L

L1
(1)
Level 1. See cache.

(2)
A standard CEPT data rate.

(3)
A GPS signal, at 1575.42 MHz.
L2
(1)
Level 2. See cache.

(2)
A GPS signal, at 1227.60 MHz.
L2TP
Layer 2 Tunneling Protocol. A member of the TCP/IP protocol suite at the Link layer, used to create VPNs for privacy. The original version was based largely on Microsoft’s PPTP, and connected nodes via PPP. The 2005 L2TPv3 can host multiple connections, and supports other connection protocols besides PPP.
L3
Level 3. See cache.
l337 (or l337 h4x0r)
Deliberate mangling of “elite” (or “elite hacker”); other spellings exist, including 1337, 31337, 133t, l33t, and leet. An Internet subculture that emerged in the 1990s, marked by self-proclaimed hyper-competence with computer and network technology, hostility to DRM and n00bz, and, above all, an aggressively ungrammatical jargon that substitutes numbers or symbols for letters, and ends nouns with “z” to indicate plural. L337-speak is not the same as text-speak, although there’s some crossover. The original intent of the jargon was to escape the notice of keyword filters and searches.
L4
Level 4. See cache.
L4S
Low Latency, Low Loss, and Scalable Throughout. A standard in development by the IETF as of 2023 to improve Internet performance by reducing latency of packet forwarding. It involves active queue management (AQM) by routers, and an IP extension called Explicit Congestion Notification (ECN) that allows IP packet headers to carry information about delays they’ve encountered.
LabVIEW
A National Instruments graphical programming environment and toolset, for the creation of virtual instruments (VI) that perform traditional hardware operations in software. The language, G, provides automatic memory management and support for multi-threading, and can link in modules written in text-based languages. Unlike languages used in other visual programming environments, G is itself visual; it doesn’t offer a text-only view of source code. The files have a .vi filename extension.
A LabVIEW VI typically consists of a front panel with user controls and display, a flowcharted block diagram that defines behavior, and an icon & connector pane that allows the VI to be used by another VI. A VI or set of VIs can run within the LabVIEW environment, or can be compiled into a stand-alone executable – although it doesn’t really stand alone, since it still requires an installation of the LabVIEW run-time engine. A VI can also be made to run on a FPGA or an embedded device (including one with a real-time OS), which requires a conversion module. Such a downloaded VI can communicate with its front panel on the programming PC, or can be headless, i.e., running without a user interface.
LabVIEW NXG (Next-Generation) is a newer version with additional capabilities and features, but lacking some of the original’s tools.
LAC
Link Access Control.
ladar
Laser Detection And Ranging. A general class of remote sensing applications using lasers. Typically used for 3D imaging. See also lidar.
LAMP
Linux, Apache, MySQL, PHP/Perl/Python. An operating system bundled with a set of applications & features commonly used by software developers.
LAN
Local Area Network. A network of computers that are physically close to one another, typically in the same building. The types of connections the network uses set a physical limit on the maximum distance from any given member to the nearest network switch. A larger network of multiple LANs stitched together with routers and bridges might still, loosely speaking, be called a LAN, even if the proper term is MAN or WAN.
LAN connections can be wired, wireless, or a combination of the two. The various flavors of wired (802.3) and wireless (802.11) Ethernet are by far the most common, though not the only options.
LANL
Los Alamos National Laboratory. US Department of Energy research facility in Los Alamos, NM.
LAP
Link Access Procedure (or Protocol). A number of varieties exist:
LAP-B – Link Access Procedure – Balanced. A packet-switched subset of HDLC used on the ISDN-B channel, and in OSI to control layer 2 (data link layer) functions.
LAP-D – Link Access Procedure – D-channel. The data link layer (layer 2) protocol specified by the OSI model, and used for the ISDN-D channel. Almost identical to X.25’s LAP-B, it’s a packet protocol that uses HDLC and handles error correction. A LAP-D frame consists of the 8-bit flag 01111110, the 16-bit address, the 16-bit control field, layer 3 protocol information & user data, 16-bit CRC, and finally the 8-bit flag again.
LAP-F – Link Access Procedure – Frame.
LAP-M – Link Access Procedure – Modem. The V.42 error control protocol employs LAP-M as its primary error control scheme.
LAP-V – Link Access Procedure – Video Telephone.
Laplace transform
Named for French mathematician Pierre-Simon, marquis de Laplace (1749-1827). An operation to convert a continuous time-domain system, modeled as a linear differential equation with constant coefficients, into the frequency domain, where it’s just an algebraic expression. Compare Fourier transform, Z-transform.
laser
Light Amplification by Stimulated Emission of Radiation. Invented in 1960, the laser uses some lasing medium in an optical resonator, which is essentially two tiny mirrors facing one another. It stimulates the valence electrons of the medium’s atoms or molecules to a high energy state; when they spontaneously return to the lower energy state, they emit photons, creating light that’s reflected between the mirrors. One of the mirrors is designed to leak, passing a coherent, tightly collimated beam of light that, in theory, has a single frequency determined by the valence electrons of the lasing medium. In practice, it’s not that simple.
The spaser (surface plasmon amplification by stimulated emission of radiation) is the tiniest laser known. The first created, in 2009, is 44 nm diameter but generates 530 nm wavelengths. It’s a gold core surrounded by a dye-filled glasslike shell. Incident light excites surface plasmons (fluctuations in electron density moving along the surface of the core), which then lase as they give up energy.
The lowest-power type as of 2014 is the polariton laser, which uses exciton-polaritons as the lasing medium. An exciton is an atom with an electron boosted from the valence band to the conducting band. An exciton-polariton is a quasiparticle created by the coupling of an exciton in a crystal lattice with a photon. Because exciton-polaritons are already in a high-energy state, they don’t require stimulation to cross a threshold, so they lase readily.
