board digital computers, created on the basis of discrete semiconductors and single-sided printed circuit boards, could be placed only on such large aircraft as the Il-38 and Tu-142. For aircraft of the MiG-25 and Su-24 class, they were too heavy and large.
Nevertheless, such an attempt was made – a digital computer was designed and manufactured for the MiG-25 aircraft, its appearance is shown in Fig. 5. This sample was used to practice communications in a complex and programs, for an aircraft such a digital computer was unacceptably large.
The first generation of on-board digital computers, made on discrete elements and standard printed circuit boards, despite all their significant shortcomings, solved a very important problem - they proved their potential for the industry, having been installed on board aircraft in the center of the first domestic digital control complexes. It was necessary to urgently solve the following problems.
A significant reduction in overall dimensions, weight, power consumption, and increased reliability could be found by using an element base with a higher degree of integration of semiconductors. However, in those years, the domestic electronic industry could not yet offer the necessary elements and they had to be created in-house. Therefore, the OKB "Electroavtomatika" in the laboratory of the main logic element base of the on-board digital computer under the leadership of its head B. E. Fradkin, together with the enterprise's technologists, carried out search work to create microminiature elements for the second-generation on-board digital computer, which received the name - on-board digital computer "Orbita" (hereinafter referred to as Orbita). way that the second-generation on-board digital computers (a distinctive feature of the second generation of on-board digital computers is the use of micromodules as a design and technological solution for the elements of the main logical base) formed two generations: the first generation Orbita-1 - on micromodules of our own design and production PI-64 and PI-65 and the second generation Orbita-10 - on thin-film hybrid microassemblies Trapezoid-3 developed by OKB-857 in cooperation with NIITT and produced by the Angstrem plant (both in Zelenograd).
The manufacturing process of the PI-64 and PI65 dynamic elements is shown in Fig. 6. As can be clearly seen, the electronic radio elements are initially fixed by welding on parallel conductive buses, which are then connected to a polyvinyl chloride (non-flammable) film strip, which serves as a frame. The electrical circuits of the modules are formed by purposeful perforation of certain places on the conductive buses.
Subsequently , the module blanks are rolled into a spiral and fixed on an insulating base with leads for installing the modules on the boards. The modules are filled with moisture-resistant varnish or additionally insulated with compound. Various options for this moisture protection are possible. The use of new technology for dynamic elements significantly improved the characteristics of the onboard digital computer and made it possible to implement the first generation of the second-generation onboard digital computer - Orbita-1.
Using the specified technology, the design bureau completed the design of a number of on-board digital computers for several specific head objects:
– precision navigation systems for the MiG-25 aircraft; – on-board digital computer for the Su-24 aircraft;
– for the navigation and flight control system of the Tu-144 supersonic passenger aircraft.
These modifications are shown in Fig. 7 – 9.
323
Fig. 7. Modification of the Orbita-1 onboard computer for the MiG-25 aircraft
Fig. 8. Modification of the Orbita-1 onboard computer for the Su-24 aircraft
Fig. 9. Modification of the Orbita-1 onboard computer for the Tu-144 aircraft
In total, more than 500 sets of the Orbita-1 onboard computer were produced. Micromodules created at a non-specialized instrument-making enterprise could not give radical results in reducing to reduce the weight and dimensions of on-board digital computers and increase the reliability of their operation, therefore in 1966, engineers from the Electroavtomatika Design Bureau carried out revolutionary work from a technical and organizational point of view to create a serial microminiature element base for domestic on-board digital computers.
The prerequisites for this stage appeared – in the city of Zelenograd near Moscow, construction of a number of research institutes and factories for the design and production of advanced domestic microelectronics was in full swing. The above-mentioned works were oriented towards these possibilities and, in particular, towards the Research Institute of Precision Technology (NIITT) and the Angstrem factory that had begun to operate.
The developed elements did not contain inductors, transformers or delay lines and therefore could be manufactured as micromodules. This is how the Trapeziya-3 series of micromodules (5 standard elements) appeared, manufactured using hybrid thin-film technology. The technical specifications for their design were transferred to NIITT in 1966 and their serial production was already mastered at the Angstrem plant in 1967.
The appearance of one of the Trapezium-3 series microcircuits and a typical logic board is shown in Fig. 10 and 11.
The appearance of a serial logic element base became a big event, as it allowed to quickly move on to the creation of the second generation of the second generation of the on-board digital computer – the Orbita-10. A significant reduction in the weight and dimensions of these machines finally approved the digital computer on aircraft.
Moreover, it became possible to move away from the strict weight saving and introduce some important structural improvements into the on-board computer – special multiplying and dividing devices were developed, which allowed increasing the on-board computer productivity, and the possibility of doubling the accuracy of calculations was also introduced. The general appearance of two modifications of the Orbita-10 on-board computer is shown in Fig. 12 and 13.
"
The work of the same plan was the creation of another hardware modification of the Orbita-10 BCVM for use in complexes for the Tu-22 and Tu-142 Air Force aircraft and the Tu-144 supersonic passenger aircraft. In all these complexes, the BCVMs differed only in the program loaded into them; the hardware was the same everywhere.
Additionally, another modification of the Orbita-10 computer was developed – the Orbita-10-15 computer, which was a transitional link between the large and small modifications of the computer. In its design, the existing modules were used to the maximum extent. Later, this modification was widely used in frontline aviation aircraft systems or to increase the computing power in large systems. The appearance of this computer is shown in Fig. 15. This was a successful development, and the machine immediately became popular and was used at a number of sites in the form of software modifications of the basic computer."
Through the joint efforts of OKB Elektroavtomatika, UPZ and ChPZ, more than 4,000 sets of the Orbita-10 digital computer were produced in the form of four hardware and thirteen software modifications.
This same period, the work carried out at domestic microelectronics enterprises, as well as world experience, clearly indicated the only path to development and solution of the above-mentioned problems – maximum use of solid-state element base. Guided by the current situation, work was carried out at the Electroavtomatika Design Bureau, headed by the experienced specialist B. E. Fradkin, with the participation of engineers N. T. Trenkin, V. P. Kiselev, L. I. Mogilevsky, E. M. Kadinov, A. M. Stepantsova and others, as well as together with engineers from one of the enterprises in Voronezh to modify the already existing "Tulip" series - solid-state dynamic elements with an average integration level. The resulting series of integrated circuits was named "Tulip-3". Serial production was carried out by the same enterprise in Voronezh.
The use of solid-state logic elements, miniature resistive and capacitor assemblies, as well as multilayer connecting printed circuit boards made it possible to increase the speed of the new digital computer, called Orbita-20, to 200 thousand short operations per second, reduce the number of microcircuits used by 2-3 times compared to the Orbita-10 digital computer and, as a result, double the reliability, reduce the dimensions, reduce the weight, and simplify the technology of its production.
