Well-balanced differential signals radiate less than single-ended signals do. That's one of the benefits of differential signaling. If the two complementary signals of a differential pair are perfectly balanced, a tight separation between the signal conductors can achieve an astonishing degree of field cancellation.
If, however, the two complementary signals of a differential pair are not perfectly balanced, then the degree of attainable field cancellation is limited to a minimum value determined not by the inter-conductor spacing, but by the common-mode balance of the differential pair. Because the common-mode balance of most digital drivers is not particularly good, differential pairs often radiate far more power in the common mode than in the differential mode. In such a situation, you gain no radiation benefit from squeezing the differential traces more closely together.
Figure 1 plots the theoretical radiation improvement (marked as gain) attained by a differential microstrip pair as a function of trace separation. The figure assumes that the measurement antenna resides in the board's plane, removed in a broadside direction a distance r=10m away from the traces (worst case). The equal and opposite radiation from the adjacent trace is supposed to cancel the radiation from one trace of the differential pair, resulting in a marked reduction in emissions. The depth of cancellation relates to the ratio 2πs/λ, where λ is the free-space wavelength of the highest frequency of interest and s is the separation between traces. This ratio controls the relative phase relationship of the two near-complementary waves as they leave your board. The cancellation also relates to the ratio r/r+s, which addresses the relative intensities of the two near-complementary waves as they reach the antenna. The formula in Figure 1 shows that differential cancellation improves as s decreases.
The common-mode radiation from the two traces of a differential microstrip reinforces, rather than cancels, the net result. Therefore, the common-mode radiation does not vary strongly with trace separation. You can adjust the differential-mode radiation by adjusting the trace spacing, but you can't do much about the common-mode radiation, except to install a driver with better common-mode balance.
Under FCC class B measurement conditions, the differential-mode radiation from a differential-microstrip pair with 0.5-mm separation should theoretically yield a 40-dB radiation improvement at 1 GHz over the radiation you would measure if you implemented the same signal as a single-ended layout. Smaller separations should yield even more improvement. Although that theory sounds appealing, in practice, you will rarely-if ever-achieve as much as a 40-dB improvement in overall radiation because the degree of balance available on the two outputs of your differential transmitter will limit your gains. Unless the outputs balance to better than one part in 100, a common-mode-radiation component of at least 1% of the differential amplitude will emanate from your differential pair anyway. Therefore, even with a differential spacing of zero, you could never improve the total radiation by more than 40 dB.
In plain terms, a differential-trace spacing of 0.5 mm is close enough to deliver all the EMI benefit you are ever likely to get. Because the common-mode radiation usually dominates radiation problems on digital pc boards anyway, you need not struggle to place ordinary differential digital traces any closer than 0.5 mm for any EMI purpose.
EDN takes this excerpt from the forthcoming Prentice Hall publication,High-Speed Signal Propagation: More Black Magic , by Howard Johnson, ISBN 013084408X, February 2003. Adapted by permission of Pearson Education Inc, Upper Saddle River, NJ.