LATA
Local Access and Transport Area. An area of the US within which a local phone service provider (LEC) operates. It corresponds, more or less, to an area code. Service between different LATAs must be provided by an IXC. See divestiture.
latch
Also called a transparent latch. In digital logic circuits, a component that maintains an output of binary values received from an external source to allow time for an external destination to read them. It can be switched between this latched state and the transparent state in which its outputs change to match the inputs, allowing the circuit to function as if the latch isn’t there.
LaTex
See TeX.
L-band
See microwave band.
LC
See fiber.
LCC
Leadless Chip Carrier. See JEDEC.
LCD
Liquid Crystal Display. A technology originally used for the simple alphanumeric displays of calculators, digital watches, and other instruments. It has since moved into cell phones, PDAs, notebook PCs, desktop PC monitors, TV sets – pretty much anything with a flat-panel display (FPD). The LCD is much lighter, thinner, and less power-hungry than a CRT, although it can’t yet achieve the same contrast or color depth.
Sandwiched between the front and back panels of the screen, rod-shaped molecules of a liquid crystal align themselves in response to electrical signals from a display matrix, passing or blocking light from the back of the panel. The active-matrix LCD (AMLCD), also called a thin film transistor (TFT) display, has a transistor for each color (red, green, blue) of each screen pixel, unlike the older passive-matrix LCD (PMLCD). The transistors provide better control over the light passing through the screen – they filter the light rather than simply passing or blocking it, enabling brighter, more accurate colors and faster updates without ghosting. AMLCD therefore has the better display, which has made it standard in the FPD market despite PMLCD being cheaper and using slightly less power.
LED TVs are LCDs that use an array of LEDs for backlighting, rather than the CCFLs standard in earlier sets. Typically, the LEDs are arranged in a strip running around the edge of the panel, and use light guides to distribute the light evenly across the back of the screen. Quantum dot technology is expected (2012) to improve the color depth and power efficiency of LCDs.
LCoS
Liquid Crystal on Silicon. A HDTV technology that uses reflective liquid-crystal layers on CMOS chips.
LCX
Low-voltage CMOS Extension, possibly, but the Low Voltage Logic Alliance (Fairchild, Motorola, and Toshiba) say it’s not an acronym. An extension of the CMOS logic family that operates at 3V but tolerates 3.3V or 5V I/O. A 1997 addition, VCX, works at 2.5V but tolerates 3.3V I/O.
LDAP
Lightweight Directory Access Protocol. A member of the TCP/IP protocol suite at the Application layer, this slimmed-down version of X.500 is used as the Internet standard for server-to-server directory services. Directory services help you name, describe, find, and protect resources over networks, creating a logical network rather than just duplicating the physical distribution. Each OS and application uses its own directory, so a DAP is needed to access them all.
LDC
Long Distance Carrier.
LDO
Low Drop-Out. A type of DC linear voltage regulator that uses an open collector or open drain rather than an emitter follower design, and can operate efficiently even with a low differential between Vin and Vout.
LDPC
Low-Density Parity Check. A family of error-control codes conceived in 1962-1963 and forgotten for over 30 years, because they require intensive computations that weren’t practical for 1960s technology. Theoretically able to come within 0.0045 dB of a channel’s Shannon limit, but they have long coding delay. LDPC, along with turbo code, constitutes a new, third class of FEC called iterative codes.
Least Astonishment, Law of
From “The Tao of Programming” (1987), by Geoffrey James. The design principle that a system should respond in the way least astonishing to the user.
LEC
Local Exchange Carrier. A local phone service provider – a company that owns or operates local lines and switching equipment within a single LATA. It can be a BOC (an incumbent LEC, or ILEC), or one of the many local companies outside the Bell system (a competitive LEC, or CLEC). See divestiture.
LED
Light-Emitting Diode. See diode. For LED TVs, see FPD; for LED lighting, see light.
LEMO
A Swiss company founded in 1946 by on Mouttet. “LEMO” usually means their family of compact, shielded, multi-contact metal connectors with self-latching push-pull coupling.
LEMO connectors
LEMO connectors
LEO
Low Earth Orbit. An orbital altitude of roughly 500-1500 km, above the atmosphere but below the first Van Allen belt. This orbit has low eccentricity, 30°-90° inclination, a period of 1-2 hours, and a speed of about 7.3 km/s. Signals passing through a LEO satellite incur a 10-40 ms round-trip delay.
LET
Light-Emitting Transistor. First demonstrated in 2004, still experimental in 2015. It can modulate emitted light at much higher frequencies than LEDs can achieve.
Leyden jar
Named for Leiden University (the Netherlands), where it was developed in the 1740s. A simple capacitor made from a jar or similar container lined inside and outside with metal foil. The outside foil is grounded. An electrode inserted through the non-conducting lid is in contact with the inside foil, allowing it to be charged – traditionally via static electricity from a glass or amber rod rubbed with fur. The jar discharges when contact occurs between the electrode and the grounded outer foil.
Leyden jar diagram
Leyden jar
The image shows a positively charged jar, because that’s easier to achieve. The electrode is (obviously) a conductor, which, by definition, readily gives up valence electrons. However, it is possible to negatively charge a Leyden jar, e.g., by using a plastic rather than a glass rod.
LF
(1)
Low Frequency. 30-300 kHz. See RF.

(2)
Line Feed. See CR/LF.
LFE
Low Frequency Effects. See sound.
LFSR
Linear Feedback Shift Register. A clocked binary data register that shifts data serially through its m stages to the output, starting with some non-zero initial fill condition. The repeating output pattern is called a linear recursive sequence (LRS), and its length N can’t exceed 2m – 1.
