Digital Integrated Circuits

This section encompasses monolithic integrated semiconductors of all types, including microprocessors, memory and other integrated devices.

Semiconductor Logic

The earliest solid state computers employed discrete germanium diodes and transistors. Improvements in materials and fabrication lead to an industry switchover to silicon devices, which were then progressively reduced in size and increased in integration. The earliest integrated circuits were hybrid devices, which utilized multiple discrete diode and transistor dies interconnected with fine gold wire and encapsulated in a single metal canister.

The 1960s saw the introduction of the monolithic integrated circuit, in which multiple transistors and diodes were etched upon a single die. Monolithic integration improved through the 1960s, mostly through the efforts of Fairchild Semiconductor and Texas Instruments. This continued until the early 1970s when various manufacturers began offering single-chip arithmetic-logic units, giving birth to the microprocessor. Intel's consumer-accessible 4004 is often credited as the first true microprocessor, but other devices of similar integration were already being used by the US military at the time of the 4004's introduction.

Semiconductor Memory

RAM, along with the microprocessor, introduced the fourth generation of computers, an era we are still living in today. The two main classes of RAM are static RAM (SRAM) and dynamic RAM (DRAM), the primary memory of virtually all computers built since the mid-1970s. Static RAM uses transistors to store the bits, while dynamic RAM uses capacitors, resulting in a cheaper RAM that requires periodic refreshing to retain its contents. Static RAM requires no such refreshing, and will retain its contents as long as power is supplied.

Another important type of semiconductor memory is the PROM, programmable read-only memory. Unlike RAM, PROMs are non-volatile and will retain their contents either forever or until a specific erase procedure is executed, depending on the type of PROM. PROMs come in various single-write and rewritable types, the most readily visible type being the UV-erasable/electrically rewritable UV-EPROM. UV-EPROMs (and UV-EPROM-equipped microcontrollers) are easy to identify by their square, round or rectangular erase window, through which the monolithic die is typically visible. PROMs have been largely replaced by modern flash memory.

Magnetic Bubble Memory

Magnetic bubble memory is a non-volatile data storage medium invented at Bell Labs in 1967. Bubble memory uses a thin magnetic film on a garnet substrate, which forms cylindrical domains when constricted under a magnetic field. These domains, or bubbles, each store one bit of data. The bubbles are created by a generator signal, pushed around the film in racetrack-like loops, and eventually detected by a sense amplifier. Unlike semiconductor memories, bubble memory is sequential access, rather than random access. Conceptually, it is like a tiny magnetic diskette and drive, but with no moving parts. Instead of the disk moving, the bits move.

Bubble memory was a promising technology that was positioned to replace all other forms of memory, but was quickly outpaced by semiconductor memory (speed and cost) and hard disk drives (capacity and cost). The computer industry was already shaped around a well-established split between fast, small-capacity main memory and slower large-capacity storage, and bubble memory was squeezed out in both categories as a costlier, inferior alternative.

Bubble memory may still be occasionally used in some military and industrial applications, where extreme ruggedness and nonvolatility are a requirement. Bubble memory is inherently radiation-hardened: it can withstand the photoelectric effects of a nuclear event, when implemented with specialized nonstop logic circuits which prevent partial loop rotation.

Former manufacturers of magnetic bubble memory include Hitachi, Intel, Motorola, Rockwell and Texas Instruments. In 2008, IBM re-branded the bubble memory concept as 'racetrack' memory, a sort of nano-scale bubble memory which uses an array of tiny wire loops to carry the magnetic domains.

AMD 7-6967
 AMD 7-6967

This rare device is an unidentified, early AMD IC with a gold and white ceramic sandwich carrier and gold leads. The size of the IC suggests that it might be a ROM.

[View Detail]

AMD Am2900
 AMD Am2900

Devices included in this entry:

AMD Am2901BPC bit-slice ALU (40-pin plastic DIP)
AMD Am2903DC bit-slice ALU (48-pin cerDIP with ceramic cavity lid)
AMD Am2903DC bit-slice ALU (48-pin cerDIP; pictured in thumbnail)
AMD Am2910DM-B 12-bit address sequencer (40-pin cerDIP)
AMD Am2940DC DMA generator (28-pin plastic DIP)
AMD Am2960DC 16-bit error detection and correction unit (48-pin cerDIP)


AMD Am2900 is a family of 4-bit bipolar logic devices which function together as a bit-slice processor. The Am2901 and Am2903 arithmetic-logic units are the core of the series, with other tasks being handled by about 50 different Am29XX devices. The Am2900 family was used in the design of various DEC and Data General minicomputers, and is widely regarded as the most successful bit-slice architecture.