The Orbita-20 digital computer, which uses integrated circuits as its main logical element base, is a third-generation machine. (..)
The preliminary design of the Orbita-20 computer was successfully defended in 1971, a year after the start of its development.
In addition to the widespread use of the technology of modifications of the basic BCVM Orbita-20, in a number of BCVM applications it was possible to use software modification while maintaining the unchanged hardware. In this case, only the contents of the memory devices were subject to change - this was the most progressive solution supported by a successful composition of hardware. It was possible to develop the number of software modifications to 21 for a specific hardware modification of the BCVM, which can be considered a record for specialized equipment.
In subsequent years, the transformer-based, flashed-through, special-wire, permanent memory device of the Orbita-20 digital computer (ROM) was replaced everywhere with a more advanced semiconductor reprogrammable ROM with rewriting using ultraviolet erasure of information and programmers. After the loaded programs had been fully worked out, the reprogrammable ROM was replaced with a one-time programmable one using the method of burning fuse jumpers on the crystal of the memory microcircuits. In this way, greater reliability of preserving the programs loaded into the memory was achieved.
The external appearance of the basic modification of the Orbita-20 digital computer is shown in Fig. 16, and a typical logic panel with integrated circuits “Tulip 3” installed on them is shown in Fig. 17.
The device panels are visible. The block on the left is the secondary power source. On the right are two easily replaceable ROM blocks. By replacing these blocks, software modifications of the machine were created.
Fig. 16. Basic model of the Orbita-20 onboard computer with the top covers removed.
Fig. 18 shows the external appearance of a large modification of the Orbita-20 on-board computer. Such machines were used in the on-board complexes of heavy aircraft such as the Tu-22 and its modifications, and other similar ones. Such complexes required a significant (for those times) volume of memory devices, which increased the number of identical memory units.
The front of application of the Orbita-20 digital computer turned out to be wide. Below, as an illustration, is a table of developed and serially produced hardware and software modifications of this digital computer. A number of projects were brought to working samples, but were not released later for objective reasons (not shown in the table).
In total, more than 15 thousand sets of various modifications of this onboard digital computer were produced.
The high level of development is also evidenced by the fact that a large number of different modifications of the Orbita-20 onboard computer have been successfully operated at the main facilities of the Russian Air Force and Civil Aviation and beyond to this day, i.e. for more than 30 years.
Table of modifications of the Orbita-20 digital computer, produced in the period 1973 – 1990.
My comments:
NK - navigation (K)Complex
PrNK - Targeting- Navigation Complex
Orbita-20-23 - seems to be not used as main computer(s) of Mig-29. Initially it was intended to be used on that plane, but around 1980 , they decided to be use C-100, what requires porting of software. This requiresThe name "23" suggest that it was used in mig-23 or rather on Mig-27. Orbita-20-6 was used in CEI - central display system - see below.
Orbita-20-22 was used on Su-17M4/-22M4
Orbita-20-750 - it was intened to be usded on Ka-50 (4 computers).
Some version also Also Su-25T (2 computers)
Fig. 20 shows the indication system for the MiG-29 aircraft, which uses a modification of the Orbita-20 digital computer. In the figure, the three left blocks are the Orbita-20-6 digital computer. In the center are the cockpit indicators: the navigation and tactical information indicator and the direct vision indicator (TV).
Continue...
Next article describes CWM-80... This computer was thought as Oribita-20 replacement..
These onboard digital computers were supposed to appear no later than the time when the capabilities of the Orbita-20 onboard digital computer would be exhausted.
In an effort to maintain continuity in the development of their subject matter, as well as systematically improving the characteristics of their products, the employees of the Electroavtomatika Design Bureau began developing a digital computer on a large-scale integrated circuit (LSI) in 1973. In preparation for this work, functional circuit diagrams for these LSIs were developed in-house with an orientation toward the production capabilities of the enterprises of the Ministry of Electronic Industry.
The fourth generation digital computer was named TsVM80 (another name is “Gamma”). During its design, processor sections of the 1804 series of integrated circuits and integrated circuits of semiconductor memories were used – semi-permanent with electrical recording of information and ultraviolet erasure and permanent with the burning of fusible jumpers in the semiconductor structure.
This development was supposed to, as before, for the third-generation Orbita-20 on-board computer, become the basis for the industry as a central computer for various aviation complexes and for controlling individual systems as peripheral computers, as well as for simulators.
One of the distinguishing features of the fourth generation of on-board computers was that the command system in them was compatible with ground-based general-purpose computers, such as the M6000, CM4, and a standard system bus used in general-purpose computers began to be used to combine the functional modules included in the computer.
The latter circumstance allowed on-board programs to be tested on laboratory stands equipped with universal computers, which significantly reduced the software development time. This was essentially the first attempt to use proven industrial technologies in specialized digital computers.
The combined structural diagram of such a stand, called the programmer's workstation (PWS), is shown in Fig. 22. In the structural diagram of the PWS, the term "processed digital computer" in the block The TsVM80-400 is not accurate, since it is not the TsVM that is being processed, but the program loaded into the TsVM memory.
The first generation of the TsVM80 received the index TsVM80-1ХХХХ, in which the first two characters XX define the hardware modification, and the next two – the software modification.
The basic TsVM80-1ХХХХ became its first modification – TsVM80-10300 for the navigation complex of the passenger aircraft Tu154. (...) Third generation computing machines TsVM80-3ХХХХ had panels that were half the size of the first generation and, as a result, had a smaller overall size. These digital computers were used mainly on light aircraft in the built-in equipment version. They were intended to operate as part of inertial systems, weapons control systems, and indication systems.
These digital computers were developed starting in 1981–1982, and by 1985 there were four hardware modifications, and there were more software ones.
Modifications of the CVM80-3XXXX: TsVM80-303XX is a set of two modules for operation as part of an inertial system, TsVM80-302XX – monoblock, radial type user interface according to GOST 18977-79. These on-board digital computers were used mainly at civil facilities, TsVM80-307XX – mentioned above, TsVM80-308XX – monoblock, user interface after-serial type (multiplexed) according to GOST 26765.52-87. These digital computers were used mainly at military facilities,
Fig. 23 shows one of the modifications of this digital computer. The last one is the fourth generation of the digital computer TsVM80 – TsVM80-4ХХХХ
were created in the form of hardware modifications, starting in 1985, to ensure the operation of flight and navigation systems and indication systems of heavy aircraft and to service the control and management system power plants of passenger cars Fig. 23. External appearance
airliners Tu-204 and Il-96.
