Integrated Circuits

An integrated circuit (IC), or chip, is a complete circuit of transistors, resistors and other parts built together on one small piece of silicon and sealed in a package with metal pins. This article explains what is inside a chip and how it is made, the common package shapes and how their pins are numbered, how to read the markings and the datasheet, and two classic first chips: the 555 timer and the operational amplifier.

1. A short recap

• A semiconductor such as silicon conducts only a little on its own; adding tiny amounts of other elements, called doping, makes n-type or p-type regions (recap).
• A transistor lets a small current or voltage control a larger one. The MOSFET is controlled by the voltage on its gate (recap), and transistors can be wired as logic gates (recap).
• A resistor divider makes a fixed fraction of a voltage (recap).
• A capacitor charging through a resistor takes a predictable time, set by R × C (recap). A small decoupling capacitor beside each chip supplies fast bursts of current (recap).
• A voltage regulator such as the 7805 is itself an integrated circuit (recap).

2. What an integrated circuit is

Before ICs, every transistor, resistor and capacitor was a separate discrete component, wired to the others by hand. An IC builds them all at once, side by side in a single slice of silicon, with their connections made from thin metal layers on top. That finished piece of patterned silicon is called the die.^[1]

A die is far too small and fragile to solder to, so it is sealed in a package, usually black plastic, and fine wires join pads on the die to the package's metal pins or pads.^[1] The black "chip" you see on a circuit board is the package; the circuit itself is the die inside (Figure 1).

Building parts together makes circuits tiny, cheap and reliable: the connections are made in the factory, not by hand, and parts made side by side match each other closely. One chip can hold anything from four logic gates to a whole computer processor.^[1]

A short history

On 12 September 1958, Jack Kilby of Texas Instruments demonstrated a working oscillator whose transistor, capacitor and resistors were all formed in germanium, joined by fine gold wires.^[2] In 1959 Robert Noyce of Fairchild Semiconductor patented the more practical approach: make the parts in silicon, cover them with a protective layer of silicon oxide, and connect them with metal lines laid on top of the oxide. Fairchild produced the first working ICs of this kind in 1960, and that planar method is still the basis of chipmaking today.^[3]​^[4]

In 1965 Gordon Moore noticed that the number of components per chip was doubling every year; in 1975 he revised the rate to a doubling every two years. The prediction became known as Moore's law.^[5] Intel's 4004, one of the first microprocessors (a computer's processing unit on one chip), held 2,300 transistors in 1971.^[6] Apple's A16 phone chip holds 16 billion.^[7]

3. How a chip is made

Chips are made on a wafer, a thin disk sliced from a large crystal of very pure silicon and polished extremely smooth; the most common size is 300 mm across. Each wafer carries many copies of the same chip, from a few dozen to thousands, built up together in layers.^[7] ASML, which makes the machines that print the patterns, describes the key steps like this:^[7]

1. Deposition. A thin film of conducting, insulating or semiconducting material is laid over the whole wafer.
2. Photoresist. The wafer is coated with photoresist, a light-sensitive coating.
3. Lithography. Ultraviolet light shines through a reticle, a master plate carrying the pattern for one layer, and lenses or mirrors shrink that pattern onto the resist. Where light lands, the resist changes chemically.
4. Etch. The changed resist is washed away, and the exposed material underneath is removed by chemicals or gases, leaving the pattern.
5. Ion implantation. Charged atoms are fired into the exposed silicon to dope it, which is how the n-type and p-type regions of transistors are made.
6. Repeat, then package. The cycle repeats for each layer. The finished wafer is tested, sawn into individual dies with a diamond saw, and each good die is packaged.

The whole process takes thousands of steps and more than three months.^[7] Figure 2 shows the result, much simplified: transistors at the bottom, in the silicon itself, and layers of metal wiring above them, separated by insulating glass-like layers. Advanced chips have a single layer of transistors wired together by more than a dozen layers of metal.^[8]

You will often see chips described by a process node, such as "7 nm" or "5 nm". These names once matched the size of a transistor feature, but they stopped doing so in the 1990s; today they are labels for a manufacturing generation, not a measurement.^[8] (A nanometer, nm, is a billionth of a meter.)

