Infrastructure

Structured cabling is a standardized star layout: every outlet or camera gets its own permanent cable, with no splices or taps, back to a patch panel in the rack, and the racks are linked by copper or fiber backbone cabling. The main references are the ANSI/TIA-568 series (USA), ISO/IEC 11801 (international) and EN 50173 (Europe), which define cable categories, maximum lengths and test limits. They are supported by TIA-569 for pathways and spaces, TIA-606 for labeling and documentation, TIA-607 for bonding and grounding, and ISO/IEC 14763-2 or EN 50174 for installation practice. The result is one shared infrastructure for data, IP phones, cameras and PoE, in which every link can be tested individually and equipment can be replaced without re-cabling.

For copper Ethernet, the maximum channel length from the switch to the device is 100 m, and it is the same for Cat5e, Cat6 and Cat6A at the speeds each category is designed for (for example, 10 Gbps over the full 100 m needs Cat6A). TIA-568 and ISO/IEC 11801 split these 100 m into at most 90 m of fixed cable (the permanent link, from the patch panel to the outlet) plus at most 10 m of patch cords in total at both ends. Cords lose more signal per meter, so their share must not be added to the fixed cable: 85 m of fixed cable plus two 3 m patch cords is fine, but 95 m of fixed cable is not. Measure the length along the real cable route, including rises, drops and service loops, not the straight-line distance on the drawing; PoE has the same limit.

There are three standard solutions. Put a switch (for example an industrial PoE switch in a weatherproof enclosure) part-way along the route so that each segment stays under 100 m; use a PoE extender, which regenerates the signal and adds up to about 100 m per unit but uses part of the PoE power itself, so check what is left for the device; or run fiber with SFP modules or media converters, which covers kilometers. Some PoE switches also have an Extend mode for a single camera (about 250 m at a reduced speed, typically 10 Mbps), which is not standard Ethernet. Stretching a copper cable beyond 100 m because it worked on the bench usually ends in CRC errors, links that fall back to 100 Mbps and PoE cameras that reboot.

At 100 Mbps it is technically possible, because 100BASE-TX uses only two of the four pairs (pins 1, 2, 3 and 6), and some installers use a special splitter to give the other two pairs to a second device. It is not recommended: the link falls outside the standard and cannot be certified, gigabit (which needs all four pairs) and standard PoE for both devices become impossible, and both devices are limited to 100 Mbps. The correct solution is a second cable or a small switch, preferably PoE, near the devices; while the cables are being pulled, adding a spare one costs little.

The three categories differ in the frequency they are specified for, and therefore in their margin against crosstalk and noise: Cat5e up to 100 MHz, Cat6 up to 250 MHz and Cat6A up to 500 MHz. In practice Cat5e is built for 1 Gbps (and 2.5 Gbps) up to 100 m, Cat6 carries 1 Gbps with more margin and 5 Gbps up to 100 m, and Cat6A is made for 10GBASE-T over the full 100 m; on Cat6 a 10 Gbps link is only supported up to roughly 37 to 55 m. Cat6A is also thicker and stiffer (typically 7 to 9 mm in diameter against about 6 mm for Cat6) and takes more room in conduits and trays. In ISO/IEC 11801 terms, the channels of these categories are called Class D, Class E and Class EA.

For CCTV cameras, IP phones, IP speakers and office computers, whose links run at 100 Mbps or 1 Gbps, Cat6 is the usual sensible choice and leaves a good margin for the future. Choose Cat6A for links that need 10 Gbps close to 100 m, for Wi-Fi 6, 6E and 7 access points with 2.5 or 5 Gbps ports, and for high-power PoE (802.3bt) in large cable bundles, where the thicker conductors typical of Cat6A run cooler. Cat5e is technically enough for 1 Gbps, but the cable is usually a small part of the total cabling cost and cabling stays in a building for more than 10 years, so it saves little in a new project. In every case, a correctly installed and tested system matters more than the number printed on the jacket.

