The A minus B technique does not work well when the signal of interest is smaller than the common mode signal. The user may not be able to get a stable trigger on the oscilloscope either, and poor high frequency common mode rejection ratios are also phenomena from the A-B technique. Channel gain has to be matched very carefully, otherwise the results are inaccurate. When the signal being measured is much smaller than the common mode component, both the low CMRR and lack of common mode range limit the usefulness of channel A minus channel B. In summary, this technique will only work when the signal being measured is about the same or larger in magnitude than the common mode signal, and the common mode signal is low frequency, e.g. line frequency. If the signal of interest in not time related to the common mode signal, it will not be possible to obtain a stable trigger, as the scope can not trigger on resulting math waveform. To obtain the best CMRR performance, it is necessary to precisely match the relative gain between channels. This can be done at DC by connecting both probes to a signal source such as a square wave generator and adjusting the variable gain control of one of the channels for minimum CM signal in the output. The LF compensation of the probes can likewise be adjusted for optimum low frequency CMRR. Because oscilloscope amplifiers and passive probes are not precisely matched for higher frequency gain (or attenuation), CMRR above a few kHz will be very low.