4. Packages

Packages come in two mounting styles. Through-hole packages have legs that pass through holes in the circuit board and are soldered on the other side; they are larger and easy to handle. Surface-mount packages (SMD, for surface-mount device) sit on copper pads on one side of the board and are soldered there.^[1] The common shapes are (Figure 3):

• DIP (dual in-line package): two parallel rows of legs, spaced 0.1 inch (2.54 mm) apart. It is the classic through-hole chip, and it straddles the center gap of a breadboard so each pin gets its own row.^[1]
• SOIC (small-outline IC): the surface-mount cousin of the DIP, with legs bent outward at about 1.27 mm spacing. With a steady hand it can be hand-soldered. SSOP and TSSOP are smaller, thinner versions.^[1]
• QFP (quad flat package): legs on all four sides, often 0.4 to 1 mm apart, for chips with many connections.^[1]
• QFN (quad flat no-lead): like a QFP with the legs removed; small pads sit on the underside edges. Many QFNs have a large exposed pad in the middle of the underside, soldered to the board to carry heat away and often to connect ground.^[1]​^[9] The RP2040 microcontroller, for example, is a 56-pin QFN.^[10]
• BGA (ball grid array): a grid of solder balls covering the underside, used for processors and other chips with hundreds of connections; Amkor's plastic BGAs go up to 1,521 balls.^[11] BGAs are normally soldered by machine in a reflow oven.^[1]

Many simple chips come in several packages. TI sells the same 74HC00 logic chip as a 14-pin DIP (19.3 mm long) and as a 14-pin SOIC (8.7 mm) or TSSOP (5.0 mm), among others.^[12] Some ICs use packages you have met before: the 7805 regulator comes in the three-legged TO-220 (recap), and tiny chips use transistor-style packages such as SOT-23 (recap).

5. Pin numbering

Every pin of an IC has its own job, so you must know which is which. Packages mark pin 1 with a notch at one end, a small dot or dimple beside pin 1, or both. Looking down on the top of the chip, the pin numbers then count up counterclockwise around the package.^[1] On an 8-pin DIP with the notch at the left, pin 1 is bottom left, pins 1 to 4 run left to right along the bottom, and pins 5 to 8 come back right to left along the top (Figure 4). The same counterclockwise rule applies to SOIC, QFP and QFN packages.

Datasheet pinouts are drawn as a "top view". Seen from the underside of a board, the pattern is mirrored, a common source of mistakes.

BGA balls use a grid instead. Rows are lettered and columns numbered, so each ball has a name such as A1 or C7, and the A1 corner is marked on the package. Letters that are easy to confuse with numbers, such as I, O, Q, S, X and Z, are skipped.^[13]

6. Units and markings

The writing on a chip is mostly a part number. Take TI's SN74HC00N (Figure 5):^[12]

• SN is TI's prefix.
• 74 marks the standard version, rated −40 to 85 °C; the wider-range 54 version is rated −55 to 125 °C.
• HC is the logic family, a group of chips sharing one circuit design, voltage range and speed. HC is a CMOS family (often read as "high-speed CMOS") that runs from 2 to 6 V.
• 00 is the function: four 2-input NAND gates.
• N is the package: a 14-pin DIP. The same chip in SOIC is marked only "HC00", because the package is too small for the full number.

Chips may also carry a maker's logo and a lot or date code that traces the production batch.^[12] The "74" numbering goes back to TI's 7400 series of logic chips, launched in plastic packages in 1966, which soon won over half of the logic market.^[14] The numbering is still used today.

ICs have no unit of their own, the way resistors have ohms. Datasheets use the familiar units: volts for supplies, milliamps (mA) or microamps (µA) for currents, nanoseconds (ns) for timing, and degrees Celsius. Packages are named by type and pin count, such as "SOIC-8".

7. Common types: analog, digital and mixed-signal

Analog ICs handle signals that can take any value within a range, such as the voltage from a microphone or a temperature sensor. Amplifiers, comparators, timers and voltage regulators are analog chips.^[1]

Digital ICs handle signals that are treated as just two states, high (1) and low (0). Logic gates, counters, memory, microcontrollers and processors are digital.^[1] Most digital chips today use CMOS (complementary metal-oxide-semiconductor) circuits, built from pairs of MOSFETs: a p-channel transistor that connects the output to the supply and an n-channel one that connects it to ground. Only one is on at a time, so in a steady state almost no current flows.^[15]

"High" and "low" are ranges, not exact voltages. With a 4.5 V supply, a 74HC00 input counts anything below 1.35 V as low and anything above 3.15 V as high; in between, the result is not guaranteed (Figure 6).^[12] Each gate takes typically 9 ns to respond at that supply.^[12]

Mixed-signal ICs contain both. The RP2040 microcontroller on the Raspberry Pi Pico is mainly digital, with two processor cores and 264 kB of memory, but it also has an analog-to-digital converter (ADC), which measures a voltage on four of its pins and turns it into a 12-bit number.^[10] Power ICs, such as regulators and motor drivers, mainly handle current rather than information.