In the TsVM 80 of the fourth TsVM80-308XX generation, a command system was implemented compatible with the CM4 general-purpose computer, which by that time had become the most widely used in industrial applications. (...) For example, Fig. 24 shows the basic TsVM80-400 and several of its modifications: for monitoring the operating modes of engines of mainline passenger aircraft – TsVM80-401XX(Fig. 25), for use in flight and navigation systems – TsVM80-402XX (Fig. 26) and TsVM80-404XX (Fig. 27)
CVM-80-40... was used in commercial planes: Ił-96 and Tu-204 as navigation computers (modern looking like ...) , I cut description (one can see on page) , only paste some pictures:
Continue ... the next paragraphs of describes development of late 80 and early 90.. as we know... there was collapse of Soviet Union and funding.. hard to describe shortly... The next generation based on microprocessors , VLSI.. some epizodes..were using 386/486...
The development of the basic digital computer family models: SB3541 and SB3542 should be considered separately. This work was a reflection of another attempt by the state to introduce the development of specialized digital computers in the "defense five" ministries into some organizational framework and to unify this work. This attempt, like several earlier similar attempts, was not crowned with success. In this case, the main reason was the state restructuring of the national economy that began during this period and its negative consequences.
SB3541 and SB3542 were an additional generation of the fourth-generation digital computers. It can be considered that SB3541 is at the origins of the creation of digital computers in the Electroavtomatika Design Bureau since the 90s.
The basis for the development was the decision of the Military-Industrial Commission of the Military-Industrial Commission under the Central Committee of the CPSU and the Council of Ministers
USSR No. 456 of 16.12.1986. Based on this decision, the working group of the Military-Industrial Complex formulated general technical requirements for the work called "family of basic computers", briefly - SBXXXX, where the symbols X formed a specific type of digital computer. The work was distributed between a number of ministries, the aviation industry was assigned two types of on-board SB: the monoblock SB3541 and the single-board digital computer for embedding in equipment - SB3542.
Based on the analysis conducted in the late 1970s - early 1980s, a program was developed to create a family of unified SBECMS for use on mobile objects of all classes. This program was approved in 1984 by a decision of the State Commission. In accordance with it, LNPOEA began work on creating unified SBECMS - SB3541 and SB3542 with the "Elektronika-32" type architecture, and the Argon Research Institute - SB5140 with the "POISK" architecture. Unfortunately, this program was not completed. Only individual machines were developed (and then with a significant delay in terms) - SB3541 based on MPK1839, SB5140 and SB5580 based on BMK 1537KhM2. The experience of work on the creation of the SBEVM was taken into account, and the idea of interspecific unification itself received its logical conclusion in the development of the Baget family of computers (lead developer NIISI RAS), which also includes the aircraft-use machines Baget-53, Baget-52, Baget-63 and Baget-62.
"Furthermore, despite the gradual reduction of funding from the Ministry of Defense and the growing difficulties in carrying out the work, the OKB team found the strength to continue developing the SB3541 (basic upper-level digital computer) and the SB3542 (basic digital computer for embedded applications) digital computers. These products are shown in Fig. 32 and 33.
The digital computer was based on the domestic microprocessor set of the 1839 series, developed by NIITT, Zelenograd; the architecture and command system were compatible with the VAX11/750 general-purpose computer.
Thanks to the applied microprocessor set MP1839, the speed of the digital computer was increased several times compared to the 80-4XXXX digital computer.
The SB3541 digital computer under development had a performance speed of up to 5 million operations per second with the prospect of increasing the performance speed to 10 million operations.
On the basis of SB3541 digital computer, modifications of the 90-5XX digital computer were subsequently developed, which ensured the construction of multi-machine upper-level computing systems. The 90-5XX digital computer was supposed to be used in prospective complexes for MiG aircraft.
Further work was stopped due to the complete cessation of their funding by the Ministry of Defense.
Nevertheless, the groundwork created during the development of SB3541 and SB3542 was used in the development of the specialized multiprocessor SCVM VM94 (up to five processors) for the ground-based defense system.
During these years – 1994 and later – a development was carried out that did not leave a noticeable trace, but its specificity deserves attention. It was a large complex work, but
we are interested in its component part – the on-board computer. The customer's requirement to use imported components: processor, memory, interfaces from Intel was unusual for the specialists of OKB "Electroavtomatika".
As a result of the design work, a prototype of a digital computer was created that has the following characteristic features:
– the processor was presented in the form of a single-board micromachine, which used the integrated circuit of the i80486 microprocessor,
– the radial and serial channel controller was designed based on the i80386 integrated circuit,
– the system bus was assigned a parallel type bus
ISA (Industrial Standard Architecture) is actually a standard bus for personal computers such as IBM PC and compatibles,
– for the first time in the practice of designing on-board digital computers, a very progressive technology of using mezzanines on the ISA bus was applied, (...)
The developers themselves assessed the work as quite conservative - by this time the LSI processor and controller i80486 and i80386, as well as the ISA bus, could not be considered advanced solutions, they could be replaced by other more advanced ones. On the contrary, the mezzanine technology and the use of the environment LabVIEW is progressive to this day.
However, most of these decisions were dictated by the customer and were implemented in full.
The resumption of development and modernization of aviation complexes in the late 90s required the developers of digital computers to further increase their computing power.
An important step was to change the structure of the operating part of the arithmetic unit of the new digital computers. Following the trends of the development of the microelectronic base, taking into account world experience, the developers switched from using the universal architecture LSI to RISC processors. These LSI were supposed to be reproduced in Russia; they were easily available for import.
In 1998–2000, in order to reduce costs and development timeframes and use of previously created resources, OKB specialists "Electroavtomatika" designed a processor module using the R3081 LSI (in 2003 it was replaced by the 1890 series LSI, developed by the Research Institute of the System Intelligence of the Russian Academy of Sciences) with a speed of 25-50 million operations per second, compatible with modules previously developed for the 90-5XXXX digital computer. The development of such a module allowed a sharp increase in the computing capabilities of the 90-5XX digital computer. The digital computer with the operational part of the arithmetic unit with the RICS architecture was assigned the code TsVM 90-6XX.
More than 10 modifications of the 90-6XXXX computer were developed with different memory capacities and numbers of external interfaces.
The developed digital computers are used in aviation on the aircraft Su24, Su25, An74, Yak130.
(...)
These blocks combined not only computing functions, but also specialized ones.
