tech

The 8008 Microprocessor That Sparked the Microprocessor Revolution.

Intel's pioneering 8008 microprocessor was first produced over 50 years ago. This was Intel's first 8-bit microprocessor and the ancestor of the x86 processor family that readers may be using now. (The author could not find a high-quality chip photo of the 8008, so they personally opened up a chip and took some detailed photos.) These new chip photos are in this article, accompanied by a discussion on the internal design of the 8008.

The following photo shows the small silicon chip inside the 8008 package (click to get a higher resolution photo). Readers can hardly see the wires and transistors that make up the chip. The surrounding squares are 18 pads, which are connected to the external pins through tiny metal wires.

The text "8008" can be seen on the right edge of the chip, and the text "© Intel 1971" can be seen on the bottom edge. The abbreviation "HF" appears in the upper right corner, representing Hal Feeney, who was the logic designer and physical layout person of the chip. Other key designers of the 8008 include Ted Hoff, Stan Mazor, and Federico Faggin.

Advertisement

Inside the Chip

The following diagram highlights some of the main functional modules of the chip. On the left is the 8-bit Arithmetic/Logic Unit (ALU), which performs actual data calculations.The ALU uses two temporary registers to save its input values. These registers occupy a significant area on the chip, not because they are complex, but because they require large transistors to drive the signals through the ALU circuit.

Below the registers is the carry-lookahead circuit. For addition and subtraction, this circuit computes all eight carry values in parallel to enhance performance. Since the low-order carry depends only on the low-order bits, while the high-order carry depends on multiple bits, the circuit block is triangular in shape.

The triangular layout of the ALU is unusual. Most processors stack the circuitry for each bit into a regular rectangular pattern (bit-slice layout). However, the 8008 has eight blocks (one for each bit), which are haphazardly arranged to fit the space left by the triangular carry generator, and the ALU supports eight simple operations.

At the center of the chip is the instruction register and the instruction decode logic, which determines the meaning of each 8-bit machine instruction. Decoding is done through a programmable logic array (PLA), which is an arrangement of gates used to match bit patterns and generate the appropriate control signals for the rest of the chip. On the right side is the memory block. The 8008's seven registers are located in the upper right corner, and in the lower right corner is the address stack, which consists of eight 14-bit address words. Unlike most processors, the 8008's call stack is stored on the chip, not in memory. The program counter is just one of these addresses, making subroutine calls and returns very simple, and the 8008 uses dynamic memory for this storage.

The physical structure of the chip closely resembles the block diagram in the 8008 user manual (below), with the various blocks on the chip almost in the same positions as in the diagram.

What does the photo of the chip show? It can be seen as three layers. The chart below shows a close-up of the chip, pointing out these layers. The top layer is the metal wires, which are the most prominent feature. In the details below, these lines are mostly horizontal, with the polysilicon layer beneath the metal, appearing orange under a microscope.The foundation of a chip is the silicon wafer, which appears as a purplish-gray in the photo. Pure silicon is essentially an insulator, and its regions are doped with impurities to create semiconductor silicon. Due to its position at the bottom, the silicon layer is hard to distinguish, but a black line can be seen at the boundary between doped and undoped silicon. Several vertical silicon "lines" can be seen in the photo.

Transistors are the key components of a chip, and a transistor is formed when a polysilicon line passes through doped silicon. In the photo, the polysilicon appears as a brighter orange color when forming a transistor.

Why is it an 18-pin chip?

One inconvenience of the 8008 is that it only has 18 pins, which makes the chip slower and more difficult to use. The 8008 uses 14 address bits and 8 data bits, so 18 pins are not enough for each signal. Instead, the chip has 8 data pins, which are reused over three cycles to transmit the low address bits, high address bits, and data bits. Computers using the 8008 require many support chips to interact with this inconvenient bus architecture.

There is no sufficient reason for designing the chip with only 18 pins. Other manufacturers commonly use 40-pin or 48-pin packages, but 16 pins were "held as a sacred rule" at Intel, and they reluctantly increased to 18 pins. A few years later, when the 8080 processor came out, Intel had accepted 40-pin chips. The 8080 processor became more popular, partly because it adopted a simpler bus design allowed by the 40-pin package.