The category of a link is set by its weakest component: cable, keystone or plug, patch panel and patch cords must all be of the same category (a principle of TIA-568). Cat6 cable with a Cat5e keystone or patch cord is in practice a Cat5e link and fails a certification test to the Cat6 limit; for 1 Gbps it usually still works, but with less margin. The same rule applies to shielding: shielded cable and keystones next to an unshielded patch cord lose most of the benefit of the shield. Decide the project's target category (for example Cat6 or Cat6A) at the start and record it, buy all components of that category, and test to the same limit.

Cat7 (Class F in ISO/IEC 11801, 600 MHz) is a fully shielded S/FTP cable; TIA does not define it, and when Cat7 cable is terminated with ordinary RJ45 connectors the link can only be rated to the level of the RJ45 hardware (in practice Cat6A), so it gives no extra Ethernet speed. Its real advantage is stronger shielding in electrically noisy places. Cat8 (2000 MHz) is designed for 25GBASE-T and 40GBASE-T over channels of at most 30 m and is mainly used between servers and switches in data centers. For CCTV cameras, IP phones and office networks, a properly installed and tested Cat6 or Cat6A system is entirely sufficient.

In CCA (copper-clad aluminum) cable the conductor is aluminum with only a thin copper coating; it is cheaper and lighter, but its resistance is about 1.5 times that of pure copper of the same size. TIA-568 limits the DC resistance of a conductor to 9.38 Ω per 100 m, which CCA usually exceeds, so even if it passes a speed test the link cannot be certified (Certification); aluminum is also brittle and oxidizes, so conductors break when bent or punched down and the contacts in keystones and RJ45 plugs loosen over time. With PoE it gets worse, because cable loss and voltage drop rise with resistance: if a channel that just reaches the 12.5 Ω worst case of 802.3at were built with CCA, its resistance would be about 18.75 Ω, the cable loss 6.75 W instead of 4.5 W, the voltage drop 11.25 V instead of 7.5 V, and the voltage at the camera would fall below the 42.5 V minimum. In practice the camera restarts at night when its IR or heater switches on.

The most reliable test is a resistance measurement: join the two wires of one pair at the far end and read the loop resistance with a multimeter at the near end, after subtracting the resistance of its leads. For 100 m of 23 AWG copper cable this is about 14 Ω and for 24 AWG about 17 Ω (at 20 °C; about 4% more for every 10 °C warmer), the standard limit is about 19 Ω, and CCA of the same size reads about 1.5 times higher, roughly 21 Ω for 23 AWG. A simpler test is to scrape a conductor lightly with a blade: copper is copper-colored all the way through, whereas CCA shows silvery aluminum under the copper layer. Also read the jacket print (category, conductor size such as 23 AWG, standard and length markings), ask for the manufacturer's test report, and be suspicious of a cable that is very light or very cheap.

Horizontal cable, the cable run inside walls and trays, has a single solid conductor in each wire with lower loss, made to be punched down once onto keystones and patch panels. Patch cords use stranded conductors, which are flexible but lose more signal per meter (TIA-568 allows for about 20% more for 24 AWG and about 50% more for 26 AWG), which is why a standard channel accepts only 10 m of cords and stranded cable must not be used for long fixed runs. Conversely, do not use solid cable for patch cords that are moved often, because solid conductors crack after repeated flexing; buy factory-made cords, or for a fixed device such as a camera use a plug designed for solid conductors.

UTP is enough for most office, CCTV and VoIP projects and is easier to install. Shielded cable is worth it where there is strong electrical noise (factories, motors and variable-speed drives, lift machine rooms, radio transmitters or heavy power cables) or in dense 10 Gbps bundles, where it reduces alien crosstalk. In ISO/IEC 11801 names, the letter before the slash describes the overall shield and the letters after it describe the pairs: F/UTP has one foil around all the pairs, U/FTP a foil on each pair, and S/FTP an overall braid plus a foil on each pair. A shield only works if the whole path is shielded (cable, keystone or plug, patch panel and patch cord) and bonded to ground.