The design bureau produced more than 20 different types of such blocks, in particular:
– multifunctional color indicators (MFCI); – specialized units for various applications,
example block BFVI - block for generating visual information. As an example, Fig. 36 shows the cockpit of a light combat aircraft with two MFCs and one control panel installed in it
Based on the 90-5XX and -6XX digital computers (and their modules), the Electroavtomatika Design Bureau has developed a wide range (more than 40) of various computing equipment, control and indication devices for a wide range of end objects: modifications of MiG and Su aircraft, helicopters, long-range aircraft, S-80, Tu, ground systems, etc.
Next paragraphs contains tables with parameters of computers
263/264, Orbita-10 -20 -CVM-80-3... CVM-80-4 CVM-90-...
All are known...except CVM-90-... so I copy them here:
Next there are general compositions architecture of those systems... also interesting ... but too much for this post...
The ends of this article (dated on 2010) relates to some development using massive parallel computing... and describes state of art of that time. Those systems are probably used on current generation of Russian equipment ..
The on-board digital computers developed in the 1990s no longer meet the requirements of advanced projects, and therefore, starting in 2000, the Electroavtomatika Design Bureau began work on creating a number of on-board digital computers that would ensure the implementation of the requirements of aviation systems for fifth-generation aircraft.
In the first direction, design and functional modules have been developed using high-speed domestic processors of the Multkor 1892VM3 type (developed by the State Unitary Enterprise Scientific and Production Center ELVIS) and the KOMDIV type: 1890VM2T and 1890VM3T (developed by the Scientific Research Institute of System Engineering of the Russian Academy of Sciences).
In the second direction, a prototype of a promising multiprocessor system has been created. The system consists of three computing modules united by a PCI system bus, which provides fast data exchange between computing modules; network interfaces are used as an external interface.
Ethernet type sys.
The total performance of such a system is at least 0.6 billion op/s. Figure 37 shows a prototype of this system.
The proposed computing system is essentially open, uses standard system buses and, as its practical implementation, provides the user with the resources of a multiprocessor system containing three levels of performance.
The first level is high performance up to 50 million op/s for solving traditional tasks of on-board integration and performing host functions.
The second level is higher – up to 200 million op/s for processing rapidly changing signals (radar and similar information).
The third level is ultra-high performance up to 2-3 billion op/s for solving problems of increasing the level of “On-board intelligence” of complexes (see above).
All three levels can act together.
It should be noted that the first level has been mastered by OKB Elektroavtomatika in the form of a number of operating prototypes, some of which have been delivered to the customer.
At the second level, a standard module based on a microprocessor was designed.
The third level is represented by a prototype of a fast matrix processor, on which a performance of 1.3 10^9 op/s was obtained. The appearance of the matrix processor is shown in Fig. 38. It consists of three boards: two boards of the actual matrix processors and one control board.
Next (hopfully final) post I will write some words about C100 computer (used for example in Mig-29 and Su-27 base variant) . This might put slightly a new light on its performance. I will compare them with Orbita-20 - based from some architecture, and performance point of view..
LukaszK many thanks for the above posts, truly fascinating stuff, even if like i pointed earlier the technical intricancies are beyond my head. TsVM-80 is the computer used in Su-27 right? Strangely the book doesn't seem to mention the Su-27 using the TsVM-80-30XX series.
LukaszK many thanks for the above posts, truly fascinating stuff, even if like i pointed earlier the technical intricancies are beyond my head. TsVM-80 is the computer used in Su-27 right? Strangely the book doesn't seem to mention the Su-27 using the TsVM-80-30XX series.
Su-27 does atleast use C-100. However I wouldn’t be surprised if another computer was required for some purpose due to all the additional systems it has compared to MiG-29, such as expanded Navigation system and Narcissus datalink control setup.
LukaszK many thanks for the above posts, truly fascinating stuff, even if like i pointed earlier the technical intricancies are beyond my head. TsVM-80 is the computer used in Su-27 right? Strangely the book doesn't seem to mention the Su-27 using the TsVM-80-30XX series.
According to C100 (1983)- its specification is well know, for example from Russian Computer Museum, that was referred several post above.
Comparing to Orbita-20 (1974) it has somehow more memory RAM/ ROM , but this also depends on particular version of Orbita.
According to speed - it seems that C100 brings no progress: Orbita-20 was stated 200k/op s and C100 - depending on sources 200k op/s 180k op/s (most common) or 170k op/s.
The second question is why Orbita was being considered as choice a decade later, while C100 was avaiable?
It was used in Su-25T (2 computers) or Ka-50 (4 computers TsVM-20-751).
Actually, I do not know the exact answers on above questions.. only some speculations.
Some technical details about both computers:
If you want get some knowlege - Orbita was described somehow in more details in (links already pasted above)
(BTW it is said to be used in Mig-31 system, but as we know (?) that it was rather somehow upgraded Argon-15K.
But changes may be only in performance not architecture. BTW It is said that Mig-31 also use Orbita-20, probably for display processing, some diagrams related to composition of computers are also presented )
Some description of C100 is in book describing Mig-29 avionics. I attached part that describes C100 (and Orbita-20, and another processor) as attachment. Computers used on Mig-29 (export variant) are described from page 29, and C100 computer from page 31.
Some general notes:
Orbita family and so on... was created by company "Electroavtomatika" and C100 and Argon-15 (and other) by company "Argon".
in all sources, well, they might be some errors, or at least inconsistencies. So I try to take all information with some distance and If possible it is better to confirm not in single source.
This version has weight 38kg, volume 63.5dm3, used power 300VA.
(Comparing to C100 : weight 32kg, mass 48 dm3, power 275W)
What is somehow strange, Orbita-20 family has range of parameters that differs an order magnitude: for example mass from 8.5kg to 60kg...
That is strange. I understand that later variants used solid state memories. Compact memories required much less power than ferrite , ... so cooling might be also reduced and then more compact power supply, ... all this contributed to decrease of weight.
But still hard to explain such mass reduction ( 8.5 from 39 kg in case of ЦВМ-Орбита 20-23) without changes in central processing unit.
So possible explanation is that also a central processor was refreshed and based on somehow new element base. That also results in increased processing speed. But this is only speculations
Although in the "Electroavtomatika article" they mentioned about some "exotic computer" based on Orbita-20
A unique modification of the BCVM was a transitional variant between the Orbita-20 BCVM and the next generation BCVM. The product was a multi-purpose three-channel computing system with the possibility of flexible hardware and software switching of channels. The appearance of the BCVM with the cover removed is shown in Fig. 21.
Here, for the first time in the practice of OKB "Electroavtomatika", a conductive method of cooling the on-board computer unit was used. The work was brought to the point of delivery of prototypes, but was subsequently closed due to the cessation of work at the customer's main facility.
Then there is referenced fig 21 - what seems to be CVM-80. But this might be a mistake, and also on fig19 there is some "exotic system"
Anyway, maybe this was that designed to be used in new aircraft (Ka-50, Su-25T) ?]