AMD Am2900 Family Data Book (PDF)

[View Detail]

Burr-Brown DAC80 & DAC812CM
 Burr-Brown DAC80 & DAC812CM

Devices included in this entry:

Burr-Brown DAC80 D/A converter (24-pin custom cerDIP; pictured in thumbnail)
Burr-Brown DAC812CM D/A converter (24-pin hermetic metal DIP)


The Burr-Brown DAC80 and DAC812 are monolithic 12-bit D/A converters designed for industrial applications. The DAC812CM is a high speed device, with much higher accuracy and conversion speed ratings than the DAC80. The DAC80 is housed in an unusual hermetic side-brazed ceramic carrier, while the DAC812CM is sealed in a hermetic metal canister.

Burr-Brown DAC80 & DAC80P Datasheet (PDF)
Burr-Brown DAC812 Datasheet (PDF)

[View Detail]

Cypress CY7C341-35RC
 Cypress CY7C341-35RC

The Cypress CY7C341 is a UV-erasable switch matrix PLD with 12 arrays and 192 logic cells, in an 84-pin ceramic PGA. Despite the Cypress branding, the die of this device was fabricated by Altera.

Cypress CY7C341 & CY7C341B Datasheet (PDF)

[View Detail]

Fairchild F8 / 3850 Series
 Fairchild F8 / 3850 Series

Devices included in this entry:

Mostek MK3850N-3 ALU (40-pin plastic DIP)
Mostek MK3853P SRAM interface (40-pin cerDIP; pictured in thumbnail)


The Fairchild F8 (3850 series) is a somewhat odd 8-bit microcontroller. The parent device is the 3850 8-bit ALU with 64 bytes of scratchpad RAM and two I/O ports. Instruction decoding is decentralized: instead of the 3850 decoding its own instructions, the program storage unit, PIO and memory interface chips handle the decoding.

[View Detail]

Ferranti Interdesign DTC228B
 Ferranti Interdesign DTC228B

The DTC228B is an unidentified chip by Ferranti Interdesign, a rare and historically significant manufacturer.

[View Detail]

Fujitsu MB61H303A
 Fujitsu MB61H303A

The Fujitsu MB61H303A is a somewhat unusual device from a time before the era of programmable logic devices. The MB61H303A is a 3900-gate CMOS gate array, which would be configured by the user on a computer-aided engineering workstation. The design for the metalized interconnect would then be sent to Fujitsu for fabrication. Demand for such devices rapidly evaporated after the introduction of UV-erasable PLDs.

Fujitsu CMOS Gate Arrays Datasheet (PDF)

[View Detail]

Hewlett-Packard 1818-2xxx Series
 Hewlett-Packard 1818-2xxx Series

Devices included in this entry:

HP 1818-2056A (28-pin grey trace cerDIP)
HP 1818-2251A (28-pin cerDIP)
HP D1818-2800 (28-pin perforated grey trace LCC)
HP A1818-2817 (28-pin perforated grey trace LCC)
HP C1818-2833 (28-pin perforated grey trace LCC; pictured in thumbnail)


HP is notorious for anonymous ICs. A HP part number tells you nothing, except that it is a part used by HP in one of their systems. Some HP parts are proprietary in-house devices, while others are simply re-branded off-the-shelf chips from another vendor such as Intel or Mostek. Part number cross references can be found for HP's renumbered third party devices, but their in-house parts are either poorly documented or completely unidentified.

The 1818-2xxx devices pictured here are believed to be proprietary 28-pin ROMs for various HP systems, but their exact function is unknown.

[View Detail]

Hewlett-Packard 5061-30xx Series
 Hewlett-Packard 5061-30xx Series

Devices included in this entry:

HP A5061-3010 hybrid microprocessor (82-pin ceramic LCC)
HP 5061-3011 No-EMC variant hybrid microprocessor (82-pin ceramic LCC; pictured in thumbnail)
HP D5062-3001 AEC variant hybrid microprocessor (107-pin ceramic LCC)
HP D5062-3001 AEC variant hybrid microprocessor (107-pin ceramic LCC)
HP C5061-3012 TACO tape drive controller (56-pin ceramic LCC)
Hewlett-Packard A5061-3020 shift register (89-pin ceramic LCC)
Hewlett-Packard C5061-3055 ROM (40-pin ceramic LCC)


The 5061-3010 is an extremely advanced 16-bit, 10 MHz hybrid microprocessor, developed as the CPU of the Hewlett-Packard 9825A desktop calculator. Though it originated at HP's calculator division in Loveland, Colorado, the 5061-3010 is no mere calculator chip. When HP sets out to design a microprocessor, they come back with a 100kg minicomputer, squeezed into a volume about the size of a pack of cigarettes, heatsink included.