Power and data paths in the chip

The data bus provides a data stream for the chip. The chart below shows the 8-bit data bus of the 8008, represented in rainbow colors with 8 data lines. The data bus connects to the 8 data pins on the outer upper half of the chip, running between the Arithmetic Logic Unit (ALU) on the left, the Instruction Register (IR) in the middle, and the registers and stack on the right. It is divided into two parts on the left, extending along both sides of the ALU.The red and blue lines illustrate the power routing. Power routing is an underestimated aspect in microprocessors, as power is routed through metal layers due to the low resistance of the metal. However, since early microprocessors only had one layer of metal, power distribution had to be carefully planned to ensure that paths do not cross. The chart above shows the blue Vcc line and the red Vdd line. Power is supplied through the Vcc pin on the left and the Vdd pin on the right, then branches into fine, interwoven lines to power all parts of the chip.

Register File

To provide a detailed view of the chip's appearance, the author has magnified the register file of the 8008 in the photo below. The register file consists of a dynamic random access memory (DRAM) cell grid of 8x7, with each storage cell using three transistors to save a bit. (Transistors can be seen as small rectangles, where the orange polycrystalline silicon appears slightly brighter in color.) Each row is one of the seven 8-bit registers of the 8008 (A, B, C, D, E, H, L). On the left, you can see seven pairs of horizontal lines: the read select lines and write select lines for each register. At the top, you can see eight vertical lines to read or write the content of each bit, as well as five thicker lines to provide Vcc. The use of DRAM as a register (instead of the more common static latch) is an interesting choice. It is speculated that Intel chose DRAM because they were primarily a memory company at the time.

How PMOS Works

The 8008 uses PMOS transistors. Simply put, a PMOS transistor can be seen as a switch between two silicon lines, controlled by the gate input of polycrystalline silicon. When the gate input is low, the switch closes, and it can pull its output high. If you are familiar with NMOS transistors used in microprocessors like the 6502, PMOS may be a bit confusing because everything is reversed.

A simple PMOS NAND gate can be constructed as follows. When both inputs are high, the transistor is off, and the resistor pulls the output low. When either input is low, the transistor will conduct, connecting the output to +5V. Thus, the circuit implements a NAND gate. To be compatible with 5-volt TTL circuits, PMOS gates (and therefore the 8008) use an uncommon voltage supply: -9V and +5V.Due to technical reasons, resistance is actually implemented through transistors. The following diagram shows how the transistors are connected to act as pull-down resistors, with the details on the right showing what the circuit looks like on the chip, with the -9V metal wire at the top, the transistors in the middle, and the output being the silicon wire at the bottom.

The History of the 8008

The intricate story of the 8008 began with the Datapoint 2200, a popular computer introduced in 1970 as a programmable terminal. (Some consider the Datapoint 2200 to be the first personal computer.) The Datapoint 2200 did not use a microprocessor but instead used a board-level CPU constructed from individual TTL chips. (This was the standard way to build CPUs in the era of minicomputers.) Datapoint and Intel decided that this circuit board could be replaced with an MOS chip, and thus Intel began the 8008 project to manufacture this chip. Soon after, Texas Instruments also agreed to manufacture a single-chip processor for Datapoint. Both chips were designed to be compatible with the 8-bit instruction set and architecture of the Datapoint 2200.

Around March 1971, Texas Instruments completed their processor chip, known as the TMC 1795. After delaying the project, Intel completed the 8008 chip around the end of 1971. For various reasons, Datapoint rejected both microprocessors and built a faster CPU based on updated TTL chips, including the 74181 ALU chip.

Texas Instruments tried to market the TMC 1795 processor to companies such as Ford, but without success, and eventually abandoned this processor to focus on the lucrative calculator chip market. On the other hand, Intel marketed the 8008 as a general-purpose microprocessor, which ultimately led to the birth of the x86 architecture. Although Texas Instruments was the first to launch an 8-bit processor, Intel successfully brought their chip to market, pioneering the microprocessor industry.The chart outlines the "family tree" of the 8008 and its related processors. The architecture of the Datapoint 2200 was utilized in the TMC 1795, Intel 8008, and the next generation Datapoint 220011. Consequently, four completely different processors all employed the instruction set and architecture of the Datapoint 2200. The Intel 8080 processor is an improved version of the 8008, significantly expanding the instruction set of the 8008 and rearranging machine code instructions for greater efficiency. The 8080 was used in pioneering early microcomputers, such as the Altair and Imsai. After the development of the 4004 and 8080, designers Federico Faggin and Masatoshi Shima left Intel to construct the Zilog Z-80 microprocessor, which improved upon the 8080 and became very popular.

The leap to the 16-bit 8086 processor was not as gradual. Although most 8080 assembly code could be converted to run on the 8086, it was not a simple task, as both the instruction set and architecture underwent fundamental changes. Nevertheless, some features of the Datapoint 2200 still exist in today's x86 processors. For instance, the Datapoint 2200 had a serial processor that processed one bit of a byte at a time. Since the least significant bit had to be processed first, the Datapoint 2200 was little-endian. For compatibility, the 8008 was also little-endian, and Intel's processors have remained so to this day. Another feature of the Datapoint 2200 was the parity flag, as parity calculations were important for terminal communications. The parity flag has continued through to the x86 architecture.