Anyway.. if this lower mass is true (I mean 8.5kg) such class computer (if not a mistake... it is repeated in two sources...).. well this might be some argument, why such compact computer, even if its processing speed was the same, was considered as choice in some newer designs like in Su-25T or Ka-50.
What is also worth mention here - is speed. 200k ops/s seems to be not much, but note that the same speed in register - memory operation., what is not always the case.
Some another argument, why Orbita-20 might be an attractive choice in ... might be ... the convenience and easiness of programming. As we see below... C100 - has much wider set of commands including custom one (POISK archtecture), and represents CISC type of machine. https://en.wikipedia.org/wiki/Complex_instruction_set_computer
But it might be dificult to get a full potential of such machine.
Another, factor might be , just familiarity with system.
Command set of Orbita-20, its architecture were well known ... it has some similarity starting from CVM264, Orbita-10.
Many solutions from 70 tees were based on Orbita-20. Designers/programmers were familiar with that system.
They developed, tested and introduced systems based on that computer. Also there were written algorithms and programs - tested and worked on real hardware.
Why to use a new computer if existing solutions works well and performance of Orbita-20 was enough? While memory constrains might be relaxed with using a new solid state memories..
Ok, back to topic... some key points related to Orbita design:
The TsVM-Orbita 20 is made using the principle of block design based on unified functional blocks.
The TsVM-Orbita 20, the external appearance of which is shown in Fig. 2.2…2.3, structurally represents a supporting frame 4 (Fig. 2.2), on the base of which the following blocks are placed:
BP20 - power supply 1 ,
BVC20 - digital computing unit 2 ,
BPP20 - block of constant memory 3 .
Fig.2.2. External appearance of the TsVM-Orbita 20 (front view).
1 – power supply unit BP-20; 2 – digital computing unit BVC-20; 3 – constant memory unit BPP-20; 4 – mounting frame; 5 – puller; 6 – shock absorber APNM4; 7 – shock absorber APNM5; 8 – handle; 9 – control connector.
The read-only memory device ROM is designed to store numbers and commands that make up the program of the problem being solved and has the following technical characteristics:
transformer-type storage device on U-shaped cores,
The execution of a typical command can be divided into the following steps:
command selection- at the address specified by the program counter, the command is retrieved from memory;
command decoding - finding out its meaning, selecting operands from registers;
performing an operation , if necessary accessing memory - calculating the physical address;
memory access ;
memorizing the result .
Commands from the ROM are selected sequentially, in ascending order of numbers, by the command address coming from the UAC (processor). The selected command goes to the UAC (processor), where it is divided into two parts;
- the address part containing the number of the RAM (ROM) cell from which the number (constant) must be selected ;
- an operation code which, after decoding in the UAC (processor) in the form of control signals , is used in the digital computer when performing operations.
Based on the control signals of the UAU (processor), arithmetic operations are performed on one or two numbers.
The first number (constant) is selected from RAM (ROM). The RAM number is the information converted by the input converters of the UVV, coming from the system sensors or the result of intermediate operations. The ROM constants are programmed into the storage device in accordance with the program.
The second number is the result of the previous operation, which is stored in the UAU (processor). The results of the operations can be converted by the output converters of the UVV and output to the system receivers if there are corresponding control signals and addresses.
"To increase the speed of the TsVM-Orbita 20, the UAU (processor) provides for combining commands, i.e. simultaneously with the execution of the "K" command, the "K+1" command is decoded and the "K+2" command is selected from the ROM."
This is nothing more than pipelinig https://en.wikipedia.org/wiki/Instruction_pipelining
That means that operations might be done in cycle of processor. What is not clear - that along fetching command into pipeline, if data from memory was also fetch in advance.. The same number of ops in register - register (R-R) and register-memory (R-M) suggests that, but I found no evidences.
System of command was fixed, and contains about 76 commands.
After some reading... I can say that operation with operand from RAM took 1 cycle, operands with accessing constants (data) from ROM 2 cycles. Another observation - there are not too many internal registers in ALU... (actually 2 ) , but as access to memory is in one cycle - this is not big problem (however storing information in RAM needs to be done separately)
Processor had separate device for multiplication / division.
Below schema of processor. On the left - elements for memory access with PC (program counter). In the middle - processor control unit - to store and decode instruction https://en.wikipedia.org/wiki/Instruction_pipelining
That is more or less all about Orbita-20.
Now some words about C100. It is more difficult to describe - as - the way how it works - it not so clear/straightforward.
But first some data. From Mig29 technical description book:
2.3.2. Brief characteristics of the BTsVM C1OO.02.
BTsVM C1OO - made according to the backbone-modular principle.
Operation of the BTsVM units is controlled at two levels - software (operator) and microprogram.
G. Main technical characteristics of the BTsVM C1OO.02. From
1. Type - multi-address, synchronous, parallel action.
2. Number system - binary; the code of negative numbers
is represented in the additional code. ^
3. Bit capacity numerals - 16 binary digits (including the sign digit) - word; the fill is fixed before the most significant numerical digit after the sign. The digital computer can operate with bytes (8 binary (binary) digits) and double words (32 binary digits).
4. Operator (command) capacity - variable.
5. Micro-command capacity - 36 data bits and 4 control bits.
6. Internal organization of the digital digital computer operation - microprogrammed. *
7. Number of operators (commands) - no more than 256; number of micro-commands - 128. ,
8. Performance (number of short operations per second):
on short operations of the "register-register" type (RR) - not less than 800000; on short operations of the "register-memory" type - not less than
250000; on operations of the multiplication type with loading of operands sending the double-bit result to memory - not less than 80000.
9. The capacity of the RAM is 4096 18-bit words, including two control bits; the capacity of the long-term memory:
for programs - 64K 9-bit words, including one control
bit; for the microprogram - 8K 40-bit words, including eight control bits;
10. The addressing method is direct, relative and indirect.
11. The access cycle time: to the RAM - not more than 2.4 μs; to the long-term memory - not more than 1.2 μs.
12. The type of information exchange with external devices:
exchange of binary serial bipolar code in an asynchronous manner or by the READY signal with a maximum exchange rate of 1.1 Mbit/s;
exchange parallel 16-bit code with a maximum exchange rate of 200,000 words/s.
13. Number of interrupt classes - 7.
(...)
18. On-board computer weight - no more than 32 kg.
19. On-board computer volume - no more than 48 dm3.
20. Mean time between failures - no less than 1000 h; average on-board computer recovery time - no more than 30 min (according to TU).
(...)
