Differential Signals

Noise reduction on differential signals requires two copies of the difference signal. The best way to do this is to split the two polarities of the signal into two copies is by using passive power splitters or dividers (as shown in the following figure). This method requires four oscilloscope channels. You then configure the four individual channels as two differential channels from within the Channel Dialog Box.

This connection method does not add noise or jitter to the measurement because it uses passive devices. In addition, the subtraction of two digitized signals further reduces the contribution of the oscilloscope channel's voltage noise an additional 3 dB.

Another effective method of duplicating a differential signal is to first convert the two polarities of the differential signal into a single difference signal using a balun transformer, and then split the difference signal into two copies using a power splitter/divider (see the following figure). Again, using passive devices does not add noise or jitter to the measurement, but physical balun transformers do introduce errors that the subtraction of oscilloscope channels does not. It is important to use a balun that has adequate insertion loss and common-mode rejection over the full bandwidth of the measurement. PCI Express clock jitter measurements, for example, may only measure a 100 MHz clock, but they require a full measurement bandwidth of 5 GHz.

You can also use buffers or amplifiers to fan-out your differential Signal Under Test (SUT) if the buffer or amplifier does not add appreciable jitter of its own to the measurement. Differential buffers used with differential measurements offer a little more benefit than they do for single-ended measurements because they also perform the difference operation.