Yes. The shield carries induced noise to ground; if it is not continuous or not bonded, the noise has nowhere to go, and a floating shield can act like an antenna and even perform worse than UTP. In practice, use shielded keystones and patch panels, terminate the foil and drain wire properly at every connector, bond the patch panel to the rack, and connect the rack with a green-yellow copper wire to the ground busbar of the rack room, which is itself connected to the building's main grounding terminal. A shielded copper link between two separate buildings is not recommended, because their ground potentials can differ and current can flow along the shield; use fiber there.

No. The PVC jacket of indoor cable is not UV-resistant; in sunlight it hardens and cracks within a few years, and moisture then gets in and ruins the pairs. Outdoors, use cable with a UV-resistant polyethylene (PE) jacket, usually black, a double jacket where mechanical damage is possible, and a gel-filled or water-blocked design for wet ducts and underground runs. PE jackets are usually not flame-retardant, so run outdoor cable only a short distance into a building and change over to indoor cable near the entry point. KDT's Cuprium range includes an outdoor Cat6 S/FTP cable with a sunlight-resistant polyethylene jacket (NC-Cuprum 6SFTPO).

Gel-filled cable has the spaces between the pairs filled with a water-blocking gel, so any water that gets in cannot travel along the cable; it suits wet, underground ducts, but the gel must be cleaned off when terminating and it is not made for indoor use. Direct burial is only allowed for cable built for it (direct-burial rated, usually with a thick PE jacket, gel or water-blocking tape and sometimes armor), and even then it is better to put it in a duct (for example HDPE), which protects it from digging and rodents and lets you replace it later. A typical depth is at least about 60 cm, more under roads; lay warning tape above the duct, keep away from power cables, seal the duct ends against water, and add a pit or pull box at bends and on long straight runs. Between two buildings, armored fiber is usually a better choice than copper.

Both jacket types are designed to limit flame spread, but PVC gives off dense black smoke and corrosive hydrogen chloride gas when it burns. LSZH (low smoke zero halogen) jackets produce little smoke (tested to IEC 61034) and no halogen acid gas (IEC 60754), which keeps escape routes visible and does less harm to people and equipment. Many fire regulations prefer or require LSZH in hospitals, airports, tunnels and metro stations, data centers and tall or crowded buildings; in Europe, cables are also given fire classes under the CPR and EN 50575 (for example Dca or Cca). LSZH is usually a little stiffer and more expensive, and it is suitable outdoors only if it is made for UV and moisture.

Both are allowed by TIA-568 and perform exactly the same; the only difference is that the orange and green pairs swap places (pins 1 and 2 with pins 3 and 6). What matters is to use one scheme at both ends of every link and throughout the project and to record it in the documentation, because mixing them on one link makes a crossover cable and mixing them across a network makes troubleshooting confusing. Today's switches, cameras and computers support Auto-MDI/X, so crossover cables are no longer needed. T568B is the more common choice in commercial projects, and keystones usually carry both color codes.

Yes; TIA-568 recognizes this as a modular plug terminated link (MPTL): a keystone on the patch panel at the rack end and a field-installed RJ45 plug at the device end, for fixed devices such as cameras, access points and IP speakers. Choose a plug made for solid conductors and for your cable category (Cat6 plugs for Cat6 cable, preferably with a load bar or an end-open, tool-less type), use a good crimping tool and keep the untwisted length of the pairs to about 13 mm at most. Outdoors, put the connection in a junction box or a waterproof connector, and test the link with a certification tester using an MPTL adapter. KDT's Cuprium range includes Cat6 RJ45 plugs in UTP and shielded versions, including end-open and tool-less models.