22. On-board computer element base: main (silicon integrated circuits of 134, 133, 136, 130, 106, representing potential systems of elements with transistor-transistor logic (TTL) based on the logical circuit (NOT/OR-NOT); a special element base that provides connections both between the elements of the on-board computer and between the on-board computer and external devices and systems.
This description was already known.
The new one is From "340 pages book", that This confirms C100 performance...
My comments:
First of all, C100 executes commands based on micro-program. This is somehow typical in modern processors - instead of hardware circuits - what and how command is implemented - is described in set of commands. https://en.wikipedia.org/wiki/Microcode
On C100 - command code was described by 1 byte - so 256 command were possible: half of them were fixed: the second half were design specific, it means - that for specific tasks , there could created custom command that perform specific operation (POISK)
This way, there might be complex commands those did quite complex/long sequence of processing in multiple processor cycles.
Method of executing of such micro-programs - is ... complicated..
Each command is 36 bit word: and in short... allows to set controls to separate blocks of processors , with micro-instruction code for each block (up to 5 codes).
Each block of processor - has own decoder - that decodes that micro-instruction command into control signals,that control this block. This allows - to somehow -to do parallel work by each block in respect to others.
For example - arthritic - logical unit (ALU) might do operation (like adding) while in parallel, there might be done multiplication and some data transfer to/from memory.
How much this can be done parallel - it is not clear for me, as probably micro-command have own limitation (for example - in command there was pointed some internal block's registers.
Each micro-command was done in a single processing cycle.
Each block might have - own registers (also to be pointed in micro command).
Micro-command memory contains 8K of such 40 (36) bits micro-command... (4bits are for error correction code) ... effectively 36 bits is used.
This is comparatively a lot: 4.5 bytes per micro-command, and 8K of such commands.. this means... 36kB of separate ROM only for defining micro-commands..
Comparing to ROM of programs... what was 64kB (or 128kB - depending on sources) (not counting CRC bits...).
So more than 20% of whole ROM is to define micro-commands. So micro-commands probably were important and constitute for performance.
Having possible 256 commands and 8K - this means, that , on average - a single CPU command (assembler) conists of 32 micro-commands. Of course - some of them, were simpler (probably built in) those might took from 1 up to let say several cycles.
This left memory for others - to be possible to much much longer.
How this actually was in practice.. well.. program needs to be decomposed , and some processing need to be decoded as single assembler command. That command, should be somehow optimized - on hardware level (micro-instructions) - having in mind possibility of making concurrent processing by block's processor.
In theory - this would be possible to get a much from hardware, in practice... well.. who knows...
BTW C100 had the second co-processor for input-output operation.
Description of work is in "340 book". on page 189. I read, somehow understood, but I will skip at this point... only some selected fragments:
All micro-operations are coded, and micro-command decoders are introduced into
the blocks, ensuring the recognition of each micro-command by its code and the formation of the corresponding control signals. Since most blocks
of the computer can operate in parallel, control micro-commands are fed to these blocks simultaneously. This provides greater, in comparison
with the hard logic circuit, possibilities for combining
the stages of command execution.
The central processing unit (CPU) of the Ts-100 digital computer includes:
arithmetic logic unit (ALU);
arithmetic multiplication-division unit (MADU);
program control unit (PCU);
microprogram control unit (MPU);
data control unit (DCU);
random access memory unit (RAMU);
data switch unit (DSU).
Information exchange between the units is carried out via four main highways: the command and constant highway (LE), the microcommand highway (LF), the address highway (LA) and the data highway (LD).
The AL and BAUD units together constitute the arithmetic-logical unit and are used to perform arithmetic and logical operations on bits, bytes, words of single and double length.
The possibility of placing micro-instructions of several blocks in a volume of 36 bits and
the desire to combine the operation of as many blocks as possible in time led to the fact that in the processor micro-instruction format, the KSB fields of the block micro-instructions are combined into the processor structure code (PSC) field. (...)
All in all - comparing computers - based on number of instructions per seconds, itself, it is hard task (like comparing apples to coco-nuts). Well it is hard to take any conclusions. For sure, C100 was much more advance system than Orbita-20.
The cited numbers (I mean 170k ops/s) for sure are true, but, probably are result of averaging of performance from a real application code. While some instruction might took several cycles of processor and be really complicated, its productivity might be significantly higher, than a pure number suggests.
Its productivity, well I suppose were much higher than Orbita-20.
But here -it is still hard to do any conclusions: from one hand number of simple operations R-R is 800k that is 4x highier. Still not clear how many registers were accessible in ALU: if several (as in modern CPUs) - some operation might be done in R-R mode without frequent access to memory.
Moreover it was possible to combining some work of processor's blocks- like did multiplication while addition and access to memory, in single cycle. That suggests productivity might be even higher than 4x. On the other hand cited performance R-M was only 250k, not far from Orbita-20 in this respect (R-RAM).
According to my memory, on Su-27 they wanted to use single Argon A-15Z (like Zukowski). But then they changed to a pair of C100.
Similarly on Mig-29 initially they wanted to use Orbita-20. But then also switched to C100. However, they changed also to a pair of C100. But as we see not everybody followed the same path - for some reasons.
Initially, it was planned to use a Research Institute of Radio Electronics (NIIRE, Leningrad) common advanced digital computer design. Slippage of schedules on development of this led to adoption of NITsEVT's Argon-15 as used on the MiG-31 instead. This was modified to Argon-15Zh standard for this application, which was component N001-05 in the radar. Several early N001 test radars used the Argon15Zh.
Meanwhile, Mikoyan were adopting NITsEVT's newer A2009SIS (Ts-100) computer for the N019 radar.
The Argon-15Zh/N001-5 was 5 times more effort to produce than the Ts-100, according to the Oktyabr plant tasked with producing both computers. Performance of the Argon-15 and Ts-100 was broadly comparable, but Ts-100 used an entirely Soviet element base and was lighter and easier to make. The POISK architecture was optimised for the specific tasks and very efficient and every effort was made in its design to make it ruggedised and reliable for airborne applications. It used medium-scale integration LSIs.
As part of the unification effort between N001 and N019, the A2009SIS (Ts-100) computer was adopted for the N001, to be replaced by the more advanced A2009BIS in 1981.
Other digital computers used on the Su-27:
GNIIRS Electronika-NTs in the TKS-2 comms system
Electroavtomatika's Orbita-20-51 in the SEI-31 integrated display system made by Electroavtomatika (funny, that).
The flight/navigation system was supposed to use the GNIIRS Electronika-NTs, but the VNIIRA A313 Manyovr-V used on the MiG-31 was used as an interim and seems to have stuck.