The core of the 5061-3010 consists of three NMOS chips: the binary processor chip (BPC), extended math chip (EMC) and the input/output controller (IOC). The BPC is basically a single-chip version of a HP minicomputer, utilizing the core 211X set of 59 instructions. The EMC handles another 15 instructions, dealing mostly with BCD math, and the IOC performs 12 I/O instructions. Each NMOS die is approximately 4.7mm on a side, the total surface area about 66mm^2. HP's NMOS II fabrication process, specifically developed for the manufacture the 5061-3010, could not support larger dies without annihilating chip yields.

HP also manufactured a number of different 5061-30xx hybrid NMOS support chips in leadless chip carriers, including specialized controllers and memory devices. Additionally, the hybrid package style was chosen for the HP-80 financial calculator hybrid ROM (1813-0024), and the HP-65 programmable calculator hybrid CPU (00065-60218).

[View Detail]

Hewlett-Packard MC2
 Hewlett-Packard MC2

Devices included in this entry:

HP 1AA1-6003 (48-pin ceramic gold trace LCC; pictured in thumbnail)
HP 1AA6-6004 (48-pin ceramic grey trace LCC)


The 1AA1-6003 and 1AA1-6004 are variants of Hewlett-Packard's silicon-on-sapphire MC2 processor, originally designed for the ill-fated HP 300 Amigo microcomputer and later used to greater success in the HP 3000 Series 33 minicomputer. The original HP 300 chipset consists of the 1AB2-6003 processor control unit (PCU), 1AB3-6003 register, address, skip and special (RASS), and the 1AB4-6003 register, arithmetic and logic unit (RALU).

While the 1AA1-6003 is currently unidentified, the 1AA1-6004 is known to be a MC2 processor modified as a HP-IB controller.

[View Detail]

Intel 1101A
 Intel 1101A

Devices included in this entry:

Intel C1101A (16-pin cerDIP; pictured in thumbnail)
Microsystems International ML1101A (16-pin cerDIP)


Introduced in 1969, the Intel 1101 256-bit SRAM is the first commercial implementation of silicon gate MOS VLSI. The 1101 established PMOS as the technology of choice, versus Intel's 3101 64-bit Schottky TTL SRAM which was developed in parallel. The successful PMOS technology was then used in the breakthrough 1103 DRAM device, dubbed the 'core killer'.

The 1101A is an improved pin-compatible variant of the 1101 which requires fewer power supplies.

Intel 1101A & 1101A1 Datasheet (PDF)

[View Detail]

Intel 1103 & 1103A
 Intel 1103 & 1103A

Devices included in this entry:

Intel 1103 (18-pin plastic DIP)
Fairchild 1103A (18-pin cerDIP; pictured in thumbnail)


Introduced in 1970, the Intel 1103 is the first commercially available DRAM. The 1103 is a 1kbit PMOS device based on a three-transistor memory cell, designed in 1969 by Honeywell engineer William Regitz. The 1103 was Intel's breakout product, displacing core memory as the predominant technology and establishing Intel as the leading pioneer in semiconductor RAM.

Perhaps the most noteworthy customer of the Intel 1103 was Hewlett-Packard, who assisted Intel in improving the 1103 design and implemented the 1103 as the primary memory in their early 9800-series desktop calculators.

Intel Memory Design Handbook - August 1973 (PDF)

[View Detail]

Intel 1702A & 8702A
 Intel 1702A & 8702A

Devices included in this entry:

Intel B1702A UV-EPROM (24-pin ceramic sandwich DIP; pictured in thumbnail)
Intel B8702A UV-EPROM (24-pin ceramic sandwich DIP)


Introduced in 1972, the Intel 1702 is the first commercially available UV-EPROM, capable of storing up to 256x8 bits. The 1702 was invented by founding Intel employee Dov Frohman, after discovering that stored charges in the broken gates of defective ICs would alter their behavior. By introducing an insulator that breaks down while exposed to ultraviolet radiation, an erasable and rewritable device could be manufactured.

The Intel 1702A is a revised variant of the original 1702 with a much faster programming time. While the 1702 draws a peak current of 1.2A during programming, the 1702A draws only 200mA, allowing for a much higher programming duty cycle. This reduces the programming time from 20 minutes on the 1702 to only two minutes on the 1702A.

