The data backbone behind every lighting control system, camera and BAS sensor. The standard is more specific than the market thinks, and more generous where everyone over-specifies. Here is what ANSI/TIA-568 and ISO/IEC 11801 actually say, and what a 50 degree ceiling void does to all of it.
Four ideas to carry into the next specification conversation you have.
A Category is a frequency range in MHz, not a data rate: Cat 5e is 1 to 100 MHz, Cat 6 is 1 to 250 MHz, Cat 6A is 1 to 500 MHz. And Cat 6 at 10 Gbps is 37 m to 55 m depending on the alien crosstalk environment: 37 m is what you can promise. There is also no TIA Category 7, and there never was.
The strongest case for screened cable in the Gulf is thermal, not electrical. ISO/IEC 11801 de-rates horizontal length above 20 °C at 0.2%/°C for screened but 0.4 to 0.6%/°C for unscreened. At 50 °C a 90 m run becomes 77.4 m unscreened and 84.6 m screened. Screened bundles also measure about 29% cooler under full PoE.
100 m is a 102 m insertion-loss budget, not a length. The cord allowance is (102 minus the link length) divided by a de-rating factor: 1.2 for 24 AWG, 1.5 for 26 AWG, 1.95 for 28 AWG. Beside a 90 m link that is 10 m of 24 AWG cord but only 6.2 m of 28 AWG, which is also hard-capped at 15 m by the 25 ohm DC loop resistance limit.
Category does not gate PoE. ANSI/TIA-568.0-E Table 5 lists IEEE 802.3bt Types 1, 2, 3 and 4 against category 5e, 6, 6A and 8, all at 100 m, and IEEE's own requirement is Class D or Category 5 or better with channel DC loop resistance at or under 25 ohms. What a higher category buys is thermal margin and efficiency, not permission.
Per ANSI/TIA-568.0-E and 568.2-E (October 2024) and ISO/IEC 11801. Note that 568.2-E extended DC resistance unbalance limits down to Cat 5e, so product certified to 568.2-D may not comply with the current revision.
| Category | Bandwidth · max length | PoE Type |
|---|---|---|
| Cat 5e | 1 Gbps · 100 m | PoE Types 1-4, to 100 m. Category does not gate PoE |
| Cat 6 | 10 Gbps · 37-55 m | Reach depends on the alien crosstalk environment. Promise 37 m |
| Cat 6A | 10 Gbps · 100 m | The practical choice for a PoE-heavy plant, for thermal margin |
| Cat 8 | 25/40 Gbps · 30 m | A data-centre top-of-rack cable by design, not a horizontal one |
| Horizontal run at 50 °C, unscreened | 77.4 m | ISO/IEC 11801 de-rating, down from 90 m at 20 °C |
| Horizontal run at 50 °C, screened | 84.6 m | Screened de-rates at a third of the rate. The Gulf argument |
| Patch cord beside a 90 m link | 10 m at 24 AWG | 6.2 m at 28 AWG, and 28 AWG is capped at 15 m regardless |
| Field-test length threshold | 99 m / 110 m | ANSI/TIA-1152-A allows the limit plus 10% |
| PoE heat per cable, Type 1 to Type 4 | 15x | 7.5x per pair, and Type 4 energises twice the pairs. A first-order model |
| Separation from 3 lighting circuits, Cat 6A U/UTP | 20 mm | EN 50174-2:2018, A = S × P. 0 mm in solid steel conduit, not 300 mm |
Six decisions made at design time that the whole building lives with.
True, and the standards give the opposite lifecycle advice to the usual pitch. ISO/IEC 11801 says the system is "anticipated to have a usable life in excess of 10 years", not 25, and singles out the backbone as the layer to design for the life of the building. The horizontal layer is the one you cannot easily replace, which is a better argument for specifying it well than any warranty is.