The most common are pulling too hard (the limit for a four-pair cable in TIA-568 is 110 N, about 11 kg of force), bending too tightly (the installed bend radius should be at least 4 times the cable diameter, about 25 mm for a 6 mm Cat6 UTP cable, usually 8 times for shielded cable, and the specifications of KDT's Cuprium Cat6 cables list 50 mm), crushing the cable with tight cable ties or staples, and kinking it around sharp edges. Untwisting the pairs by more than about 13 mm at the termination, walking on cables laid on the floor and leaving no spare length are also frequent. Use hook-and-loop (Velcro) ties loosely, support the cable on trays or J-hooks rather than on ceiling tiles, and leave a service loop, typically about 3 m in the rack room and 0.3 m at the outlet.

A common design rule (used in TIA-569 and in electrical codes) is to fill a conduit to no more than 40% of its internal cross-section when it carries three or more cables. For example, a conduit with a 25 mm internal diameter has a cross-section of about 491 mm² and 40% of that is about 196 mm²; a 6.1 mm Cat6 cable takes about 29 mm², so 6 cables fit (about 6.7). Take the internal diameter from the conduit's specification, because the nominal size is not the internal size. Keep each section between pull boxes to at most 30 m with no more than two 90° bends, and do not fill a tray at the start so that there is room for future cables.

Measure the shortest and longest runs from the rack along the real cable route and take their average; add about 10% for detours, rises and drops, and add the slack at both ends (about 3 m in the rack and 0.3 m at the outlet). Example: the shortest run is 15 m and the longest 75 m, so the average is 45 m and each run needs 45 × 1.1 + 3.3 = 52.8 m. A 305 m box gives 5 complete runs (305 ÷ 52.8 ≈ 5.8), because horizontal cable must not be spliced and the leftover is only useful for shorter runs, so 40 outlets need 8 boxes. Mixing long and short runs on the same box reduces the waste.

Run data cables in their own tray or conduit, or in a separate compartment of a divided tray; electrical safety rules also do not allow data cable in the same conduit or box as mains wiring without a barrier. The required distance depends on the cable type, the containment and the power of the circuits, and EN 50174-2 and ISO/IEC 14763-2 give the calculation; as a common rule of thumb, about 30 cm is kept in open parallel runs beside ordinary building circuits, and a grounded solid metal divider or metal conduit cuts this to half or less. Cross power and data cables at 90 degrees, and keep about 1.2 m away from strong noise sources such as motors and transformers and about 30 cm from fluorescent lamp ballasts.

Rack equipment is 19 inches (482.6 mm) wide and its height is measured in rack units: 1U = 1.75 inches = 44.45 mm (EIA-310 and IEC 60297), so a 42U rack has 42 × 44.45 = 1866.9 mm of mounting space. Add up the units you need and leave 30 to 50% for growth: 3 patch panels (3U), 3 cable managers (3U), 2 switches (2U), a fiber patch panel (1U), an NVR (2U), a rack-mount UPS (2U) and a PDU (1U) make 14U, so a 22U rack is a sensible choice. Make the rack at least 10 to 15 cm deeper than the deepest device: 600 mm is enough for patch panels and ordinary switches, but for an NVR full of disks, a rack-mount UPS or a server, check the device depth in the datasheet and expect to need a rack that is 800 to 1000 mm deep.

Put heavy items such as the UPS and battery packs at the bottom of the rack for stability, and the patch panels and switches in the middle to upper part where they are easy to reach, with a cable manager between each patch panel and switch. Make all equipment breathe in the same direction (usually cool air in at the front and hot air out at the back) and close empty units with blank panels so that hot air does not return to the front of the rack. Route power cables down one side of the rack and data cables down the other, label both ends of every patch cord, and leave clear working space, typically about 1 m in front of the rack and at least 0.6 m behind it. Wall-mount racks suit only small sites with light equipment; a heavy UPS, an NVR full of disks and servers belong in a floor-standing rack.