[View Detail]

Intel 2608 & 2708
 Intel 2608 & 2708

Devices included in this entry:

Intel B2708 UV-EPROM (24-pin ceramic sandwich DIP)
Motorola MCM2708L UV-EPROM (24-pin cerDIP; pictured in thumbnail)
National Semiconductor MM2708Q UV-EPROM (24-pin cerDIP)
Texas Instruments TMS2708JL UV-EPROM (24-pin cerDIP)
Texas Instruments TMS2708JL UV-EPROM (24-pin ceramic sandwich DIP)
Intel B2608 OTP EPROM (24-pin ceramic sandwich DIP)
Intel D2608 OTP EPROM (24-pin ceramic sandwich DIP)


Introduced in 1975, the 2708 is one of Intel's second-generation UV-EPROM devices. Unlike the 1702, which was fabricated using a PMOS manufacturing process, the 2708 is fabricated in NMOS, and has a much faster 350ns access time. In addition to the 2708, Intel manufactured the 2608, which is functionally identical to the 2708 but sold as a one-time-programmable (OTP) EPROM, with a permanently masked erasure window. The 2608 was cheaper to manufacture than the 2708, as the absence of erasure function testing lowered production costs.

[View Detail]

Intel 2716
 Intel 2716

Devices included in this entry:

Intel C2716 UV-EPROM (24-pin cerDIP; pictured in thumbnail)
Mostek MK2716T UV-EPROM (24-pin cerDIP)
Texas Instruments TMS2716JL UV-EPROM (24-pin cerDIP)
Texas Instruments TMS2516JDL UV-EPROM (24-pin cerDIP)


The Intel 2716 is a third-generation EPROM device, with a capacity of 2048x8 bits. The 2716 is among the most widely adopted of early EPROMs, and was a popular choice of arcade gaming hardware developers in the late 1970s and early 1980s.

At the same time Intel was developing the 2716, Texas Instruments was working on their own 2716 design, which required more supply voltages than the Intel device. Intel's 2716 surpassed the TI 2716 in sales, and TI soon released the 2516, a clone of the Intel 2716. This unusual circumstance still causes confusion to this day, as TI's 2716 is not pin-compatible with Intel's 2716, while the 2516 is entirely equivalent.

[View Detail]

Intel BPK-72
 Intel BPK-72

Devices included in this entry:

Intel 7110-1 bubble memory (20-pin ceramic LCC; pictured in thumbnail)
Intel D7220-1 bubble memory controller (40-pin ceramic sandwich DIP)
Intel D7230 current pulse generator (22-pin ceramic sandwich DIP)
Intel D7242 sense amplifier (20-pin ceramic sandwich DIP)
Intel D7250 coil pre-drive (16-pin ceramic sandwich DIP)


The Intel BPK-72 family consists of the 7110 high density 1-megabit bubble memory device and its associated support chips. The 7110 has 256 data loops of 4096 bits each, arranged as 2048 512-bit pages. The 7110 has separate input and output tracks, as well as other advanced architectural features.

The example pictured here is a 7110-1, the first incarnation of the 7110, with a 20-pin leadless carrier and medium ambient temperature tolerance. 'A' variants (7110A-1, 7110A-4, 7110A-5) have standard tinned through-hole pins. Variants with the -4 suffix (7110-4, 7110A-4) have low temperature tolerance, while -5 variants (7110-5, 7110A-5) have high temperature tolerance. Additionally, there are 'AZ' types with through-hole pins (7110AZ-1, 7110AZ-4) whose performance characteristics are currently unknown.

It should also be noted that some early documentation refers to the 7110-4 as the 7110-0 or simply 7110 with no suffix, and the 7110-5 is listed as the 7110-2. It is unknown whether this abandoned numbering scheme ever made it to production.

Intel subsequently developed the 7114-1, a 4-megabit module in a 20-pin package with tinned through-hole pins.

Intel Magnetic Bubble Storage Data Catalog (PDF)
BPK 72 Prototype Kit Datasheet (PDF)
BPK 72 Prototype Kit User's Manual (PDF)
Intel Solutions Bubble Memory Application Note (PDF)
7110 1-Megabit Bubble Memory Datasheet (PDF)
7220-1 Bubble Memory Controller Datasheet (PDF)
7230 Current Pulse Generator Datasheet (PDF)
7242 Dual Formatter/Sense Amplifier Datasheet (PDF)
7250 Coil Pre-Drive Datasheet (PDF)
7254 Quad VMOS Drive Transistors Datasheet (PDF)

[View Detail]

Intel MCS-4
 Intel MCS-4

Devices included in this entry:

Intel P4002-1 40-byte RAM (16-pin plastic DIP)
Intel D4004 microprocessor (16-pin ceramic sandwich DIP; pictured in thumbnail)
Intel P4008 8-bit address latch (24-pin plastic DIP)
Intel P4009 I/O interface


The Intel 4004 is generally considered to be the world's first microprocessor. The 4004 was designed by Federico Faggin, an employee of Intel who was predominantly responsible for the design of the Intel 4040, 8008 and 8080, and later went on to found Zilog and develop the incredibly successful Z80 microprocessor. Originally intended as a proprietary calculator chip, the 4004 went on to great commercial success. The 4004 now lives on as a collector's item, fetching truly obscene prices for certain carrier variations.

In addition to the devices pictured here, the MCS-4 chipset includes the 4001 ROM and 4003 shift register.

[View Detail]

Intel MCS-48
 Intel MCS-48

Devices included in this entry:

Intel P8021 microcontroller (28-pin plastic DIP)
Intel C8035-8 C-1 microcontroller (40-pin cerDIP; pictured in thumbnail)
Intel P8039 microcontroller (40-pin plastic DIP)
Siemens SAB8039P microcontroller (40-pin plastic DIP)
Intel P8041A microcontroller (40-pin plastic DIP)
Intel C8741A microcontroller (40-pin windowed cerDIP)
Intel MD8741A/B microcontroller (40-pin windowed ceramic sandwich DIP)


Introduced in 1976, MCS-48 is Intel's first family of microcontrollers. MCS-48 devices have a modified Harvard architecture with onboard RAM, and some models have an integrated program ROM as well. The first three MCS-48 devices, the 8035, 8048 and 8748, were introduced in 1976 and implemented in NMOS.

MCS-48 was later replaced by the popular MCS-51 family of microcontrollers.

MCS-48 Microcomputer User's Manual (PDF)

[View Detail]

Intel MCS-51
 Intel MCS-51

Devices included in this entry:

Intel P8031AH (40-pin plastic DIP)
Intel P80C32 (40-pin plastic DIP)
Intel C8751H-88 (40-pin cerDIP; pictured in thumbnail)
Signetics SC87C51CCF40 (40-pin windowed ceramic sandwich DIP)


The Intel MCS-51 is family of powerful Harvard architecture single-chip microcontrollers, introduced in 1980. The parent device is the 8051, with 128 bytes of RAM, 4kB of onboard ROM and a UART interface. Like the earlier 8048 microcontroller, there are numerous variants of the 8051. The 8031 and 8032 lack an onboard ROM, while the 8751 includes a user-programmable EPROM. Early devices are fabricated in NMOS, while later devices are CMOS, indicated in the part number with an intervening 'C'.

[View Detail]

Intel MCS-80
 Intel MCS-80

Devices included in this entry:

Intel C8080A microprocessor (40-pin cerDIP; pictured in thumbnail)
Intel D8080A microprocessor (40-pin ceramic sandwich DIP)
Intel D8238 system controller (28-pin ceramic sandwich DIP)
Intel C8316A ROM (24-pin cerDIP)
Intel Am9080APC microprocessor (40-pin plastic DIP)


Introduced in 1974, the 8080 is Intel's third microprocessor, and their second 8-bit microprocessor. The 8080 is an extension of Intel's original 8008 architecture, and the direct predecessor of the 8085, 8086 and 8088. Unlike the PMOS-based 4004 and 8008, the 8080 is implemented in NMOS.

While the first-generation Intel microprocessors were best suited to small scale systems like calculators and terminals, the 8080 was selected as the CPU of a number of early microcomputers, such as the Altair 8800 and IMSAI 8080.

[View Detail]

Intel MCS-85
 Intel MCS-85

Devices included in this entry:

AMD Am8085APC microprocessor (40-pin plastic DIP)
Intel D8085A microprocessor (40-pin ceramic sandwich DIP)
Intel P8085A microprocessor (40-pin plastic DIP)
Intel P8085AH-2 microprocessor (40-pin plastic DIP)
Intel D8202A DRAM controller (40-pin ceramic sandwich DIP)
Intel D8755A UV-EPROM (40-pin windowed ceramic sandwich DIP)
Toshiba TMP8755AC UV-EPROM (40-pin windowed cerDIP; pictured in thumbnail)


Introduced in 1976, the Intel 8085 is a direct descendant of the 8080, with simplified power requirements and a slightly amended instruction set. The 8085 also incorporates the functions of two 8080 support devices: the 8224 clock generator and the 8228 system controller. The Intel MCS-85 family includes the 8085 microprocessor and a large number of support chips.