Fibonacci [5,3,2,1] LFSR
Fibonacci [5,3,2,1]f LFSR
Galois [5,4,3,2] LFSR
Galois [5,4,3,2]g LFSR
There are two ways of building a LFSR. They assign opposite orders to the stage weights, as shown in the figure. The Fibonacci implementation, or simple shift register generator (SSRG), uses out-of-line configuration. It feeds a modulo-2 sum of its output g0 and selected intermediate stages (tap points) back to the input gm. The Galois implementation, also called a multiple-return or modular shift register generator (MRSRG or MSRG), has an in-line configuration that uses the output gm as the input g0, and also modulo-2 adds it to the tap points. It’s preferred over the Fibonacci form because it uses fewer gates, making it faster. There’s more than one way of representing these implementations.
The tap-point configuration is commonly described by a generating polynomial (GP) of the form [111101] or [x5 + x4 + x3 + x2 + 1], to continue the example from the figures above. The GP has the same value for either LFSR type. Certain sets of feedback tap points will result in the special case N = 2m – 1, which is called a maximal-length sequence or m-sequence. The initial fill determines the phase of an m-sequence, but not its length.
A Galois and a Fibonacci LFSR with identical, maximal tap points and the same initial fill will produce mirror-image outputs. Correctly modifying the initial fill of one or the other will make the outputs identical and phase-matched.
An m-sequence is always balanced, containing 2n–1 ones and 2n–1 – 1 zeros. Furthermore, its normalized auto-correlation function R[k] = 1 when shift k = rN (r is any integer), but R[k] = –1/N when k ≠ rN. If the tap points are non-maximal, then the choice of initial fill can change LRS length. See Gold code, PN, tap points.
LFTR
Liquid Fluoride Thorium Reactor. A type of molten salt reactor that uses a mix of thorium and uranium-233 as the dissolved fissile material.
LGA
Land Grid Array. For the IC chip form factor, see JEDEC. For the Intel CPU socket series, see Socket #.
LH2
Liquid Hydrogen (H2). The most powerful liquid fuel for rockets, paired with liquid oxygen (LOX).
LHCP
Left-Hand Circular Polarization. See polarization.
LICC
Low-Inductance Ceramic Capacitor.
lidar
Light Detection And Ranging. A general class of remote sensing applications using optical-frequency sources – both lasers (ladar) and other types of light emitters. Lidar is much more susceptible to precipitation than radar, which simultaneously limits its uses and gives it value for atmospheric applications. At certain frequency ranges, it penetrates foliage well, which makes it a useful tool for military units and archaeologists.
Li-Fi
Light Fidelity. Creating a LAN with infra-red or visible light. It offers the potential for higher data rates than Wi-Fi technology, but suffers from range and propagation problems. Some proposed designs combine the signal carrier with room/office lighting.
LIFO
Last In, First Out. A type of shift register that handles data this way. Sometimes called a push-pop register (see stack). It could also be called a FILO, but it isn’t. Contrast FIFO.
light
Electromagnetic (EM) radiation detectable by the human eye. The visible spectrum covers frequencies of roughly 430 to 750 THz, or wavelengths of 0.7 to 0.4 µm. It’s conventionally divided into red (0.700-0.620 µm), orange (0.620-0.592), yellow (0.592-0.578), green (0.578-0.500), blue (0.500-0.446), and violet (0.446-0.400). That makes ROYGBV rather than ROYGBIV, since indigo is no longer treated as distinct from blue and violet. IR and UV are just outside the visible spectrum.
So much for light as a physical phenomenon. As for light as a technology...
Incandescent lamps, developed in the 19th century, run electrical current through a fine wire filament inside a bulb filled with an inert (non-combustible) gas, causing the filament to glow brightly from the heat of electrical resistance without burning up. In modern bulbs, the filament is an alloy of tungsten (chosen for its high melting point and toughness), and the gas is some combination of argon, nitrogen, and occasionally krypton or xenon.
Manufacturers and vendors have traditionally ranked incandescent bulbs by the watts of power they consume. The meaningful numbers, however, are lumens of light emitted and light quality. The latter is rated in terms of both color temperature in Kelvin (the temperature a blackbody radiator would have to be to radiate the same hues – daylight is about 6500 K, compared to an incandescent bulb’s 2700 K to 3000 K) and an international standard called the color rendering index (CRI). People can be very sensitive to light hue. They find yellowish or even reddish light more appealing, and characterize it as warmer, although it’s cooler on the color temperature scale. Light closer to the blue end of the scale is better for reading and some other tasks, but is more disruptive to the circadian rhythms of people and animals exposed to it at night.
Incandescent lighting is highly inefficient. It radiates no more than 10% of the power it consumes as light, with the rest emitted as heat. For this reason, many countries, including the US, have attempted to end the manufacture and sale of standard incandescent bulbs. The potential replacements:
Halogen lamps are a type of incandescent. The inert gas contains a small amount of one of the highly reactive halogen gases (usually vapor of bromine or iodine), and the bulb – made of quartz or heat-tempered glass – is small to confine the gas close to the filament. In the halogen cycle, vaporized tungsten from the filament reacts with the halogen to form a halide compound, which in turn breaks down on the filament to re-deposit the tungsten and free the halogen. This produces very bright light more efficiently than a standard incandescent bulb, but the bulb also gets significantly hotter.
Fluorescent lamps pass current through a thin mercury vapor suspended in low-pressure inert gas (argon, krypton, neon, or xenon), causing excitation of valence electrons in the mercury atoms. As the electrons decay to their lower states, they emit ultraviolet light. A coating on the interior of the bulb absorbs the UV light and fluoresces at visible wavelengths. Fluorescent lights achieve better than 20% efficiency, and last longer than incandescents. On the down side, they contain mercury, and many people dislike the hue of fluorescents – even those with color temperature matching incandescent light. Also, for each bulb/tube, the fixture must include a starter circuit that vaporizes mercury in the tube when it’s first turned on, and a ballast circuit to limit current and adjust its frequency. The ballast sometimes makes an annoying buzz.