Practically all the electrical power used by rack equipment turns into heat: each watt equals 3.412 BTU/h, so a rack drawing 800 W produces about 2,730 BTU/h, close to a quarter of a 12,000 BTU/h air conditioner. With a PoE switch, the power delivered to the cameras becomes heat at the cameras, so only the switch's own consumption and the losses of its power supply heat the rack. Keep the air at the equipment inlets at about 18 to 27 °C (the ASHRAE recommended range); a closed cabinet needs fans, and the airflow required is roughly 3 × power (W) ÷ allowed temperature rise (°C) in m³/h, for example about 60 m³/h for 200 W with a 10 °C rise. Never put a rack in a closed cupboard without ventilation, and close empty units with blank panels (such as KDT Cuprium BL-1U/2U/4U).

Convert the rack load to amperes: current = power (W) ÷ voltage (V) for a PF close to 1, so a rack drawing 800 W at 220 V takes about 800 ÷ 220 ≈ 3.6 A. Keep the continuous load to at most about 80% of the breaker's rated current; for example, on a 16 A breaker (16 × 220 = 3520 VA) that is about 2800 VA at most. Give the rack a dedicated circuit with a sound ground and do not share it with heavy or noisy loads such as air conditioners, motors and lifts. Plug the PDU directly into a wall outlet or the UPS, do not chain PDUs one after another, and respect the rated current printed on its label. The number of sockets is not the limit; the total current of the devices must stay below the PDU's rated current (typically 10 or 16 A), so the 800 W rack above, drawing about 3.6 A, fits well within a 10 A PDU.

Solid horizontal cable is meant to be terminated once and then left alone; a patch panel gives each cable a fixed, labeled termination point, and all moves are done with flexible patch cords. A change then only means swapping a cord, a faulty port or switch can be replaced without touching the building's cabling, every permanent link can be tested on its own, and the rack stays tidy. A common layout puts a 1U cable manager between each 24-port patch panel and its switch, joined by short patch cords (typically 0.3 to 1 m). KDT's Cuprium range includes 24-port 1U metal patch panels (NP-24U for UTP keystones, NP-24UD for UTP and shielded keystones, and NP-24L with front-access keystone installation) and 1U cable managers (CM-1U with a cover and CM-1UB with a brush).

TIA-606 says to give every cable, patch panel port, outlet, rack and pathway a unique identifier and to write the same identifier on both ends of the cable, on the outlet and patch panel port, and in the documentation. A simple, common pattern is floor-room-panel-port; for example 2A-B07 means floor 2, rack room A, patch panel B, port 7. Labels must be legible and durable (printed cable labels, not marker pen), placed near the end of the cable and still visible after installation. Keep a table or an as-built drawing (As-Built) with each cable's identifier, length, both ends and test result; TIA-606 defines four classes of administration, from a single rack room up to several buildings or sites.

The first IP digit (6) means completely dust-tight and the second is the water rating: IP65 is tested against water jets from a 6.3 mm nozzle (about 12.5 liters per minute) and IP66 against powerful jets from a 12.5 mm nozzle (about 100 liters per minute) (IEC 60529). For most urban wall installations IP65 is enough; IP66 suits industrial sites, hose-down cleaning or heavy wind-driven rain. The IP rating holds only if cables enter from below through cable glands that fit the cable diameter and unused holes are plugged, and a fan without a proper filter and cover lowers the rating. KDT's Cuprium range includes wall-mount outdoor enclosures KR-40X, KR-80X, KR-160X and KR-240X rated IP65, and double-wall enclosures KR-0802FDW and KR-7504DW-DRWP rated IP66 with provision for a 9×9 cm fan.