8. How to read a datasheet

A datasheet is the maker's official, free document for a part: what it does, its limits, how to wire it and its packages. Read it before using any chip. Most follow a similar order (Figure 7); here is TI's NE555 datasheet:^[16]

• Page 1: features, applications and description. A one-page summary. The NE555's says it can time "from microseconds to hours" and its output can sink or source up to 200 mA.
• Pin configuration and functions. Top-view drawings of each package and a table of what each pin does.
• Absolute maximum ratings. Limits that must never be exceeded, even for a moment, or the chip may be permanently damaged. For the NE555 the supply must never exceed 18 V. These are not working conditions.
• ESD ratings. How much static discharge the pins survive in standard tests (section 14).
• Recommended operating conditions. The range in which the chip is designed to work: for the NE555, a supply of 4.5 to 16 V and output current up to 200 mA. Design within these.
• Electrical characteristics. Tables of measured performance under stated test conditions, given as min (minimum), typ (typical) and max (maximum). Design with the min or max values; "typical" is only what an average part does. With a 5 V supply the NE555 draws typically 2 to 3 mA, but up to 6 mA.
• Typical characteristics. Graphs of behavior against voltage, current or temperature.
• Detailed description and application information. How it works, formulas and example circuits.
• Power supply and layout recommendations, such as a 0.1 µF ceramic capacitor from the supply pin to ground.
• Mechanical, packaging and ordering information. Exact package dimensions and the full part numbers with their markings.

Also check which variant and package each number applies to: one datasheet often covers several versions with different numbers.^[16]​^[17]

9. First example: the 555 timer

The 555 timer was designed by Hans Camenzind for Signetics and launched in 1971. His original design had 23 transistors, 16 resistors and 2 diodes, and the chip has ended up in kitchen appliances, toys and spacecraft.^[18] It is still made by several companies and remains a favorite first IC.

Inside, the 555 has three main blocks (Figure 8):^[16]

• An internal divider that sets two reference levels: one-third and two-thirds of the supply voltage.
• Two comparators, circuits that compare two voltages and report which is higher. One watches the trigger pin (TRIG, pin 2) against the one-third level; the other watches the threshold pin (THRES, pin 6) against the two-thirds level.
• A flip-flop, a tiny memory circuit that holds one of two states. When TRIG falls below one-third of the supply, it sets the output (pin 3) high. When THRES rises above two-thirds, it resets the output low and switches on the discharge transistor (DISCH, pin 7), which connects pin 7 to ground.

Pin 4, RESET, forces the output low, so tie it to the supply when unused. Pin 5, CONT, gives access to the two-thirds level.^[16]

One-shot (monostable) mode

In monostable mode, a short low pulse on TRIG starts one timed output pulse. A capacitor C charges through a resistor R_A until it reaches two-thirds of the supply, which ends the pulse. The pulse length is about

t = 1.1 × R_A × C

so 100 kΩ and 10 µF give about 1.1 seconds (calculated). Because both the charging and the threshold scale with the supply, the time barely depends on the supply voltage.^[16]

Oscillator (astable) mode

In astable mode the 555 triggers itself and runs continuously, making a square wave, a signal that switches between high and low. TRIG and THRES are joined to the capacitor, which charges through R_A and R_B and discharges through R_B alone, swinging between one-third and two-thirds of the supply (Figure 9). TI gives:^[16]

t_high ≈ 0.693 × (R_A + R_B) × C
t_low ≈ 0.693 × R_B × C
f ≈ 1.44 ÷ ((R_A + 2R_B) × C)

Worked example: an LED blinker. With R_A = 10 kΩ, R_B = 68 kΩ and C = 10 µF (all calculated):

• t_high = 0.693 × 78 kΩ × 10 µF ≈ 0.54 s
• t_low = 0.693 × 68 kΩ × 10 µF ≈ 0.47 s
• f = 1.44 ÷ (146 kΩ × 10 µF) ≈ 0.99 Hz, about one flash a second.

