4.15 Topology Discovery - Rev D0

This device supports Topology Discovery according to the OPEN Alliance 10BASE‑T1S Topology Discovery Specification. Topology Discovery is a method to measure the signal propagation time between two nodes on the same 10BASE‑T1S segment and, therefore, the relative distances and order of the nodes on the segment. If the speed of propagation through the medium is known, typically about 5 ns/m, the distance between the two nodes can be calculated.

Topology Discovery a system diagnostic, meaning that before the measurement can begin, a reference and a measured node must be determined and coordinated. During this time, all other nodes must avoid all transmissions. This includes disabling of PLCA to prevent transmission of BEACON symbols from interfering with the measurement. Furthermore, it is recommended to place the reference node on the physical end of the transmission line. This way, each measured node will have a unique signal propagation time relative to the reference node.

Topology Discovery works by the reference node transmitting a single pulse onto the network. The measured node detects this pulse and transmits a pulse in return. The reference node receives this returned pulse and transmits another pulse. This cycle of “bouncing” a pulse back and forth between the reference and measured nodes is repeated for a fixed, known time interval. The distance between the devices may then be calculated using the duration of the measurement, the number of pulses received and the internal delays of both nodes.

The polarity of the transmitted pulse is randomly varied each transmission cycle through the use of a scrambler. This is done to provide rejection of noise and To prevent a baseline wander on the line, a 1B/2B coding is applied to the output. This means that for every positive pulse sent, a negative pulse is sent in the cycle afterward and vice-versa.

Upon reception, the 2B pulse pair is internally deserialized into a single 1B pulse. A descrambler is used verify that the polarity of the incoming 1B pulse is as expected and correct. Prior to beginning an actual measurement, both descramblers must be trained by the transmissions from the other node. The first 60 pulses are therefore reserved for this synchronization and not counted as part of the measurement. If the scramblers and descramblers cannot be synchronized within this time, or if the polarity of an incoming pulse does not match the predicted polarity due to noise or error, then the measurement will result in a failure.

To activate Topology Discovery, a ‘1’ must be written into the Topology Discovery Enable (TDEN) bit in the Topology Discovery Control (TDCTL) register. If a ‘1’ is written into the Reference Node Select (REFN) bit in the TDCTL register, then the node is selected as the reference node. Otherwise, it will be treated as a measured node.

Topology discovery is performed in two steps. Only the final step determines number of pulses exchanged between the reference and measured nodes as previously described. First, however, the internal delay of the device must be measured. By measuring the internal delay, any variability in timing due to silicon process, temperature, and voltage may be removed increasing the accuracy of the measurement of the signal propagation time between the two nodes. The internal delay of a node is the delay between the first edge of a receiving pulse and the first edge of the corresponding pulse being transmitted by that same node. To measure the internal delay of a node, all other nodes must be configured so that they do not transmit anything onto the network including PLCA BEACONs. The node measuring its internal delay is then permitted to begin its delay measurement by enabling topology discovery and writing a ‘1’ into the Internal Delay Measurement Start (INTDLYSTRT) bit in the TDCTL register. Because only this one node is transmitting onto the network, the other nodes are able to listen in while this node’s internal delay is measured. Similar to the distance measurement, the device sends pulses onto the network. However, instead of waiting for a response from another node, it listens to its own transmitted pulses to trigger the follow-up pulse. This is repeated for a fixed amount of time as configured in the Distance Measurement Duration (DISTMESDUR) field of the TDCTL register and the number of pulses is counted. The internal delay is calculated by dividing the duration of the measurement by the number of pulses counted.

The resulting internal delay count is contained in the Topology Discovery Internal Node Delay Result -Low (TDINTDLYRESL) and -High (TDINTDLYRESH) registers. In manual mode, this internal delay measurement must be communicated from the measured node back to the reference node in its Topology Discovery Measured Node Delay Measurement Result -Low (TDMNDLYRESL) and -High (TDMNDLYRESH) registers for it to calculate the signal propagation time between the two nodes. However, in automatic mode, the reference node acquires values in these registers by listening in while the measured node performs its internal delay count. In this case, the internal delay count of the measured node as observed by the reference node is automatically stored in the TDMNDLYRESL and TDMNDLYRESH registers.

The signal propagation time measurement is initiated with the reference node transmitting a single pulse onto the network. The measured node detects this pulse and transmits a return pulse. When the reference node receives this returned pulse it will respond by transmitting another pulse. This cycle is repeated for a fixed time interval defined in the Distance Measurement Duration (DISTMESDUR) field of the TDCTL register. When the Distance Measurement Done (DISTMESDN) bit in the Topology Discovery Status (TDSTS) register is set, the signal propagation time measurement phase has completed and the number of pulses received and counted by the reference node is stored in the Topology Discovery Distance Measurement Result -Low (TDDISTRESL) and -High (TDDISTRESH) registers. The signal propagation time between the two nodes may then be calculated using the duration of the measurement, the number of pulses received and the internal delays of both nodes. The resulting signal propagation time may be divided by the speed of propagation through the medium, if known, to derive the physical distance between the nodes to within a number of centimeters.

Additional Topology Discovery details may be found in the Topology Discovery for 10BASE-T1S Systems application note.