Connect the metal body of the rack, shielded patch panels and equipment with a green-yellow copper wire, as short as possible, to a ground busbar in the rack room, and connect that busbar with a separate conductor to the building's main grounding terminal; this bonding structure is described in TIA-607, EN 50310 and ISO/IEC 30129. Grounding means equalizing potential: a separate ground for cameras or the network that is not connected to the building ground is not compliant and is dangerous, because during lightning or a fault a potential difference appears between the two systems and damages equipment, so all ground electrodes must be bonded together. Never use the neutral wire as a ground. Measure the electrode resistance with a ground-resistance tester; a common target is 5 Ω or less, and IEC 62305 recommends less than 10 Ω for a lightning-protection electrode.

Lightning does not need to hit the camera directly: a nearby strike induces voltage in long power and network cables, while an Ethernet port has only about 1.5 kV of isolation (an IEEE 802.3 requirement), and a copper cable between two points with different ground potentials also becomes a path for current. Protection has several layers: a low-resistance, bonded ground for the mast, camera housing and rack; network and PoE surge protective devices (SPD) at both ends of the cable and a power SPD (Type 2) in the panel that feeds the camera and switch; fiber instead of copper on long outdoor runs and between buildings; and a lightning rod if the mast is the highest point, because an SPD cannot stop a direct strike. Route cables without loops and away from the down conductors of a lightning rod.

A network SPD must be rated for the cable category and link speed (for example Cat6 and 1 Gbps) and for PoE (passing 48 to 57 V DC and the required power, and for 802.3bt a four-pair version), and it should be tested to IEC 61643-21; an SPD without PoE support will not work on a PoE link or may be destroyed by it. Install one at each end of the outdoor cable, near the camera and at the switch or rack the cable reaches, and connect both with a short, thick ground wire (preferably 0.5 m or less in total) to a bonded ground; a long or thin ground wire is almost useless. The power SPD and the network SPD must share a common ground so that no potential difference is left between them. An SPD wears out after absorbing a large surge, so check its status indicator after every storm and replace it if needed.

VA (volt-amperes) is apparent power and W (watts) is real power: W = VA × power factor (PF). In a small inexpensive UPS the PF is usually about 0.6 to 0.7, so 1000 VA gives only 600 to 700 W, while in an online UPS and newer models the PF is about 0.9 to 1. Add up the mains power of all the equipment in watts and choose a UPS whose watt rating is such that this total is at most 70 to 80% of it (for growth, the spin-up current of disks and battery aging); check the VA too. Example: an NVR at 60 W, an eight-port PoE switch with a 64 W PoE load and 10 W for the switch itself (74 W, which at 90% power-supply efficiency draws about 82 W from the mains) and a modem and router at 15 W come to about 157 W; 157 ÷ 0.7 ≈ 225 W, so a 600 VA UPS rated 360 W is enough, but a 1000 VA unit is better for a longer runtime.

A simple estimate: runtime (hours) ≈ (battery voltage × Ah × 0.6) ÷ load (W); the factor 0.6 covers inverter efficiency, usable depth of discharge and capacity loss with age. Example: two 12V 9Ah batteries in series (24 V) hold 24 × 9 = 216 Wh, so for a 150 W load the runtime is about 216 × 0.6 ÷ 150 ≈ 0.86 hours, or roughly 50 minutes. It is not linear: half the load usually gives more than double the runtime, so use the manufacturer's runtime chart for exact figures. VRLA batteries usually fall to 80% of their capacity after 3 to 5 years (sooner in a hot room) and must then be replaced; for longer runtime use an external battery pack made for the same UPS, or a larger UPS.

Yes. PoE current produces heat in the resistance of the conductors (P = I² × R); in the worst case allowed by the standards, about 4.5 W is lost in 100 m of cable for 802.3at (30 W), 9 W for 802.3bt Type 3 (60 W) and 18.7 W for 802.3bt Type 4 (90 W). In a large, tightly packed bundle the heat cannot escape and the cables in the middle run hotter, and cable attenuation rises with temperature (about 0.4% per °C above 20 °C for UTP and about half of that for shielded cable), so the link stays reliable over a shorter length. TIA's TSB-184-A and ISO/IEC TS 29125 cover this; in practice, use Cat6A or cable with thicker conductors for high-power PoE, split large bundles into smaller groups (for example at most 24 cables), lay the cables loosely on open trays, and avoid long runs through hot spaces or thermal insulation.