[View Detail]

Intel MCS-86
 Intel MCS-86

Devices included in this entry:

Intel D8086 microprocessor (40-pin ceramic sandwich DIP; pictured in thumbnail)
Intel D8088-2 microprocessor (40-pin ceramic sandwich DIP)
Intel C8206 polynomial error detection circuit (82-pin ceramic LCC)
Intel C8207-2 DRAM controller (82-pin LCC)


Introduced in 1978, the Intel 8086 is the first widely adopted 16-bit microprocessor, and the first implementation of what would come to be known as the x86 architecture. A year after the 8086, Intel released the 8088, an 8086 variant with a 8-bit external data bus. IBM's selected the 8088 as the CPU of their first IBM personal computer in 1981 a decision which would have far-reaching effects to this day. Virtually all modern personal computers are built upon an x86-based microprocessor, manufactured by either Intel or AMD.

The MCS-86 includes the 8086 and 8088 microprocessors, as well as a long list of support devices.

[View Detail]

Intel i860
 Intel i860

Introduced in 1989, the Intel i860 (80860) is a 32-bit VLIW RISC microprocessor with an integral 64-bit FPU and graphics unit. Described by Intel as a "Cray on a Chip", The i860 was intended for use in supercomputers and high end workstations, competing against MIPS and SPARC processors from the era. Unfortunately, the i860 was plagued with performance problems related to its poorly optimized compilers. By the time its compilers had improved enough to provide reasonable performance, other RISC architectures had already surpassed the i860. The i860 failed to achieve commercial success, and was discontinued in the mid-1990s.

The i860 shown here is an A80860XR-40, a 40MHz variant packaged in a 168-pin ceramic PGA.

[View Detail]

Intel i960
 Intel i960

The Intel i960 is a 32-bit RISC microprocessor, introduced in 1985. The i960 began as the 960MX, a CPU originally designed for the failed Intel / Siemens BiiN fault-tolerant multiprocessor computer. While the BiiN project was a disaster, earning the derogatory label Billions Invested In Nothing, the i960 was a commercial success, widely adopted in the early 1990s as an embedded microcontroller.

[View Detail]

Mostek MK4332D-3
 Mostek MK4332D-3

The Mostek MK4332D-3 consists of a pair of MK4116E 16kB DRAMs in 18-pin leadless carriers, mounted to an 18-pin ceramic breakout DIP to produce a single 32kB 'RAM-PAK' device. Such hybrid devices were designed to ease adoption of the new leadless carrier technology, by providing a LCC footprint with conventional through-hole pins.

[View Detail]

Mostek MK5011P
 Mostek MK5011P

Introduced in 1972, the MK5011P is an early variant of the original MK6010 single-chip calculator, developed by Mostek for the Busicom LE-120A handheld calculator. Busicom, a Japanese calculator manufacturer which had previously built large four-function desktop machines based on TTL and core memory, sought to employ LSI technology in their upcoming products. Busicom collaborated with Intel to develop the MCS-4 (4004) chipset, the first commercially available microprocessor. They simultaneously contracted Mostek to develop the MK6010, the first single-chip calculator. These two devices would set the stage for VLSI's dominance, spawning thousands of siblings, descendandants and competing products through the 1970s and onward.

The MK5011P was used in a number of similar four-function desktop calculators, including the Colex 1200LP, Privileg 1200 and Unitrex 1200.

[View Detail]

Mostek MK5017BA
 Mostek MK5017BA

The Mostek MK5017P is an early PMOS clock controller with multiplexed display control, widely used in digital clocks in the mid-1970s. The MK5017P is capable of 50Hz or 60Hz operation, has an internal tone generator and requires only one power supply. The MK5017P was initially distributed in three variants: AA (alarm clock), AN (clock radio) and BB (calendar clock). The example pictured here is a MK5017BA, used in the Heathkit GC-1092D Time-Calendar Clock. The difference between the BA and BB calendar types is currently unknown.

Mostek MK5017P Datasheet (PDF)
Heathkit GC-1092D Time-Calendar Clock Schematic (PDF)

[View Detail]

Mostek 3870 Series
 Mostek 3870 Series

Devices included in this entry:

Mostek 3870 microcontroller (40-pin plastic DIP)
Mostek MK38P70/02H microcontroller (40-pin piggyback cerDIP)
Mostek MK38P70/02J microcontroller with MK2716T UV-EPROM (40-pin piggyback cerDIP; pictured in thumbnail)
Fairchild 3871EPC PIA (40-pin plastic DIP)
Mostek 3875 microcontroller (40-pin piggyback plastic DIP)


The Mostek 3870 is an advanced single-chip implementation of the Fairchild F8 multi-chip microcontroller architecture. Fairchild and Mostek were simultaneously developing single-chip solutions for F8, but Mostek's 3870 microcontroller was brought to market first, and was superior to Fairchild's less ambitious '3859' design. Fairchild subsequently canceled their own project and assumed second-source production of the 3870 instead.

