Characterization of High Speed Bus

Once the PCB has been designed and tested, the high speed signals should undergo characterization. Characterization of high speed bus involves taking measurements related to the high speed signals. Characterization will prove the robustness of the design. The purpose of characterization is

• To ensure that the simulation results are in accordance with the observed results
• To ensure that the signal observed are within the specification.
• To find the voltage and current margin.
• To measure any path loss for critical signal

Correlating Simulation Result with Characterization Results

One of the first tasks a signal integrity engineer is expected to do after the board first boots up – is to take scope shots of the signals that were simulated. The signal should be captured using XY data. The SI engineer then compares the characterization results with the simulation results. A highly correlated waveform validates the SI engineer’s simulation methodology as well as the fact that the design rules were followed in the PCB design. A mismatch in the simulated and observed result will trigger need for further investigation. It can indicate wrong methodology followed in simulation, a wrong simulation model, or an error in the circuit or the PCB design. Efforts must be made to seek the truth behind the discrepancy. Efforts must be made to streamline the process where the simulated and observed waveforms match.

We must ensure that the signal observed is within specification - This will require understanding of the electrical specification of the system. It is possible to check the set up and hold time, overshoot, amplitude and any other specification using high speed oscilloscope.

While taking the scope shot for correlation, we must also take into account the bandwidth limitation of the oscilloscope and the probe. If the oscilloscope manufacturer can provide the model of the probe, integrate it into your simulation. Most often the oscilloscope scope shots are taken at some distance away from the silicon receiver pads. The oscilloscope scope shot is taken on the motherboard which is at some distance away from the BGA receiver pads of the IC. We should simulate the waveform at the test point in addition to the one at the receiver. The waveform simulated at the test point should be compared with the observed voltage at the receiver test point. The distance between the test point and the receiver silicon pad consists of a small length of the PCB trace, the BGA pin parasitics and the package traces.

Previous                    Next