CFL (compact fluorescent lamp) bulbs fold the long, straight tube of the standard fluorescent light into a small volume comparable to that of incandescent bulbs, which they’re intended to replace. Like other fluorescents, they’re more efficient and longer-lasting than incandescents. The newer integrated CFL bulbs have a built-in ballast circuit, so they use the same socket as incandescents too. However, besides being relatively expensive, they still use toxic mercury vapor, and they prompt the same complaints about light quality as other fluorescents. Furthermore, their long-life advantage diminishes if they’re turned on and off frequently.
LED lights are more rugged, more efficient, and more long-lived than CFLs, and contain no mercury. Initially, they were much more expensive, but that has changed. It’s hard to scale them up to higher-luminosity outputs in the standard bulb form-factor. This is because, although they produce little waste heat, they also have low tolerance for heat and must be able to disperse it quickly. Also limiting their output is that, above a certain current level, LEDs suffer declining efficiency as more electrons collide to generate waste heat rather than completing the energy level transitions that produce light. Unlike other light sources, LEDs require DC power. (AC-driven LEDs do exist but have poor efficiency.) The usual solution is to build an AC-to-DC driver circuit into the LED bulb, or into the lamp fixture. A non-profit called the Design Lights Consortium (DLC) provides ratings for LED lighting products.
The neon light is more for signs and displays than for illumination. It’s a type of cold-cathode lamp, an ancestor of the technology in the PDP. It consists of a sealed tube with an electrode at each end, filled with any of a number of gases. A very high voltage across the electrodes ionizes the gas to make it a charge carrier, which fluoresces at wavelengths (hence colors) characteristic to the gas: neon is red, helium yellow, mercury vapor blue, carbon dioxide white, etc.
The table below gives watts (W) of power consumed and lumens (lm) of light emitted for some roughly comparable incandescent, CFL, and LED bulbs, assuming 120 VAC power.
Incandescent CFL LED
25W, 210 lm 7W, 375 lm 6W, 300 lm
40W, 460 lm 9W, 600 lm 7.5W, 430 lm
60W, 890 lm 13W, 850 lm 12.5W, 806 lm
75W, 1180 lm 20W, 1280 lm 17W, 1100 lm
100W, 1750 lm 26W, 1570 lm n/a
150W, 2850 lm 32W, 2000 lm n/a
Through long use, the lighting industry has standardized lightbulb shapes and sockets. The shape code is one or more letters, followed by a number representing the bulb’s widest diameter. The number means eighths of an inch in the US; elsewhere, it often (not always) changes to represent millimeters. The table below lists some of these shape codes.
Code Meaning Description
A Arbitrary The classic shape, sometimes described as a pear or teardrop
B Bullet, or Bulged  
BR Bulged Reflector A variant of R with a more pronounced bulge to the top
BT Bulged Tube or Blown Tubular or Blunt Tip  
C Candle Shaped like a candle flame, used for chandeliers & decorative lighting
CIR Circular or Circline A ring-shaped bulb for fluorescent lamps
E Ellipsoidal Egg-shaped; not to be confused with the E## standard for threaded socket & base
ER Ellipsoidal Reflector Longitudinal cross-section is a truncated ellipse rather than the cone of the R type
F Flare or Flambeau Flame-shaped & wavy, for decorative lighting
FL Flood Sub-type of R with a wider beam
G Globe Common for lighting bathroom mirrors
MR Multi-faceted (or Mirrored or Metallic) Reflector A type of R with mirror coating
PAR Parabolic Aluminized Reflector More rugged than the basic R, common for outdoor fixtures
R Reflector Internal coating reflects light out the top of the bulb, for directional lighting; see also SP and FL
S Straight-sided, or Sign  
SP Spot Sub-type of R with a narrower beam
T Tubular Common for fluorescent lights with bi-pin base at each end
There are also standards for the base of the bulb and the corresponding socket. The E## socket/base standard is the familiar threaded type, where E stands for Edison and ## is the diameter of the base in millimeters, measured across the outer threading. In the US, thanks to Thomas Edison’s patents, the most common bulb is the A19-E26: a pear-shaped, 2 3/8" diameter bulb with 26 mm diameter (just over 1") threaded base. E26 and the European E27 are both referred to as a medium Edison screw, and are usually interchangeable. Other base/socket standards include GU## (twist-and-lock), BA## (bayonet, e.g., the standard flashlight bulb), and the bi-pin connector used for many fluorescent tube lamps.
Bulb with E26 base
E26 base
Bulb with E14 base
E14 base
Bulb with BA9s base
BA9s base
Bulb with GU24 base
GU24 base

Common US bulb base types

lightning
Through one or more poorly understood mechanisms, violently agitated clouds develop widely separated regions of electrical charge: negative ions at the cloud base, and positive ions higher up. This is at least in part because liquid water molecules more readily hold negative charge, while ice crystals take on a positive charge. Negative charge in the lower cloud repels negative charge in the Earth’s surface, leaving ground and water beneath the storm with a net positive charge.
Within the cloud, ice crystals with polarized charge ionize air molecules, creating charge carriers that form the stepped leader – a fast-moving, intricately branching flow of nitrogen and oxygen ions with a bluish-purple glow that extends from the cloud in a series of erratic jumps. Its random, jerky motion occurs because it’s not moving toward the ground, but rather away from the cloud, being repelled by the negative charge of the cloud base but initially too far from the ground to be attracted by the latter’s positive charge.