[View Detail]

Motorola 68000
 Motorola 68000

Devices included in this entry:

Motorola MC68000L8 microprocessor (64-pin cerDIP; pictured in thumbnail)
Hitachi HD68HC000P10 microprocessor (64-pin plastic DIP)
Motorola MC68450L10 DMA controller (64-pin cerDIP)


The Motorola 68000 is a powerful 32-bit CISC microprocessor with a 16-bit external data bus. Development of the 68000 began in 1976, as the Motorola Advanced Computer System on Silicon (MACSS) project, with the goal of developing a powerful new architecture which was not backwards-compatible with Motorola's 6800 8-bit microprocessor. This was a risky business decision at the time, as Intel's methodology was to maintain various levels of backwards compatibility through their entire flagship microprocessor lineage.

Despite a lack of compatibility with the 6800, the 68000 was a popular device, widely implemented in Unix-based workstations and other professional desktop systems, such as the Hewlett-Packard 9000 series computers. The 68000 eventually penetrated the consumer market, and was selected as the CPU for the Apple Lisa and Macintosh computers. By the early 1990s, Texas Instruments was using the 68000 as the CPU for their TI-89 and TI-92 graphing calculators.

[View Detail]

RCA 1248849-001
 RCA 1248849-001

The RCA 1248849-001 is an unidentified IC with a handwritten part number, housed in a 24-pin military flat pack. This device was undoubtedly used in an aerospace or defense system, but its function is unknown.

[View Detail]

RCA CD4000 Series
 RCA CD4000 Series

Devices included in this entry:

RCA CD4012AK dual four-input NAND gate (14-pin planar DIP; pictured in thumbnail)
RCA CD4019AK multiplexer (16-pin planar DIP)


The RCA CD4000 series is generally considered to be the first complete family of CMOS logic circuits, and includes functional equivalents of many TTL devices. Despite the much lower power requirements of CMOS, SSI TTL vastly outperforms SSI CMOS in terms of speed, and the CD4000 series did not offer large-scale integration, the primary advantage of a matured MOS technology.

[View Detail]

Smith-Corona Marchant I Calculator LSI
 Smith-Corona Marchant I Calculator LSI

Devices included in this entry:

TI TMC1771SC calculator LSI (40-pin ceramic LCC; pictured in thumbnail)
TI TMC1772SC calculator LSI (40-pin ceramic LCC; pictured in thumbnail)
TI TMC1773SC calculator LSI (40-pin ceramic LCC; pictured in thumbnail)
AMI 479A calculator LSI (40-pin ceramic LCC)
AMI 480B calculator LSI (40-pin ceramic LCC)


The Texas Instruments TMC177x series, used in the Smith-Corona Marchant I portable calculator, is the fourth calculator chipset produced by TI. The TMC177x series slightly predates the historically important TMS0100 series calculator chips, and is probably similar in design to the handful of other TMC17xx and TMC18xx LSI chipsets which were developed for Canon and other manufacturers in 1970 and 1971. Chips of this type were also used in the SCM Cogito 410 desktop calculator, but were replaced with ceramic DIPs in the Cogito 412 and 414.

The primitive leadless carrier is the most striking feature of these ICs. Note the asymmetrical design, built to interface with preexisting edge connector components. More modern leadless designs feature pins exiting the carrier on two or four sides, and require a custom leadless socket.

The function of the AMI ICs is unknown, but they share the same custom SCM carrier.

[View Detail]

Standard Microsystems CRT-7004-001
 Standard Microsystems CRT-7004-001

The Standard Microsystems CRT-7004-001 is an ASCII character generator, capable of generating 77-bit 7x11 bitmap characters or segment characters. Unlike intelligent alphanumeric LED displays of the era, which typically generate only 64 characters, the CRT-7004-001 can generate the full 128-character 7-bit ASCII character set.

