Why Should Ethernet Cable Runs Stay Under 100 Meters?

This is a classic networking question. The 100-meter figure is not arbitrary; it results from the combined effects of physical transmission characteristics and Ethernet protocol mechanisms. Let's break it down from two perspectives: why signals cannot travel indefinitely and why the protocol imposes a distance limit.

Twisted-pair Ethernet cabling transmits data in the form of electrical signals. During transmission, these signals encounter three major challenges.

Signal attenuation

Copper conductors have electrical resistance. As a signal travels along the cable, some of its energy is dissipated as heat due to this resistance. The insulating material between the conductors also introduces dielectric losses. Together, these effects cause the signal amplitude to decrease continuously with distance.

At 100 meters, signal attenuation is typically around 20 dB. Beyond 100 meters, attenuation increases significantly, and the signal-to-noise ratio (SNR) at the receiving end may become insufficient for reliable signal detection, resulting in bit errors.

Crosstalk

An Ethernet cable contains four twisted pairs. When signals travel through each pair, they generate electromagnetic fields. According to Maxwell's equations, a changing electric current produces a magnetic field, and a changing magnetic field can induce current in adjacent pairs.

As the transmission distance increases, the cumulative effects of this coupling can become more significant. Higher-frequency signals are particularly susceptible to interference.

Signal distortion and propagation delay

The cable has distributed capacitance and inductance, which produce a low-pass filtering effect on the signal waveform. Higher-frequency components attenuate more rapidly, distorting the waveform and making square-wave edges appear rounded.

Signals also require time to propagate through the cable. Typical Category 5 UTP cable has a propagation delay of approximately 5.56 ns/m, resulting in a one-way delay of about 556 ns over 100 meters.

Physical signal attenuation determines whether a signal can be transmitted reliably, but the Ethernet protocol's CSMA/CD (Carrier Sense Multiple Access with Collision Detection) mechanism is the direct basis for deriving the 100-meter figure.

Ethernet originally used a shared-medium mechanism. When two devices transmitted data simultaneously, their signals could collide on the transmission medium.

To prevent collisions from causing undetected data loss, Ethernet requires a transmitting device to be able to detect a collision before it finishes transmitting the minimum-length frame. Otherwise, the sender may assume that transmission was successful even though the receiving end has received a corrupted frame.

The receiver discards the frame because it fails the error check, while the retransmission mechanism is not triggered because the sender did not detect the collision.

Step 1: Calculate the Bit Time

Fast Ethernet (100BASE-TX) operates at 100 Mbps.

Bit time = 1 / 100 Mbps = 10 ns

Step 2: Calculate the Minimum Frame Transmission Time

The minimum Ethernet frame size is 64 bytes, or 512 bits.

Minimum frame transmission time = 512 × 10 ns = 5,120 ns

This 5,120 ns represents the maximum allowable round-trip collision delay.

Step 3: Calculate the Maximum Transmission Distance

The propagation delay of Category 5 UTP cable is 5.56 ns/m. Therefore:

Collision-domain round-trip distance = 5,120 ns / 5.56 ns/m ≈ 921 m

Maximum length of a single cable segment = collision-domain round-trip distance / 2 ≈ 460 m

Step 4: Apply the 5-4-3-2-1 Repeater Rule

The Ethernet repeater rule, known as the 5-4-3-2-1 rule, specifies a maximum of five segments and four repeaters in a single collision domain, with additional engineering margin required.

After accounting for these factors, the maximum length of a single twisted-pair cable segment is ultimately standardized at 100 meters.

When the transmission distance exceeds 100 meters:

• The signal propagation delay exceeds the 5,120 ns round-trip collision-delay limit.
• The sender may be unable to detect a collision before finishing transmission of the minimum-length frame.
• A frame corrupted by a collision may finish transmitting and reach the receiving end.
• The receiver discards the frame because it fails the error check, but the retransmission mechanism is not triggered.

The result: Frames may be silently lost, causing network slowdowns, disconnections, or transmission interruptions.

You might wonder: Gigabit Ethernet (1000BASE-T) operates at ten times the speed of Fast Ethernet, giving it a bit time of just 1 ns. Based on the calculation above, shouldn't the maximum distance be much shorter?

In practice, Gigabit Ethernet and faster Ethernet technologies commonly use full-duplex switched networks. Each port has dedicated bandwidth, so collisions between multiple devices sharing the same medium are no longer an issue, and collision detection is no longer the primary constraint.

However, the 100-meter limit remains because signal attenuation and waveform distortion become the dominant concerns. At 1 Gbps and higher speeds, signals operate under more demanding signal-integrity requirements, making 100 meters the safe upper limit for maintaining signal quality.

Structured cabling standards, such as TIA/EIA-568 and ISO/IEC 11801, specify the following requirements:

• Permanent horizontal cabling: No more than 90 meters.
• Total channel length, including patch cords: No more than 100 meters.

In practical installations, factors such as connector termination losses, cable bends, and ambient temperature must also be considered. It is recommended to keep cable runs within 80–90 meters.

Standard copper Ethernet cable with 0.5 mm conductors may theoretically reach approximately 112 meters, but the standardized limit is set at 100 meters to provide an adequate margin.

Low-quality cables, such as copper-clad iron or copper-clad steel, may begin to experience unstable transmission at distances of just 60–70 meters.

When Ethernet cables are routed in the same conduit as power cables, strong electromagnetic interference may reduce the effective transmission distance by 20–30%.

When the transmission rate is reduced from 100 Mbps to 10 Mbps, the bit time increases to 100 ns, and the minimum-frame transmission time increases to 51,200 ns. This significantly expands the allowable collision-domain diameter, potentially extending the practical transmission distance to 150–200 meters, depending on cable quality.

However, this does not comply with the standard and should only be considered for temporary emergency situations.

Switch cascading is the most commonly used solution. Adding a switch in the middle of the link effectively divides one long connection into multiple shorter segments, each subject to its own 100-meter limit.

The 100-meter limit for Ethernet cabling results from the combined constraints of physical signal attenuation and protocol timing requirements.

At the physical layer, resistance, capacitance, and crosstalk cause signals to attenuate and distort as they travel along the cable. Beyond 100 meters, the receiving end may no longer be able to identify the signal reliably.

At the protocol layer, Ethernet's CSMA/CD collision-detection mechanism requires the round-trip delay to remain within the 5,120 ns timing limit associated with the minimum 64-byte Ethernet frame. Combined with the 5.56 ns/m propagation delay of Category 5 UTP cable and the 5-4-3-2-1 repeater rule, this leads to a maximum distance of 100 meters for a single cable segment.

Although Gigabit Ethernet and faster Ethernet technologies commonly use full-duplex switched networks, eliminating collision detection as the primary constraint, their more demanding signal-integrity requirements mean that the 100-meter limit remains an important boundary for reliable transmission.

For practical installations, it is recommended to keep cable runs within 80–90 meters. For longer distances, switch cascading or fiber-optic conversion can provide an effective solution.

Remember: The 100-meter limit was not simply imposed by regulation; it was calculated. It reflects both the physical limitations of copper cabling and the careful balance built into Ethernet protocol design.