As a stepped leader approaches the ground, the ground sends out its own bluish-purple leaders, called positive streamers, which can reach more than 30 meters in length. When a branch of the stepped leader meets a positive streamer, they complete the channel, allowing a massive surge of current between cloud and ground that races through every branch of the stepped leader. This current, the wavefront of which travels up from the ground at one third to one half the speed of light (as electrons begin flowing, much more slowly, the other way), overloads the relatively modest conducting capacity of the ionized air channel, heating the air to tens of thousands of degrees. The air expands explosively, producing a sonic cacophony of cracks and booms.
That’s a lightning bolt. It usually flashes repeatedly as additional regions of cloud charge are drawn into the conducting channel. It can carry as much as 30 mega-amperes (MA) at 300 megavolts (MV), but very briefly.


Your browser does not support the HTML5 <canvas> element used for this animation
Lightning between clouds, or within a cloud, is more common than the cloud-to-ground form but of much less concern to people. Ground-to-cloud (positive) lightning is a rarer phenomenon in which a strike develops between a negatively charged region of ground and a positively charged cloud-top, usually after high-altitude winds have pushed the top of the cloud horizontally away from the base. In these cases, the stepped leader rises up from the ground, not down from the cloud. Positive lightning is especially dangerous, both because much higher voltage (~1 GV) builds up before the strike can cross such a distance, and because it can hit an area far outside the storm – the fabled bolt from the blue. Positive lightning is associated with some of the mysterious high-altitude photoelectric effects known collectively as TLEs.
Apparently reputable sources maintain that positive lightning can also occur due to clouds with positive charge at the base, a reversal of the accepted pattern.
Dark lightning is yet another incompletely understood phenomenon that occurs high in the atmosphere, when the electron flow in a cloud-to-cloud lightning strike has so much kinetic energy that the resulting electromagnetic waves are beyond the visible-light spectrum. This includes gamma radiation, referred to as a terrestrial gamma-ray flash (TGF). The formation of anti-matter is involved.
Lightning creation is not confined to clouds of water vapor. It can also arise from clouds of particulate matter such as those from volcanic eruptions and very large fires. The heat beneath these clouds causes the necessary turbulence to drive charge separation.
lightning rod
There are two common, contradictory misperceptions about lightning rods. One, first suggested by Benjamin Franklin (1706-1790), is that they reduce the chance of a lightning strike because they dissipate ground charge gradually into the air. Although a pointed rod does indeed dissipate charge by ionizing air molecules, the effect is far too small to change the ground’s massive electrical potential. Lightning arresters and lightning dissipaters have never been proven to work.
The other is that they actually attract lightning, which is misleading, being true only in a very localized sense. Their height is such a small fraction of the distance to the clouds that it doesn’t meaningfully increase the chance of lightning striking their location. What they do, in the event that a lightning strike does develop on a building, is to provide a low-resistance path for the surge of current so that it doesn’t pass through the structure of the building. Ships with non-metallic hulls have lightning masts, tied to conducting elements in contact with the water, and they work the same way.
The glass balls on some older lightning rods aren’t only for decoration. They tend to shatter when lightning strikes the rod, providing a visual indication that the rod and the building should be checked for damage.
Li-ion
Lithium Ion. See battery.
LIN
Local Interconnect Network. A serial, master-slave data bus protocol for automobiles, created by a consortium of carmakers. It uses 12V power, and sends 2- to 8-byte data frames at up to 20 kb/s over 40m distance. Cheaper, slower, and simpler than CAN. Compare also I2C, SPI.
linearity
An electronic or RF system is linear if, given a pair of inputs x0 and x1 that produce outputs y0 and y1, an input [x0+x1] produces [y0+y1]. Linear amplifiers are simpler but less power-efficient, so applications that require high efficiency, such as satellite comms, use non-linear devices such as the TWTA.
linker
A program that accepts one or more code object modules generated from source code by a compiler and/or assembler, and combines them (and perhaps a run-time library as well) to produce an executable.
lint
The name comes from dryer lint – stuff to be gotten rid of. It means any software tool that analyzes source code prior to compilation for errors, bad coding practices, and other issues. It began as a Unix command-line operation.
Linux
A Unix-compatible kernel created by Finnish programmer Linus Torvalds in 1991, mostly in C, but partly in processor-specific assembly language. It was adopted as the kernel for the GNU operating system. The resulting GNU/Linux OS is popularly called just “Linux”, which annoys the good people at the Free Software Foundation (FSF).
It’s the world’s most successful open-source software. Thousands of programmers contribute to each kernel release, and Linus still supervises the process. There are many flavors of the GNU OS built around the Linux kernel. See GPL.
The original release contained about 10,000 lines of program code. Version 3.3 (2012) has more than 15 million lines, which is still far less than the Windows and Mac OS kernels.
Lisp
List Processing. See programming language.
Lissajous pattern
Named for French mathematician Jules Antoine Lissajous (1822-1880). A comparison of two signals, usually generated on an oscilloscope in XY mode. The vertical axis, as usual, is the amplitude of an input signal (Y). The horizontal axis, instead of being time, is the amplitude of the other input (X). The I and Q of FSK and MSK make circular patterns. See constellation.
little-endian
See endian.
LIU
Line Interface Unit. A gateway device between a transmission line (e.g. a T-1) and its users. One side deals in synchronous serial data, and the other side in the line signal. Compare to the NIU (see HFC).
LLA
Latitude Longitude Altitude. Another name for the geographical coordinate system. Earth is not a perfect sphere, so conversion to and from ECEF requires some correction factors (the WGS84 parameters) on top of the basic trigonometry.