[View Detail]

Telesensory Systems S14001A
 Telesensory Systems S14001A

Although Texas Instruments is most commonly associated with early speech synthesizers, The Telesensory Systems S14001A was introduced in late 1975, making it the first of such devices. TSI was developing the Speech+, a talking calculator for visually impaired users, and collaborated with Silicon Systems, Inc. to manufacture a speech synthesizer based on the original speech synthesis algorithm developed by Dr. Forrest S. Mozer. The resulting device was the S14001A, a 1500-transistor PMOS device with a 4-bit D/A converter, which operates at 13KHz.

[View Detail]

Texas Instruments TIB S0004
 Texas Instruments TIB S0004

This device is an unspecified variant of the original Texas Instruments TIB0203 92,304-bit bubble memory. Like Intel's devices, TI bubble memory uses a major-minor loop architecture.

[View Detail]

Texas Instruments SN18G141
 Texas Instruments SN18G141

The SN18G141 appears to be a Texas Instruments Series SNG logic device. Series SNG is a relatively early TTL logic family, packaged in 14-pin gold-plated military flat-pack carriers. Known part numbering schemes are formatted as SNGxxx in official TI documentation, and Gxxx on the devices themselves. The SN18G141 pictured here is either a SNG variant or a print variation of the SNG141, a quad two-input NAND gate. The date code of this device is eight months earlier than the earliest known SNG devices from other collections.

Series SNF/SNG Circuits Summary (PDF)

[View Detail]

Toshiba TMM121-1
 Toshiba TMM121-1

The Toshiba TMM121-1 is an unusual Japan-manufactured EPROM of unknown specification. If you can provide any information about this device, please contact us.

[View Detail]

TRW 1003J2C
 TRW 1003J2C

The 1003J2C is a radiation-hardened 12x12-bit multiplier and accumulator, packaged in a 64-pin ceramic DIP with a large anodized aluminum heatsink on top and gold cavity lid on the underside.

[View Detail]

Western Electric SAGE Hybrid Logic
 Western Electric SAGE Hybrid Logic

Devices included in this entry:

Western Electric GF40139 / 42 / 43 / 45 / 48 / 65 (custom metal can carrier)
Western Electric 11294629 (engineering sample; pictured in thumbnail)
Western Electric 11294641 / 44 (custom ceramic sandwich carrier)
Motorola 11294641 / 50 / 51 (custom ceramic sandwich carrier)


Western Electric produced two generations of pin-compatible discrete logic ICs which were used in the Semi-Automatic Ground Environment (SAGE) defense computer from the mid-1960s through the early 1970s. The first generation of devices, GF401xx, are hybrid devices composed of multiple discrete semiconductor dies mounted on a white ceramic substrate and encased in a gold plated metal can package. Later 112946xx family devices feature a single monolithic planar die, and are packaged in a ceramic sandwich carrier.

GF401xx family devices have an unusual construction which resembles the metal can packages of Western Electric power transistors from the same era, but instead of using the long thin leads seen on most Western Electric transistors, GF401xx has 11 heavy square wire wrap pins. This basic configuration is replicated in the Western Electric 112946xx series of integrated circuits, monolithic components which are direct replacements for the GF401xx series. Unlike GF401xx, 112946xx devices were second-sourced by Motorola. These package types and pin configuration are exclusive to GF401xx and 112946xx series devices, and were used neither commercially nor in any other Western Electric components.

[View Detail]

Western Electric Integrated Circuits
 Western Electric Integrated Circuits

Devices included in this entry:

Western Electric 1C parallel-to-serial interface (16-pin ceramic flat pack; pictured in thumbnail)
Western Electric 1L (16-pin ceramic flat pack)
Western Electric 41ER (16-pin ceramic and plastic DIP)
Western Electric 618 (16-pin ceramic and plastic DIP)


Like most Western Electric components, these integrated circuits are unusually constructed and difficult to identify. The 1C is believed to be a parallel-to-serial data converter designed for use in teletype machines. The other devices listed here remain unidentified.

[View Detail]

Western Electric 557D
 Western Electric 557D

The 557D is an unusual hybrid controller IC from a Western Electric Trimline telephone handset. The 557D is constructed from a multiple layers of ceramic, with discrete semiconductors and passives bonded together within the substrate. The 557D has numerous serpentine tracks of resistive material etched onto the ceramic. Earlier versions of this device were built upon a glass substrate.

[View Detail]

Zilog Z8603RS
 Zilog Z8603RS

The Z8603RS is an 8-bit microcontroller with a 24-pin piggyback ROM socket. Unlike the MK38P70, which has a metal socket inserted directly into the carrier, the Z8603RS has what appears to be a conventional plastic board-mount socket grafted to the top. Note the unusual offset cavity.

[View Detail]

©2000-2017 The Vintage Technology Association. All rights reserved.