The high time is always longer than the low time here, because the capacitor charges through both resistors. An LED with its series resistor on pin 3 completes the blinker (recap). The timing capacitor's tolerance, which can be ±20%, limits accuracy (recap); for precise timing a crystal is used instead (article 21, Crystals and Oscillators).

The original 555 is built from bipolar transistors (BJTs) and draws a few milliamps. CMOS versions such as TI's TLC555 share its pinout, run from 2 to 15 V and draw well under a milliamp, but their outputs supply much less current.^[19]

10. Second example: the operational amplifier

An operational amplifier, or op-amp, is an amplifier with two inputs and one output. It amplifies the difference between the non-inverting input (marked +) and the inverting input (marked −), and its open-loop gain, the gain with nothing connected around it, is enormous.^[20] TI's LM358B, a common dual op-amp (two in one 8-pin package), has a typical open-loop gain of 140 V/mV, meaning 140,000 times, and at least 70,000.^[17] The name comes from the "operations", such as adding and scaling, that early op-amps performed in analog computers.^[20]

As a comparator. With so much gain, a millivolt of difference drives the output fully high or low, so a bare op-amp tells you which input is higher, for example lighting an LED when a sensor voltage passes a set level.^[20] Dedicated comparator chips such as the LM393 are designed for this job; the LM393's output is open-collector, which means it can only pull down, so it needs a pull-up resistor to a supply.^[21]

With negative feedback. Connect part of the output back to the inverting input, and the op-amp settles where its two inputs are almost exactly equal. Two simple rules then describe the circuit: the inputs draw almost no current, and the output does whatever it must to keep the two inputs at the same voltage.^[20] Resistors, not the chip, then set the gain (Figure 10):

• Non-inverting amplifier. The signal goes to +, and a divider of R2 (output to −) and R1 (− to ground) feeds back a fraction of the output. Gain = 1 + R2 ÷ R1.^[20] With R1 = 10 kΩ and R2 = 100 kΩ, the gain is 11, so 0.2 V in gives 2.2 V out (calculated).
• Inverting amplifier. The signal goes through R_I to −, with R_F from output to − and + grounded. Gain = −R_F ÷ R_I; the minus sign means the output is upside down. TI's example uses 10 kΩ and 36 kΩ for a gain of −3.6.^[17]

Real limits. An LM358B's output stops about 1.4 V below its positive supply, so on 5 V it tops out near 3.6 V; the gain-of-11 amplifier above would clip (flatten) for inputs above about 0.33 V (calculated). Its inputs work down to the negative supply but only up to 2 V below the positive one.^[17] Rail-to-rail op-amps get closer to the supplies. Speed is limited too: the LM358B's gain falls to 1 at about 1.2 MHz.^[17]

The LM358's pins are: pin 1 output A, pins 2 and 3 the − and + inputs of A, pin 4 ground or negative supply, pins 5 and 6 the + and − inputs of B, pin 7 output B, and pin 8 positive supply.^[17] Many other dual op-amps share this pinout, but always check the datasheet.

11. How ICs behave in a circuit

• Every chip needs power. Schematics often leave out the supply pins, but they must be connected. Put a 0.1 µF ceramic decoupling capacitor between each chip's supply and ground pins, as close as possible.^[16]
• Never leave CMOS inputs floating. An unconnected input drifts into the undefined band, where both transistors of the input turn on together. The chip can draw extra current, oscillate or behave randomly. Tie unused inputs to the supply or ground, directly or through a resistor of about 10 kΩ.^[12]​^[15]
• Outputs have current limits. A logic output can light a small LED through a resistor, but motors, relays and lamps need a transistor or driver (recap). Do not connect outputs directly to the supply or ground.^[12]
• Voltages must match. 5 V signals can exceed a 3.3 V chip's absolute maximum ratings, and a 3.3 V "high" may be too low for a 5 V chip. Check both datasheets.
• Chips make heat. The package limits how much heat a chip can shed, just as for regulators (recap).

