3 Inherent Noise in Mixing Segments

A 10BASE-T1S mixing segment normally contains more than 2 nodes and may include stubs; this results in more discontinuities than a point-to-point connection. Each discontinuity will contribute additional reflections and each segment will further attenuate the signal. This distortion of the signal caused by the mixing segment, is present even in the unstressed system, and will reduce the noise margin compared to a point-to-point network, as shown by the additional block in Figure 3-1.

In a multi-drop system, in addition to the components and cabling, signal quality is also affected by the topology of the system. Each node, whether the node is the intended receiver or not, introduces a slight impedance mismatch, and therefore contributes a reflection of every signal that arrives. As shown in Figure 3-2, reflections cause their own reflections. Attenuation reduces the amplitude of signals. For a received signal, every node sees a different combination of reflections, and the combination depends on which node is transmitting. A topology in which many reflections occur very close to each other can result in a system in which some nodes never receive a good signal even though other nodes may work correctly.

The signal degradation that results from the combination of node hardware, cable characteristics and topology itself can be called Inherent Noise. For any pair of nodes, inherent noise will result in a less than ideal eye diagram. With a poor choice of cable and/or topology, inherent noise can be large enough to close the eye completely so that a receiver cannot recover the signal. Unlike other noise sources, inherent noise is fixed, and it is possible to use simulation or lab measurements to determine if a topology results in poor signal quality. An example of the eye pattern measurement showing the effects of inherent noise on a system with several nodes is shown in Figure 3-3