The usual cause is voltage drop in the cable: when the camera's IR, heater or motor switches on, the current rises, the cable's voltage drop grows, and the voltage at the camera falls below its minimum (about 37 V for 802.3af and 42.5 V for 802.3at). Measure the voltage at the camera end under full load (IR on) with a PoE tester, or with a multimeter and an RJ45 breakout adapter; if it is close to these minimums, the cable is to blame. Typical causes are a long run, conductors thinner than 24 AWG, CCA, poor terminations with high contact resistance, a poor-quality patch cord or an extender in the path. Temporarily connect the camera to the switch with a short, known-good cable; if the problem disappears, re-terminate the long run or install better cable, and if it still restarts even with the short cable, check the PoE power budget of the switch.

Fiber, almost always. Copper is limited to 100 m and has two more weaknesses between buildings: the ground potentials of the two buildings can differ so that current flows along the shield, and lightning or nearby switching induces voltage in a long cable and burns the ports at both ends. Fiber is non-conductive and has neither problem; single-mode fiber with a suitable SFP module covers 10 to 80 km, and multimode OM3 and OM4 fiber is enough for 10 Gbps up to 300 m and 400 m. For the outdoor route choose armored or duct-rated cable and terminate both ends in a fiber patch panel; KDT's Cuprium range includes single-mode fiber cable from 4 to 48 cores (for example the armored KF-12CAO), fiber patch panels, pigtails and SFP modules up to 80 km (for example KG-GLC80SC). Fiber types and link-loss calculation are covered in the fiber optics topic.

A cheap LED tester only checks the wire map: opens, shorts, miswires, in better models split pairs, and sometimes an approximate length; it does not measure transmission quality, so a cable with untwisted pairs, a poor termination or CCA may show as good and then fail under load, for example at 1 Gbps or with PoE, or stay at 100 Mbps. A certification tester measures these parameters over the whole frequency range (up to 100, 250 or 500 MHz for Cat5e, Cat6 and Cat6A) and compares them with the standard's limits: wire map, length, propagation delay and delay skew, DC loop resistance, insertion loss, NEXT and PSNEXT, ACR-F and return loss. The output is a pass/fail report with the margin of each parameter for every link, which can be handed to the client and is required by the specifications of many projects.

The installer tests fixed cabling to the Permanent Link limit: up to 90 m of cable plus the connections at both ends and no patch cords, because it measures only what was built into the wall and trays and does not depend on the patch cords. A Channel test covers the whole path from the device to the switch including the patch cords (up to 100 m) and is used for final checks or troubleshooting, and a link ending in a plug (MPTL) has its own limit. In every test choose the correct limit and cable type (for example TIA Cat6 Permanent Link) and do not stop at the word pass; look at the margin of each parameter too. If NEXT fails, the cause is usually the termination: pairs untwisted too far, a badly punched keystone or plug, an unsuitable patch cord or connector, or a crushed cable; many testers show a NEXT plot with the approximate location of the fault, and the test must be repeated after re-terminating.

KDT's Cuprium passive range includes these groups: Cat6 copper cable (UTP and S/FTP for indoor use, S/FTP for outdoor use) and patch cords; keystones, RJ45 plugs, 24-port metal patch panels, cable managers and blank panels; outdoor and industrial wall-mount enclosures rated IP65 and IP66; 4-port and 8-port rack power strips (PM-04P and PM-08P) and a DIN-rail DC-to-DC converter with battery backup (KU-1948-SPU/D); and, on the fiber side, single-mode cable from 4 to 48 cores, patch panels, pigtails, patch cords and SFP modules. The specifications of each model, including cable type, dimensions and protection rating, are on that product's page in the Network Infrastructure section.