LLM
Large Language Model. An artificial neural network that’s been trained on an enormous quantity of text examples, and as a result is able to provide coherent responses to natural-language input.
LLNL
Lawrence Livermore National Laboratory. US Department of Energy research facility run by UCal in Livermore, California, east of San Francisco.
LLVM
Low-Level Virtual Machine, originally. An open-source software development kit for creating highly optimized compilers and runtime environments.
LM
LAN Manager. See NTLM.
LMDS
Local Multipoint Distribution Service. A ~10 Mb/s cellular broadband technology that uses frequencies in the 26, 28, and 31 GHz bands to provide two-way connections for voice, data, and video (wireless cable TV). A 1 GHz bandwidth uses 850 MHz for 40 MHz downstream channels, and 150 MHz for 2 MHz upstream channels. Channel rate varies. Conceived as a medium for digital subscriber TV, it has failed in that role, and the frequencies are mainly used now (2011) for linking wireless provider networks.
LMR-###
A Time Microwave Systems brand of coaxial cable with especially low losses at microwave frequencies. Common types include LMR-195 (equivalent to RG-58), LMR-200, LMR-400, LMR-600, LMR-900, and LMR-1200, with the larger numbers having smaller losses. LMR cables have a foam polyethylene dielectric, and a bonded aluminum tape outer conductor with a tinned copper braid over it. The LMR-DB versions of the cables include a moisture-resistant flooding compound that covers the braid. LMR-FR cables are fire-retardant.
LMS
Least Mean Square. An adaptive noise-canceling filter algorithm.
LNA
Low Noise Amplifier. A receiver pre-amplifier, placed very near the antenna to amplify the incoming RF signal before it can be degraded by noise in the rest of the receiver.
LNG
Liquefied (or Liquid) Natural Gas. Used for purposes as mundane as cooking and heating, and as exotic as rocket fuel. In the latter role, while it generates less power than an equal mass of liquid hydrogen, it’s easier and cheaper to store.
Raw natural gas is typically 70-90% methane (CH4), with smaller amounts of other alkane hydrocarbons in the paraffin series – ethane (C2H6), propane (C3H8), butane (C4H10), pentane (C5H12), etc. – plus trace compounds (CO2, O2, N2, H2S, and noble gases). Most of these are colorless and odorless, so suppliers add small amounts of sulfur compounds to give NG a smell. When burned efficiently with oxygen, methane produces CO2 and water.
Under standard atmospheric pressure, NG condenses to a liquid around -260° F, with about 1/600th the volume of the gaseous state. Such low temperatures being impractical, LNG is invariably also PNG (P for pressurized) to permit room-temperature storage, at pressures as high as 3600 psi (pounds per square inch).
LO
(1)
Local Oscillator. The component of a super-heterodyne radio receiver that generates a frequency tracking that of the RF carrier. The RF carrier is mixed with the LO frequency to produce the downconverted IF.

(2)
Local Office. See central office.
LOC
Lines of Code. A unit of measure for analyzing software projects. Because not all lines in a source file are of equal interest, some measurement approaches specify logical lines of code (LLOC), non-commented lines of code (NLOC), effective lines of code (ELOC), source lines of code (SLOC), or equivalent source lines of code (ESLOC), all of which mean roughly the same thing: don’t count comments. Large projects commonly inject a K for kilo, meaning thousand, in which case the acronyms become KLOC, LKLOC, NKLOC, EKLOC, KSLOC, EKSLOC.
LOD
Level Of Detail. A memory-saving technology for computer graphics. It renders virtual objects that are farther away from the user’s point of view with less precision than objects that are closer. As of 2024, LOD is used for geometric and textural data, but not for ray-tracing.
local loop
The 24 AWG copper UTP line that runs from your house to the telephone company’s central office (although nowadays, it’s probably copper only as far as the neighborhood ONU, and fiber from there on out). Getting it from your house’s telephone box to the outlets around the house is your problem, not the phone company’s. The copper part of a local loop can’t be more than about 5000m long, or the signal will be too weak, so subscribers must be within 5 km of a central office or ONU.
logical channel
A communications channel established by packet routing (e.g, IP) rather than a continuous connection.
logic analyzer
Also called a logic state analyzer. Instrument for displaying, capturing, and comparing digital signals. An oscilloscope can capture and display signals too, but the logic analyzer can apply digital logic and digital triggering to the inputs – something that the o-scope, a fundamentally analog instrument, can’t do. Contrast also network analyzer, spectrum analyzer.
logic bomb
See malware.
logic family
Any standard for using electrical signals to communicate binary data. The simpler unbalanced line standards compare a single voltage level to a fixed reference, usually ground, to determine a value of 0 or 1. The more robust balanced line or differential signaling transmits two equal and opposite voltage levels, one positive and one negative, and reads the differential between them as 0 or 1.
TTL and ECL are the oldest logic families, CMOS the most common. (Also, see the RS-### series.) Though many analog devices still use 5V power supply, hence send signals at that voltage, digital devices are increasingly at 3.3V, 2.5V, 1.8V, or lower while using ever-higher clock speeds – hence the proliferation of logic families to keep up with hardware changes. Level shifters can enable communication between different voltage levels of the same family, and in some cases between families.
BTL – Backplane Transceiver Logic. Differential inputs, supply voltage 5V or 3.3V, reference 2.1V, termination 33Ω.
CML – Current Mode Logic. ECL without the emitter follower, and termination from 25 to 200Ω.
CMOS – Complementary Metal-Oxide Semiconductor. The standard binary logic family for the microprocessors in modern computers. Supply voltage (VDD) of the original CMOS is 5V, but low-voltage CMOS (LVCMOS) instead uses 3.3V, 2.5V, 1.8V, or 1.5V. There is no reference voltage or line termination. Ground (often denoted VSS) is logic 0, and VDD is logic 1. CMOS logic circuits tend to be smaller and use less power than TTL or ECL, although their power dissipation increases with frequency. Their traditional shortcomings have been inherent input capacitance, which limits speed, and the inability to drive large output loads. Newer implementations have reduced these disadvantages.