12. Testing

• Check orientation and power first. Most "dead chip" problems are a reversed chip, a missing supply or ground, or a wrong pin. Measure the voltage between the chip's own supply and ground pins with a multimeter (recap).
• Check inputs and outputs against the datasheet. For logic, measure whether each input is clearly high or low, and compare the output with the chip's truth table. For a 555, pin 3 should switch, and pin 4 should be high.^[16]
• Use an oscilloscope for anything changing. A meter averages a fast square wave; a scope shows its shape and frequency (recap).
• Swap in a known-good part. A socket under a DIP chip makes this easy.^[1] A chip that gets hot with no load is usually damaged or wired wrongly.
• Beware of fakes. Counterfeit and relabeled chips are sold, especially through unofficial channels; buy from authorized distributors when it matters.^[22]

13. Everyday uses

ASML counts about 1.15 trillion chips sold in 2021, about 145 for every person on Earth.^[7] Phones and computers are full of processors, memory and power chips; less obvious ones run washing machines, refrigerators, air conditioners and phone chargers; TI lists all of those among the uses of the humble LM358.^[17] Small microcontrollers such as the RP2040 run hobby projects and many gadgets (see Microcontrollers),^[10] and sensor chips measure temperature, motion and light (article 18, Sensors).

14. Safety and handling

Most IC circuits run at low, safe voltages; the main danger is to the chip.

Electrostatic discharge (ESD). When two materials touch and separate, such as shoes on a floor, electrons move between them and leave a static charge, which can discharge through a chip's pins as a tiny spark. The EOS/ESD Association lists typical charges of 35,000 V from walking across carpet in dry air (10 to 25% relative humidity), falling to 1,500 V in humid air (65 to 90%).^[23] Many components can be damaged below 100 V, and damage can be latent: the chip works for now but fails early.^[23] Datasheets give an ESD rating from a standard human-body model test: ±500 V for the NE555 and ±2,000 V for the LM358B.^[16]​^[17] TI's note on that rating says 500 V allows safe manufacturing with a standard ESD control process, so the same care makes sense at home.^[16]

Simple habits follow the ESD Association's principles of keeping everything at the same potential and letting charge drain away safely:^[24]

• Touch a grounded metal object before handling chips, or wear a grounded wrist strap, and work on an antistatic mat if you have one.
• Keep chips in their antistatic bags, tubes or foam until you use them, and keep ordinary plastic, foam cups and bubble wrap away from the work area.
• Handle chips by the body, not the pins, since a finger touching the leads is exactly the discharge path the human-body model describes,^[23] and take extra care in dry rooms.
• The same habits protect MOSFETs (recap).

Wrong way round or wrong voltage. A chip inserted backwards, or supplied above its absolute maximum rating, can be permanently damaged.^[16] A damaged or reversed chip can also get hot. Check pin 1 before applying power, and if a chip gets hot or smells, disconnect power before touching it. Add fuses or other protection in projects that draw real current (article 17, Fuses and Circuit Protection).

Soldering and mains. Use a grounded soldering iron and let the chip cool before touching it. Projects that take power from the wall carry dangerous voltages before the low-voltage supply (recap).

The whole story in brief

• An IC is a whole circuit built on one silicon die, sealed in a package with pins.
• Chips are made by repeating deposition, lithography, etching and doping on wafers, then sawn apart, packaged and tested.
• Packages range from through-hole DIPs to surface-mount SOIC, QFP, QFN and BGA. Pin 1 is marked by a notch or dot, and pins count counterclockwise from the top.
• Part numbers encode maker, temperature range, family, function and package.
• ICs are analog, digital or mixed-signal; CMOS inputs must never float.
• Read a datasheet's absolute maximum ratings and recommended operating conditions first, and design with min and max values.
• The 555 makes timed pulses and square waves set by resistors and a capacitor; the op-amp amplifies with a gain set by resistors.
• Guard chips against static, and check orientation and voltage before power-up.

Coming up in this section: 17 Fuses and Circuit Protection.

Sources

All links were checked in October 2026. External links open in a new tab.

