PI and BUI Results

Meas. mode:
Jitter
Waveform type:
NRZ
PAM4
NRZ Amplitude Table
PAM Level Table

The Jitter mode Periodic Interference (PI) and Bounded Uncorrelated Interference (BUI) measurements are one component of Deterministic Interference (DI). In the results table, the following measurements are listed:

  • PI (rms) and PI (δ - δ) for all levels, or
  • BUI (rms) and BUI (δ - δ) for all levels

Whether PI or BUI is reported depends on the Separation Method setting in the Advanced tab of the Jitter Mode Measurements Setup dialog. The Separation Method setting (either Spectral or Tail Fit) determines the method used to decompose RJ and RN.

PI is a measure of the interference that is uncorrelated to the pattern, yet is periodic. Interference caused by a switching power supply is an example of periodic interference. PI (rms) is useful in the case where the PI is of a known distribution. For example, a stress test may be developed in which sinusoidal PI is intentionally injected. Assuming the injected PI is the only PI present, the system can be calibrated using the PI (rms) measurement. Dual-Dirac PI (δ - δ) is useful when the distribution is unknown. The dual-Dirac number allows various PI measurements to be compared with one another, even when the distributions are dissimilar. It is a measure of how much the PI affects the low-probability statistics of the uncorrelated interference.

BUI includes PI, but it also includes other bounded uncorrelated effects that don’t cause clear peaks in the interference spectrum. One common source with these characteristics is cross talk from adjacent traces on printed circuit boards or coupling within the chips themselves. The Spectral method of noise separation will see cross talk as noise in the interference spectrum, so it will only separate the PI peaks and attribute the other effects to RN. The Tail Fit method uses the low probability tails of the uncorrelated interference histogram to measure in a region that BUI does not contribute. This method is less accurate due to the use of a significantly fewer samples, but it allows for the separation of RN from BUI.

The δ - δ measurement algorithm is:

  1. An uncorrelated interference histogram is constructed.
  2. Random Noise (RN) is measured.
  3. Using the measured value of RN, a dual-Dirac model is fit to the histogram to obtain the PI or BUI measurement. The model is fit at a probability of 10-3 when using the Spectral separation method or an adjustable probability based on number of acquired samples when using Tail Fit.

PI or BUI (δ - δ) = UN(prob) – 2 * Q(prob) * RNrms

The RMS measurement algorithm is:

  1. An uncorrelated interference histogram is constructed.
  2. Random Noise (RN) is measured.
  3. The variance of the histogram is calculated.
  4. PI or BUI (rms) is calculated as follows:

PIrms or BUIrms = sqrt(Total Variance – RNrms2)

PI and BUI are displayed in the current vertical channel units or unit amplitude. Make the selection in the Amplitude Measurements tab of the Configure Jitter Measurements dialog.

The definition of a one or zero level can be based on an average level or on a minimum of Consecutive Identical Digits (CID). Measurements based on CID minimize the effects of ISI on the levels. Make the selection in the Amplitude Measurements tab of the Jitter Mode Measurements Setup dialog.

More Information

SCPI Command

:MEASure:AMPLitude:PI (δ - δ)

:MEASure:AMPLitude:PIR (rms)