Its important to realise that "Orbita" wasn't a single design. Orbita-1, -10 and -20 are basically 3 generations of digital computer. Electroavtomatika's TsVM-80 is basically 4th generation Orbita.
So the Su-27 doesn't have a TsVM-80 computer in it's avionics? I must have read in Y. Gordon's book that it has the TsVM-80, and also for example this bit online:
Avionics Technology has progressed in the former Soviet Unio and now Russia, but at a slower pace than in the West. And today that progress faces economic roadblocks. At major […]
www.aviationtoday.com
And also this bit saying the Su-27K has a TsVM-80-302 (have to hover on the Information button)
Its important to realise that "Orbita" wasn't a single design. Orbita-1, -10 and -20 are basically 3 generations of digital computer. Electroavtomatika's TsVM-80 is basically 4th generation Orbita.
It's been fascinating to read through the links as well a few bits online. So i have another torrent of questions (this also after going over Overscan's guide to russian avionics).
What processor was in the early N-011 with planar array antenna? The N-011M PESA has the Ts-200 with a speed of 75 million op/sec. On Overscan's it is said the Ts-200 was intended for the basic N-011, though the timing seems a bit early for me -second half of 1980s- for that kind of speed in a soviet processor?
Did the basic Zhuk indeed only used a Ts-101 or something more advanced?
The Zaslon-A is said to have an improved processor, is it an improved Argon-15 or something else? Also, what was used on Zaslon-M for MiG-31M?
There is also mention of Ts-181, Ts-175 and Ts-501 in various radar entries, any more data about them?
What processor is used on the B-004 radar of Su-34?
Lastly, any info as to the processing speed of the Solo-35 on Su-35? Ditto for the N-036 radar processor (Elbrus?).
I'm also interested as to what kind of processing was used in "Soyuz-Sintez" prototype from NPO Istok to provide such good results. And what does Russian N011M uses on Su-30SM, as Ts200 is for early variants.
Take-off has repeatedly covered
various spheres the State Ryazan
Instrument-Making Plant (GRPZ) – a
major Russian manufacturer of
airborne radars – operates
in, including its productionising of the
active phased array and development
of heliborne radar. In addition, GRPZ
develops a family of airborne digital
computers and airborne computer
systems for various applications.
Take-off’s correspondent Yevgeny
Yerokhin has been to the plant again
and seen Nikolay Andreyev, chief of
the airborne computer department of
the corporate scientific and technical
centre. What was the beginning of the work
on computers at your plant? This was an
absolutely new field for GRPZ, wasn’t it?
Let me start with the fact that our
specialised department, whose principal
task is the development of airborne digital
computers, is turning 10 in November this
year. How did everything begin? GRPZ is a
specialist in making radars for fighter jets.
As is known, there used to be partnering
among manufacturing plants throughout the
country during the Soviet times, and GRPZ
would receive airborne digital computers
for its radars from its subcontractors. With
the beginning of the economic upheaval in
the 1990s, problems with computer supplies
cropped up, and the plant was unable
to make its end-products without them.
In that situation, the most urgent task
we faced was learning to produce certain
types of dedicated computers (various
versions of the Ts100 in the first place) to
fit the radar systems we made. To learn
producing airborne digital computers as a
new product class to the company, and to
learn to support production, the corporate
scientific and technical centre set up the
airborne computer department in 1999 and
hired relevant staff. This minimal task was
fulfilled with success, but we did not stop
at that.
Development of an in-house airborne
digital computer came next, right?
Right. The Tikhomirov-NIIP research
institute in the town of Zhukovsky was
upgrading the radar fitting the Sukhoi Su-27
fighters family. Advanced high-performance
digital computers had to be introduced
to the radar. Analysis of the computers
offered by various manufacturers indicated
that none of them met Tikhomirov-NIIP’s
requirements to additional radar modes
in technical or economic terms. Against
this backdrop, GRPZ on agreement with
Tikhomirov-NIIP decided to launch
development of its own digital control
computer to fit radars in the course of the
upgrade.
The performance specification for
developing the SOLO-54 airborne digital
computer was issued in 2002. The computer was
developed extremely quickly. In cooperation
with Tikhomirov-NIIP, the SOLO-54 passed
its ground and flight tests as part of the N001V
radar fitting the Su-27SM, with the trials
wrapped up in 2004. The same year, GRPZ
launched production of the radar SOLO-54
version with an expanded interface set in
2005. The SOLO-54 radar was appreciated
by the customer, Tikhomirov-NIIP, that used
the computer in three versions of its radar –
the N001V, N001ER and N001VEP. This
was the beginning of a new field for GRPZ
to explore.
The successful completion of the
SOLO-54 and SOLO-54.01 development
stage gave impetus to GRPZ to start
development of more airborne digital
computers and airborne computer systems
mostly for airborne radar applications. The
development of new types of computers
followed several paths, e.g. development of
high-performance specialised computers
for control tasks, programmed signal
processing devices wrapped around digital
signal processors and computer systems
uniting these types of computing devices.
Development of advanced computers
relied on up-to-date system interface
and module-and-unit design standards,
including international ones. Functional
modules of airborne digital computers are
designed depending on the requirements
with the use of the Euromechanics
constructs of the 3U or 6U dimension types.
Airborne digital computer modules are
integrated in sealed cases with conductive
heat removal and external air cooling (ATR
Short construct). Such a design provides
effective protection from outside exposure
and good electromagnetic compatibility
with other electronic systems.
The scientific and technical centre is
known to have developed a number of other
airborne digital computers…
A case in point is the development of a
small-size airborne computer system to fit
the heliborne radar being developed by the
plant too. The airborne computer system
is part of the avionics of the Mil Mi-28N
attack helicopter. It is wrapped around the
Compact PCI system bus interface with the
use of Euromechanics 3U dimension type
in the 1/2 ATR Short case. The airborne
computer system embodies a control
computer with a relevant front-end interface
set, programmed digital signal processor,
radar clock driver, intermediate frequency
controlled amplifier and AD converter that
are part of the radar’s reception path.
At present, the computer is being tested,
with GRPZ productionising both the
radar and the airborne computer system.
Another stage of development of airborne
radar computers was the emergence of
the SOLO family of multiprocessor
airborne digital computers wrapped
around new architecture – a common
commuted computing environment. The
work is being done under GRPZ’s Chief
Designer Andrey Pershin and involves
the latest achievements in intermodular
connections on the basis of the PCI Express
high-capacity serial interface. The common
commuted computing environment allows
the removal of limitations on the capacity
of the multiprocessor system as compared
with known architectures wrapped around
the common system bus.