Bundle heating is modelled as the square of the current, times the pairs carrying it, so Type 3 to Type 4, 600 mA to 960 mA per pair, is 2.56 times the rise, not 60% more, and Type 1 to Type 4 is 15 times per cable. The square law is the model the IEEE cable-heating work assumed, not a law it proved, so use it as a first-order rule. The whole thermal argument reduces to one line: your budget is 60 °C minus the ambient, because that is the cable's rating. TIA assumes a 45 °C ambient and allows 15 °C of rise; ISO assumes 50 and allows 10. Measured at 1 A per pair, a 61-cable Cat 6A U/UTP bundle already rose 10.0 °C, ISO's whole budget. Two more measured findings: organised bundling runs hotter than random, and a bundle needs four to six hours to reach equilibrium, so a commissioning measurement in the first hour is measuring nothing.
It does, but TIA explicitly bans using it the way most RFPs do: "The MICE tables should not be used as performance requirements individually or in combination for components, links, or channels." MICE describes an environment so you can choose products for it. Worth knowing too that the C1 ordinary-office climatic envelope already runs to +60 °C, so a Saudi ceiling void at 50 °C is not automatically a harsher MICE class. What changes at 50 °C is the length de-rating, not the classification.
And the reason to prefer the permanent link is specific: a channel test never tests the plugs. The channel adapter's reference plane sits after the plug at each end, so field-terminated plugs are invisible to it. TIA also makes permanent link limits deliberately tighter, in writing, "to ensure that permanent links can be extended into compliant channels". Two more things a report must show: PASS* and FAIL* cannot be switched off, and the length threshold is the limit plus 10%, so 99 m for a permanent link and 110 m for a channel.
Fire is where two systems collide: a US plenum cable, the highest NEC fire class, rates Euroclass Eca, second-lowest on the European scale, and its FEP insulation and PVC jacket are both halogenated. "LSZH" is not a fire class at all, and written alone in an RFP it is unenforceable. Saudi Arabia runs both systems: SBC 401 is IEC-based, SBC 201 and 801 are NFPA-based, and floor penetrations need an F rating and a T rating of at least one hour.
ANSI/TIA-568 sets no separation distance at all: its clause is a cross-reference to TIA-569. The normative method is EN 50174-2:2018, where separation is A = S × P. S comes from the cable's segregation class (300 mm for cable of unknown performance, 100 mm for unscreened Cat 6A, 50 mm screened, 10 mm Cat 7A) and P from how many 20 A circuits run alongside (0.2 for one to three). So unscreened Cat 6A beside three lighting circuits needs 20 mm, not 300, and solid steel conduit takes every class to zero. Clause 6.2.2 removes the requirement entirely in an E1 environment for class b, c or d cable where each power circuit sits in an overall sheath carrying no more than 100 A, though not in distributor rooms, and local regulations can still require more. The only lamps with a distance of their own are discharge lamps, at 130 mm.
Every line names a document, a quantity and a unit. None of them can be answered with an adjective, which is the only test a tender clause has to pass.
a. Every link shall be certified to ANSI/TIA-568.2-E permanent link limits with an ANSI/TIA-1152-A field tester; channel results are not acceptance evidence.
b. Design length shall be de-rated for a stated 50 °C ambient using the ISO/IEC 11801 percentage rule, named by edition: 77.4 m unscreened, 84.6 m screened.
c. PoE bundle sizes shall be calculated to TIA TSB-184-A for that ambient, the pairs energised and the mA per pair of the Type installed.
d. Separation from power shall be calculated as A = S × P to EN 50174-2:2018, and fire performance stated as a Euroclass or an NEC listing, not as LSZH.
Line b is the one nobody writes. Every length in a Gulf tender today is a 20 degree number, and a link certified in a mild commissioning week can run past its insertion-loss limit in a 50 degree summer.
Permanent-link test reports on every drop are the right deliverable, and we issue them. But "TIA-certified" does not exist: TIA writes standards and certifies neither companies nor installations. Ask any contractor, including us, for what is verifiable: the tester model and its accuracy level, whether results are to permanent link or channel limits, how any PASS* results were resolved, and the ambient temperature the design assumed.
Practical lighting know-how from NLC Academy. No spam, unsubscribe anytime.
By subscribing you agree to receive NLC emails.
The standards behind every number on this page. Where a document defines a method but sets no limit, that is stated.
An 18-slide explainer, ready to post straight to LinkedIn as an image carousel. Same content, faster reading.
