Structured Cabling Channel Loss Calculator
Insertion loss, crosstalk headroom, propagation delay and skew for a twisted-pair channel or permanent link — assembled from the cable and connector coefficients rather than read off a fixed table, so your own datasheet goes straight in.
Attenuation coefficients — a√f + b·f + c/√f, dB per 100 m
Where these numbers come from
This page is sourced differently from the rest of this site, and it is worth saying so plainly. The ITU recommendations behind the other tools here are free downloads, so those figures were checked line by line against the published text. TIA-568 and ISO 11801 are paid standards. Nothing on this page is quoted from them.
Category 6 coefficients come from public IEEE 802.3 working-group documents, which reproduce the cabling specification for the benefit of the Ethernet standards work. Category 5e and Category 6A are widely published figures that could not be verified against a primary source, and the tool labels them that way rather than quietly presenting all three as equal. Every coefficient is editable: if you have the cable's datasheet, use it. That number beats any preset here, including the verified one.
How the channel is assembled
Cable attenuation per 100 m follows a three-term law, and each term is a different physical loss mechanism:
The √f term is copper loss from skin effect, which dominates. The linear term is dielectric loss in the insulation. The 1/√f term accounts for the shield where one is present and is negligible above a few megahertz. Connector insertion loss follows its own √f law.
A channel is then the sum of its parts: horizontal cable, cords, and each mated connection.
Where the famous 1.05 factor actually comes from
Channel limits are often written as 1.05 times the 100 m cable figure plus the connectors, and the 1.05 is usually presented as a constant handed down from the standard. It is not arbitrary. Stranded cordage loses up to 50 % more per metre than solid horizontal cable, so the standard 90 m + 10 m model is electrically 90 + 10×1.5 = 105 m of cable. This tool builds that up from the cord length you actually have rather than assuming ten metres of it.
Crosstalk power-sums, it does not add
The cable and every connector contribute crosstalk, and they combine on a power basis:
Two equal contributions cost 3 dB, not 6. The figure that decides whether the link carries data is not insertion loss or crosstalk alone but the gap between them, ACR: how far the wanted signal sits above the interference it has to be read against. When ACR reaches zero the link is finished, whatever the length chart says.
Why “worst case” is on every crosstalk figure here
The NEXT and ACR values on this page are built by power-summing the limit values of the cable and each connector. That is the worst crosstalk a channel made entirely of barely-compliant parts could present. Real components beat their limits, often comfortably, so a certifier will measure better than this — sometimes much better.
It also explains something that looks alarming at first. Put a textbook 100 m Category 6 channel in at 250 MHz and the worst-case ACR comes out near zero. Nothing is wrong: that is how the top of a category's band is defined. The component limits are set so that the worst permissible combination of them still just works at the highest specified frequency. Running out of headroom exactly at 250 MHz is the design intent, not a warning. What should concern you is thin ACR well inside the band, which is why the verdict above only raises it there.
Temperature, and why this matters for PoE
Copper insertion loss rises with temperature, around 0.4 %/°C for unscreened constructions and roughly half that for screened ones. A cable in a hot ceiling void at 60 °C carries about 16 % more insertion loss than the same cable certified on a cool day — enough to turn a passing channel into a failing one.
This is where cabling and power meet. A bundle carrying PoE heats itself, and the bigger the bundle and the higher the current, the larger the rise. If you are designing a dense Type 3 or Type 4 installation, the temperature you should be entering here is not the ambient one. Our PoE calculator takes the same temperature on the power side, where it costs you voltage at the far end.
Limitations
This models insertion loss, pair-to-pair NEXT and the resulting ACR, plus delay and skew. It does not cover return loss, ACR-F / ELFEXT, power-sum ACR-F, alien crosstalk between adjacent cables (the parameter that defines Category 6A and the reason it exists), TCL or coupling attenuation. Those matter, and a field certifier measures all of them. Nothing here substitutes for certifying the installed link with an instrument that holds a calibration certificate — this is for designing before you pull cable, and for understanding a result after you have.