Based on the common commuted
computing environment architecture,
two airborne digital computers were
developed to fit the Irbis-E radar under
development by Tikhomirov-NIIP for
Sukhoi Su-35 fighter. They were the SOLO-35.01 airborne digital computer
designed for radar signal processing and
a SOLO-35.02 airborne digital computer –
high-performance multiprocessor
control computer. The SOLO-35.01 has
a control data processor module and up
to four digital signal processor modules
connected via the PCI Express system
interface. The SOLO-35.02 comprises up
to four data processor modules mounting
mezzanine interface modules made to
PMC (PCI Mezzanine Cards) standard.
Both computers are made of modules of
the Euromechanics 6U dimension type in
the 1/2 ATR Short case, interconnected
via a high-speed datalink and constitute
the integrated computer system of the fire
control radar.
The computers have passed ground tests
and are being flight-tested as part of the
radar on a flying testbed at the Defence
Ministry’s Main Flight Research Centre
(GLITs).
We are developing the SOLO-21 airborne
digital computer using the same common
commuted computing environment
architecture to equip prospective aircraft.
The SOLO-21 is a high-capacity computer
comprising a multichannel ADC, several
digital signal processor modules, several
data processor modules and several interface
modules. The computer uses Euromechanics
6U dimension-type modules and PMC
interface modules and is housed by the modular
ATR Short case. The avionics suite consists
of two SOLO-21 airborne digital computers
linked with a high-performance optical line.
The development of the computer has been
mostly completed, several prototypes have
been made and their trials are beginning. Is there any competition on the Russian
airborne digital computer market and how
would you describe GRPZ’s position on the
market? What are your advantages?
Of course, there is competition, but we
have got rather solid positions and good
prospects. The SOLO digital computer
family is progressing from the rather basic –
and, nonetheless, still popular – SOLO-54
to the most sophisticated SOLO-21
computer system designed to fit the future
fighter.
As far as advantages are concerned, our
developments are based on the developed
production facilities of a manufacturer
plant. The degree of sophistication of
these technologies heavily influences the
functionality and design of our products
that, in turn, influence the evolution of our
technologies. I mean precise machining
and multilayered printed circuit boards
(PCB), including flexible/rigid PCBs,
in the first place. The establishment of
the airborne computer assembly shop
at the plant proved to be a considerable
progress in pursuing this line of work.
The workshop is fitted with automatic
surface-mounting assembly lines with
soldering quality optical and X-ray control
units, a laser installation to make metal
patterns for solder paste application, and
module moisture-protection polymer
coating equipment. To adjust, check and
test computer modules and units, the
workshop is furnished with dedicated
automated workplaces and equipment to
conduct environmental and mechanical
tests.
In addition, the advantage offered by
the plant’s airborne digital computer
development is also the development running
in parallel with the productionising efforts.
This enables us to slash the productionising
and manufacturing costs radically. Is there a chance for SOLO airborne
digital computers to be used in spheres other
than aviation?
Certainly. They have already been used
in other branches. We have developed the
BMS-M baseline mobile station for the
Army automated control system’s command
echelon. It is a sort of special-purpose PC.
We are developing five types of computers
to fit ground radars. Their constructs are
different from those of airborne digital
computers, but share the concept and
architecture. In addition, these days,
we have more orders for development of
various digital computers for the Air Force
and other services. Such a prospect and
demand for our products let us face the
future with confidence.
The SOLO-54 computer is based on the RISC processor with MIPS architecture. It has good
performance and a good set of front-end interfaces compared with other similar computers. The
SOLO-54 is housed by a compact case and requires no forced-air cooling. This is especially
important for radar modernisation when the airborne digital computer had to be integrated in a
system with the well-established layout. The real-time operating system used by the airborne
digital computer supports multithreaded applications and has a POSIX-compatible interface.
The efficient software development and debugging means available allow a drastic reduction in
the time needed to develop application software for airborne radars.
I'm also interested as to what kind of processing was used in "Soyuz-Sintez" prototype from NPO Istok to provide such good results. And what does Russian N011M uses on Su-30SM, as Ts200 is for early variants.
The heart of the airborne radar is a digital computing complex, including all system software, was developed under the leadership of A.N. Korolev, who headed department 341 of division 34 at that time. For the first time in our country, a radar data processor was created - a multi-machine computing complex based on single-board computers "Electronics-81" with a total capacity of one million operations per second, with serious system software (the operating system was developed by lead engineer Shishkin Vlad world Igorevich) and advanced test tools.
It is important to mention, that Sintez-10 was somehow experimentally program, that aims to develop technologies and being proof of concept. It was lead not by company /institute designed for creation of radar complexes (like NIIP or NIIR) but company who develops components for radars... In my opinion it was really successfully... maybe even >>too much<<.. and was "cut down". Results were pass to the companies specific for creation of radar systems...
Individual technical solutions and the element base of the radar were then used by enterprises of the radio industry in subsequent developments of radar systems, and for the development of the new Istok signal processor, a detailed technical assignment was prepared.
For testing this was ideal (cheap, COSTsolution, with good performance), but rather not directly to be moved into military systems.
They developed "own" PSP using LSI technology.
Generally, topic of late 80, and later development is interesting, but somehow difficult to find consistent sources... and it is somehow to be explored. Plenty of information can be found. But as there were many companies that developed own systems, it is difficult to develop somehow systematic view of that process. What can be done, quickly is to share some links. One can view them by his own.
There was mentioned a lot about Baquiette "platform" (familly of processors / Operating system) in reference https://kaf401.rloc.ru/ASORLD/Baget.pdf , including signal processor you mentioned. This reference was created in early 00... so maybe it does not contain somehow final iteration of that platform.
Generally post 80 development is connected with names such as "Baquiette"(platform), Solo (mentioned above) KOMDIV (famillly of Russian domestic processors). Those technologies from early /mid 00 - are currently used in designs developed in 00 like Su-35, Su-57 , Ka-50 , Su-34 and so on...
ELBRUS (VLSI processor familly). According to my understanding family of systems based on Elbrus-8 and later is a basis for recent platform like in Su-34M, Su-57M and so on.
Links:
KOMDIV (familly of processors, included radiation hardened)
The KOMDIV-32 (Russian: КОМДИВ-32) is a family of 32-bit microprocessors developed and manufactured by the Scientific Research Institute of System Development (NIISI) of the Russian Academy of Sciences.[1][2] The manufacturing plant of NIISI is located in Dubna on the grounds of the Kurchatov...
The KOMDIV-64 (Russian: КОМДИВ-64) is a family of 64-bit microprocessors developed by the Scientific Research Institute of System Development (NIISI) of the Russian Academy of Sciences and manufactured by TSMC, UMC, GlobalFoundries, and X-Fab.[1] The KOMDIV-64 processors are primarily intended...
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
