4 Evaluation Simulations

To evaluate the effects of different configurations, placement of components and component values, various schematic variants with the three coupling filters presented earlier were tested using simulations. The setup for the different testing scenarios were derived from previously available system level templates used for data evaluation. The Keysight ADS software was used to perform these simulations. The power supply coupling filters were added to each configuration and the properties and characteristics of the inductor arrays, weakly coupled inductors, and DMC were taken from the respective manufacturers.

For each of these filters a schematic was created and simulated. Not only was the type of filter tested, but the values of the inductance was also varied to see which values lead to the best performance. As a reference, the system was also simulated without using PoDL to see what effects the power transmission has on the signal quality. The results are visualized using the eye diagram of each node of the simulation and the corresponding alien noise and jitter tolerance plot.

Since there were many simulations done during the initial phase, only the most relevant results will be discussed here. An example using only two nodes on Topology 1 and an example using five nodes on Topology 3 will be shown. The example on Topology 3 will be used as a worst case scenario, the resulting eye diagram and jitter plots will be discussed.

Figure 4-1 below shows the results from a simulation consisting of two nodes and a power supply. The coupling between the devices and the network is realized using single, uncoupled inductors with a nominal value of 47 µH on both the positive and negative wires of the transmission line. The resulting eye diagram and jitter plot show that a large noise margin is left for the system to utilize.

Figure 4-1. Simulation using two 47 µH single inductors on Topology 1
: The noise and jitter plot contains the same information as the eye diagram. This plot is the superposition of each quadrant in the eye diagram normalized to show a single edge. This can also be achieved by splitting both the X- and Y‑Axis into two separate parts and rotating the result to show a rising edge. The alien noise and jitter tolerance plot can be used to quickly see the worst case scenario for all participating nodes in the measurements. This is the lowest line in the plot, highlighted with red. An offset of 60 mV is subtracted from each measurement to account for the ±30 mV signal detection threshold of the device. The y-intercept can be interpreted as the amount of jitter that is created by the system itself. This is the jitter in the system if the is no alien noise present. Since the maximum allowed jitter is 15 ns, the value of the red line at a jitter value of 15 ns indicates how much noise tolerance the system has in total. This is the amount of noise that the system may be subjected to before communications fail. If this value is below zero, it means that the system will contain at least one path failing to communicate correctly, even without external noise applied.

Figure 4-2 below shows a simulation of five nodes and a separate power supply. Here, each node uses a coupling to the transmission line consisting of two 68 µH inductors, one for the positive and one for the negative wire. The resulting eye diagram and jitter plot are shown.

Figure 4-2. Simulation using 68 µH single Inductors on Topology 3

When comparing the two results, it is obvious that the five node configuration is inherently much more sensitive to external noise than the two node configuration. This is because the topology of the system itself causes noise on the network, reducing the overall noise threshold at which communication between nodes will be disrupted. However, the noise performance is still within acceptable bounds because there is a large margin before additional noise becomes an issue, meaning that every node is able to both receive power over the transmission line and still communicate successfully with each other node in the system.

Figure 4-3 below shows the topology used to test an additional configuration with eight nodes. In this configuration the PSE was moved from the end of the cable to a location between the fourth and fifth nodes. The filter between the power supply on each device and the network was realized using a 47 µH coupled inductor on each device.

Figure 4-3. 8 Node Topology

Figure 4-4 below shows the results of the simulation of the eight node topology. The overall noise tolerance of this system is the lowest so far, but for environments with lower alien noise it is still possible to implement this system successfully.

Figure 4-4. 8 Node Topology Simulation