CTT – Center Tapped Termination. A type of SSTL. Supply voltage VCC is 5V or 3.3V, reference VCC/2, termination 10 to 100Ω.
ECL – Emitter-Coupled Logic. Common in 1970s computer ICs. CMOS is now used for the same purposes, except in a few applications that require very high speed. Differential (balanced-line) signaling, supply voltage -5.2V, 10 to 200Ω termination.
GTL – Gunning Transceiver Logic. Differential inputs, supply voltage 5V or 3.3V or 2.5V, reference 0.8V, termination 50Ω.
HSTL – High-Speed Transceiver Logic. Fast, divided into four classes. Push/pull CMOS output buffers, supply voltages 5V to 1.5V, reference 1.5V, termination 50 to 70Ω.
LVCMOS – Low Voltage Complementary Metal-Oxide Semiconductor. See CMOS, above.
LVDS – Low Voltage Differential Signaling. A constant-current standard defined by ANSI/TIA/EIA-644 and IEEE 1596.3, with high speed and low power. It’s the physical medium for industry standard FPD-Link digital video, and is also popular for space applications.
Unlike the older RS-422, it uses a nearly constant voltage – typically an offset of about 1.25V and a swing of about ±175 mV, 350 mV total. It can achieve rates of about 90 Mb/s over a 10-meter distance, and tolerates a shift of at least ±1V between driver and receiver ground levels. Line termination resistors from + to - at the receiving end are needed, both to reduce impedance-mismatch echoes and to provide a current return path. These are typically 100Ω to 120Ω, matched to the actual cable. Most of the current in the transmission line pair flows through this termination.
LVPECL – Low Voltage Positive Emitter-Coupled Logic. See PECL, below.
LVTTL – Low Voltage Transistor-Transistor Logic. See TTL, below.
PECL – Positive Emitter-Coupled Logic. ECL with a positive supply voltage (+5V instead of -5.2V) and 10 to 100Ω termination. The balanced lines must both be terminated, typically to a non-ground voltage. Low-voltage PECL (LVPECL) uses +3.3V, +2.5V, +1.8V, or +1.5V instead of +5V.
RSL – Rambus Serial Logic. A 500 MHz serial bus. Supply voltage is 3.3V, reference 2.1V, termination 15 to 50Ω.
SSTL – Stub Series Terminated Logic. Common with SDRAM up to 200 MHz. Supply voltage VDD is 5V, 3.3V (SSTL3), or 2.5V (SSTL2); reference VDD/2, termination 25 to 50Ω.
STTL – Schottky Transistor-Transistor Logic. See TTL below.
TTL – Transistor-Transistor Logic. The oldest digital logic standard, now in decline as its traditional advantages over CMOS – higher speed and higher output load drive values – erode. It’s compatible with CMOS. Supply voltage is 5V, with no reference or line termination. It reads line values 0V to 0.8V as 0 or OFF, and 2.0V to 5V as 1 or ON. Stable voltages from 0.8V to 2.0V are disallowed. A bipolar transistor is used in each logic function’s totem pole (open collector) output driver as well as in the input buffer, so that logic values are passed directly from one transistor to the other – although Schottky diodes are often used for interconnects (Schottky TTL). Low-voltage TTL (LVTTL) uses 3.3V, 2.5V, 1.8V, or 1.5V instead of 5V.
LORAN
Long-Range Aid to Navigation. An obsolete position, navigation, & timing (PNT) system run by the US Coast Guard using terrestrial transmitter stations. The final implementation, LORAN-C, covered the continental US, most of Alaska, Canadian coasts, and the Bering Sea (aided by Canadian and Russian stations). Falling into disuse after the rise of GPS, it ceased operation in 2010. Compare TACAN.
Enhanced LORAN (eLORAN), developed out of LORAN-C in the 21st century, is still active. It transmits in the 90 to 110 kHz band, and has a range of about 2000 km. Positioning accuracy can be less than 10 meters, which isn’t as good as GPS, but it has the advantage of being much harder to jam or spoof because of encryption, authentication, and the much higher transmission power of its terrestrial stations. It offers a LORAN data channel (LDC) signal that system customers can use to send data to their remote installations and equipment.
LOS
Line of Sight. This means having unobstructed open space between a transmitter and a receiver. It also refers specifically to a wireless broadband technology for connecting subscribers to the Internet, using subscriber directional antennas and a base station with a high-power transmitter. It’s expensive and has problems. See NLOS, MMDS.
LP
Low Profile. See DRAM.
LPA
(1)
Linear Power Amplifier. The final amplification stage in a multicarrier transmitter, optimized to provide a linear response. This minimizes side-lobes and non-linear effects.

(2)
Low Power Amplifier.
LPC
(1)
Linear Predictive Coding. A highly compressive source-coding scheme, mainly for speech signals, that estimates future values of the sampled signal as a linear function of past values. See linearity.

(2)
Low Pin Count. A 4-bit, 33.3 MHz PC motherboard bus to support devices with low data-rate requirements, including fans, switches, keyboard, mouse, serial ports, BIOS, and TPM. Intel introduced it in 1998 to replace legacy uses of the extended ISA bus, which, with 16-bit width and 8.33 MHz clock, has the same data rate.
LPD
(1)
Low Probability of Detection. See LPI.

(2)
Laser Phosphor Display. A new (2010) display technology that, as its name suggests, uses a laser sweeping across an array of phosphors of any size or shape. Claimed to use much less power than a LCD or LED screen. See FPD.