The 8008 is architecturally unrelated to Intel's 4-bit 4004 processor. The 8008 is definitely not an 8-bit version of the 4-bit 4004. The similar names are purely a marketing invention; during the design phase, the 8008 had a less conspicuous name, "1201."

The Position of 8008 in the History of Semiconductor Technology

Both the 4004 and 8008 utilized silicon-gate enhanced PMOS (Positive-channel Metal Oxide Semiconductor), a semiconductor technology that was only briefly used. This places these two chips in an interesting position in chip manufacturing technology.

The 8008 (as well as modern processors) uses MOS transistors. These transistors had a long road to acceptance, as they were slower and less reliable than the bipolar transistors used in most computers in the 1960s. By the end of the 1960s, MOS integrated circuits began to become more common; the standard technology of the time was PMOS transistors with metal gates. The transistor's gate was made of metal, which was also used to connect components on the chip. The chip essentially had two layers of functionality: the silicon itself and the metal wiring on top. This technology was used in many Texas Instruments calculator chips as well as the TMC 1795 chip (a chip with the same instruction set as the 8008).

The key innovation that made the 8008 practical was the self-aligned gate—a transistor that used polysilicon gates instead of metal gates. Although this technology was invented by Fairchild and Bell Labs, Intel pushed its development. Polysilicon gate transistors had better performance than metal gate transistors (for complex semiconductor reasons). Additionally, adding a polysilicon layer made signal routing in the chip easier, allowing for denser chips. The chart below shows the benefits of the self-aligned gate: the metal-gate TMC 1795 was larger than the 4004 and 8008 chips combined.

Soon after, semiconductor technology advanced again, replacing PMOS transistors with NMOS (Negative-channel Metal Oxide Semiconductor) transistors. Although PMOS transistors were initially easier to manufacture, NMOS transistors were faster, so once NMOS could be reliably manufactured, they clearly took the lead.NMOS led to the emergence of more powerful chips, such as Intel's 8080 and Motorola's 6800 (both in 1974). Another technological improvement during this period was ion implantation, which was used to alter the characteristics of transistors. This enabled the creation of "depletion-type" transistors that could be used as pull-up resistors. These transistors enhanced chip performance and reduced power consumption. They also allowed for the creation of chips that could operate using a standard five-volt power supply.

The combination of NMOS transistors and depletion-type pull-up resistors was used in most microprocessors of the late 1970s and early 1980s, such as the 6502 (1975), Z-80 (1976), 68000 (1979), and Intel chips ranging from the 8085 (1976) to the 80286 (1982).

By the mid-1980s, CMOS (Complementary Metal Oxide Semiconductor) began to dominate, combining NMOS and PMOS transistors to significantly reduce power consumption, such as the 80386 (1986), 68020 (1984), and ARM1 (1985). Nowadays, almost all chips are CMOS.

As the reader can see, the 1970s was a period of tremendous change in semiconductor chip technology. The creation of the 4004 and 8008 was a result of the intersection of technological capabilities and market demand at the time.

How to Take a Chip Die Photo

The first step is to open the chip package to expose the die. Most chips are housed in epoxy packages, which can be dissolved with dangerous acids.

To avoid boiling nitric acid, a simpler method was adopted in this article. The 8008 also has a ceramic package version (as shown above), and by gently tapping the chip along the seam with a chisel, the two layers of ceramic can be separated. The following photo shows the lower half of the ceramic package, with the die already exposed. Most of the metal pin feet have been removed, but their positions in the package are still visible. On the right side of the die is a small square that connects the ground (Vcc) to the substrate. A few tiny connecting wires can still be seen, which are connected to the die.Once the die is exposed, it can be photographed using a microscope. Standard microscopes illuminate from below, which is not very effective for die photos. Instead, the author used a metallurgical microscope, which illuminates the chip from above.

To take the photo, the author first took 48 photos with the microscope, then used the Hugin stitching software to combine them into a high-resolution image (details). Finally, the contrast of the image was adjusted to make the structure of the chip more clearly visible. The original image below (approximately what is seen through the microscope) is used for comparison.

---

Conclusion

Although the 8008 was not the first microprocessor, nor even the first 8-bit microprocessor, it was indeed revolutionary, sparking the microprocessor revolution and leading to the emergence of the x86 architecture, which dominated personal computers for decades to come.