1. SparkFun Electronics, "Integrated Circuits" tutorial (dies, packages, pin numbering, common ICs): https://learn.sparkfun.com/tutorials/integrated-circuits/all
2. Computer History Museum, The Silicon Engine, "1958: All Semiconductor 'Solid Circuit' is Demonstrated": https://www.computerhistory.org/siliconengine/all-semiconductor-solid-circuit-is-demonstrated/
3. Computer History Museum, The Silicon Engine, "1959: Practical Monolithic Integrated Circuit Concept Patented": https://www.computerhistory.org/siliconengine/practical-monolithic-integrated-circuit-concept-patented/
4. Computer History Museum, The Silicon Engine, "1960: First Planar Integrated Circuit is Fabricated": https://www.computerhistory.org/siliconengine/first-planar-integrated-circuit-is-fabricated/
5. Computer History Museum, The Silicon Engine, "1965: 'Moore's Law' Predicts the Future of Integrated Circuits": https://www.computerhistory.org/siliconengine/moores-law-predicts-the-future-of-integrated-circuits/
6. Computer History Museum, The Silicon Engine, "1971: Microprocessor Integrates CPU Function onto a Single Chip": https://www.computerhistory.org/siliconengine/microprocessor-integrates-cpu-function-onto-a-single-chip/
7. ASML, "6 crucial steps in semiconductor manufacturing" (updated October 2023): https://www.asml.com/en/company/stories/2021/semiconductor-manufacturing-process-steps
8. Samuel K. Moore, IEEE Spectrum, "A Better Way to Measure Progress in Semiconductors" (July 2020): https://spectrum.ieee.org/a-better-way-to-measure-progress-in-semiconductors
9. Amkor Technology, "Micro LeadFrame (QFN)" package overview: https://amkor.com/packaging/leadframe/microleadframe/
10. Raspberry Pi Ltd, RP2040 Datasheet (PDF): https://datasheets.raspberrypi.com/rp2040/rp2040-datasheet.pdf
11. Amkor Technology, "PBGA/TEPBGA (Plastic Ball Grid Array)" package overview: https://amkor.com/packaging/laminate/pbga/
12. Texas Instruments, SN74HC00, SN54HC00 Quadruple 2-Input NAND Gates data sheet (SCLS181H, PDF): https://www.ti.com/lit/ds/symlink/sn74hc00.pdf
13. JEITA EDR-7316C, "Design guide for semiconductor packages: Fine-pitch Ball Grid Array and Fine-pitch Land Grid Array (FBGA/FLGA)" (PDF): https://home.jeita.or.jp/tsc/std-pdf/EDR-7316C_E.pdf
14. Computer History Museum, The Silicon Engine, "1963: Standard Logic IC Families Introduced": https://www.computerhistory.org/siliconengine/standard-logic-ic-families-introduced/
15. Texas Instruments, "Implications of Slow or Floating CMOS Inputs" (SCBA004E, PDF): https://www.ti.com/lit/an/scba004e/scba004e.pdf
16. Texas Instruments, NA555, NE555, SA555, SE555 xx555 Precision Timers data sheet (SLFS022K, revised March 2026, PDF): https://www.ti.com/lit/ds/symlink/ne555.pdf
17. Texas Instruments, LMx58, LM2904 Industry-Standard Dual Operational Amplifiers data sheet (SLOS068AB, PDF): https://www.ti.com/lit/ds/symlink/lm358.pdf
18. IEEE Spectrum, "Chip Hall of Fame: Signetics NE555": https://spectrum.ieee.org/chip-hall-of-fame-signetics-ne555
19. Texas Instruments, TLC555 CMOS Timer data sheet (SLFS043K, PDF): https://www.ti.com/lit/ds/symlink/tlc555.pdf
20. Tony R. Kuphaldt, Lessons In Electric Circuits, Volume III (Semiconductors), chapter 8, "Operational Amplifiers": https://www.ibiblio.org/kuphaldt/electricCircuits/Semi/SEMI_8.html
21. Texas Instruments, LM393B, LM2903B, LM193, LM293, LM393 and LM2903 Dual Comparators data sheet (SLCS005AH, PDF): https://www.ti.com/lit/ds/symlink/lm393.pdf
22. Texas Instruments, "Anti-counterfeit": https://www.ti.com/quality-reliability/quality/anti-counterfeit.html
23. EOS/ESD Association, "ESD Fundamentals Part 1: An Introduction to ESD": https://www.esda.org/esd-overview/esd-fundamentals/part-1-an-introduction-to-esd/
24. EOS/ESD Association, "ESD Fundamentals Part 2: Principles of ESD Control": https://www.esda.org/esd-overview/esd-fundamentals/part-2-principles-of-esd-control/