LPDDR#
Low Power Double Data Rate #. Also called Mobile DDR (MDDR). A less power-hungry DRAM for mobile devices, based on the DDR memory used in PCs but with 16- or 32-bit wide bus as opposed to DDR’s standard 64 bits. The original LPDDR used supply voltage 1.8 V, dropping to 1.2 V for LPDDR2/3 and 1.1 V for LPDDR4. LPDDR4 has maximum theoretical throughput of about 32 GB/s with a 2133 MHz clock. Its expected successor is Wide I/O memory.
LPF
Low-Pass Filter. A filter that admits low frequencies and blocks high frequencies.
LPG
Liquefied Petroleum Gas. Propane in its liquid state.
LPI
Low Probability of Intercept. This refers to an RF signal that employs spread-spectrum (SS) modulation or unpredictable burst transmission to defeat interception and jamming. The same techniques can make it difficult to even notice (low probability of detection, or LPD). See DSSS, FHSS.
LPT port
Line Printer Terminal, originally. The now-obsolete parallel port, most often used to connect printers to PCs. The typical form factor is a DB-25S connector (see D-sub). Early versions of Windows allocated two interrupt requests (IRQs), IRQ 5 and IRQ 7, for LPT ports.
LRC
Same as RLC.
LRS
Linear Recursive Sequence. Also called a pseudo-random noise (PN or PRN) sequence. See LFSR.
LRU
Line-Replaceable Unit. A hardware module or component designed to be easily replaced by users if it malfunctions or needs an upgrade, as opposed to SRU.
LSA
Logic State Analyzer. See logic analyzer.
LSB
(1)
Least Significant Byte. The lowest-valued byte in a multi-byte expression. For example, the decimal number 396,584,882 becomes 0x17A367B2 in hexadecimal form, so its LSB is 0xB2. The least significant bit within a byte can also be denoted LSB, or LSb. Contrast MSB. See also endian.

(2)
Lower Sideband. See SSB.
LSC
Luminescent Solar Concentrator. A transparent panel that refracts some of the sunlight passing through it, including UV and IR rays, into photovoltaic cells along its edges. It’s used for windows that generate solar power while still functioning as visible-light windows.
LSD
Low Self-Discharge. See NiMH.
LSI
Large Scale Integration. A dated way of describing the density of an integrated circuit (IC). It means a chip with roughly 300 to 1000 elements – more than MSI, less than VLSI.
LSL
Logical Shift Left. See binary logic.
LSP
(1)
Label Switched Path. See MPLS.

(2)
Local Service Provider. A competitive or independent LEC (CLEC or ILEC). See divestiture.
LSR
Logical Shift Right. See binary logic.
LTCC
Low Temperature Co-fired Ceramic. A multi-layer PCB manufacturing technology that allows passive components to be incorporated into the board. The layers are sandwiched between ceramic substrata and heated to 850-875° C to fuse them together. This process permits the use of some materials, such as silver, that HTCC (above 1000° C) doesn’t.
LTE
Long Term Evolution. A sort-of-4G mobile phone network standard released in 2008 by the 3GPP, with the intent that it be achievable through upgrades of existing 2G GSM and 3G UMTS networks. It has up to 300 Mb/s OFDMA downlinks from cell tower to mobile device, and up to 75 Mb/s single-carrier FDMA uplinks, with either FDD or TDD to share spectrum between subscriber devices. It offers six different bandwidths, from 1.4 MHz to 20 MHz, and uses many RF bands covering roughly 700 to 2600 MHz plus a couple of outliers in the 452.5-467.5 and 3410-3600 MHz ranges.
LTE falls short of the ITU’s original 4G standard. Mobile providers had the standard changed so that they could advertise 4G service.
The follow-on LTE Advanced does meet the original standard, and is billed as “true 4G”. It has theoretical maximum 3 Gb/s downstream and 1.5 Gb/s upstream speeds, and uses MIMO to boost gain. As of 2013, it’s not available in the US.
LTE-capable devices have a reputation for burning through batteries quickly because many providers currently (2013) require dual-band operation: the unit must transmit on both the LTE frequencies and the older 3G frequencies.
LTI
Linear Time-Invariant. A system is linear if, given a pair of inputs x0 and x1 that produce outputs y0 and y1, an input [x0+x1] produces [y0+y1]. A system is time-invariant if time-shifting the input x produces an identical shift in the output y. A signal must be LTI to be reliably handled by DSP.
LU 6.2
Logical Unit 6.2. A protocol for IBM’s SNA. It provides peer-to-peer communications between programs without the participation of the server or mainframe.
LUT
Look-Up Table. A fundamental component of logic devices. It defines one or more output signals based on the states of a set of input signals.
LVCMOS
Low Voltage Complementary Metal-Oxide Semiconductor. See logic family.
LVDS
Low Voltage Differential Signaling. See logic family, or, for the digital video standard known as LVDS, see FPD-Link.
LVDT
Linear Variable Differential Transformer. An inverse solenoid – it converts linear mechanical displacement into a voltage proportional to the displacement, so it can be used to measure small movements. Compare RVDT.
LVPECL
Low Voltage Positive Emitter-Coupled Logic. See logic family.
LVTTL
Low Voltage Transistor-Transistor Logic. See logic family.
LWIR
Long-Wave Infra-Red. See IR.
LWR
Light Water Reactor. See reactor.
LXI
LAN Extensions for Instrumentation. Essentially GPIB over a network. Debuted 2005.
LZW
Lempel-Ziv-Welch. A data compression algorithm patented by Sperry Corp. in 1985, based on the public-domain LZ algorithm. It depends on frequently repeated 8-bit groups, so it works best for 8-bit text files.