Gain structure and noise performance in active microphone splitter systems
The use of an active microphone splitter system in place of the traditional passive transformer-based splitter provides a number of clear advantages. These include easier control over microphone powering, headphone monitoring facilities, and metering, in addition to the fundamental advantage of improved line drive capability. The combination of low output impedance and higher signal level mean that an active splitter is potentially capable of quieter performance and better noise immunity than a passive one, as well as minimizing high-frequency losses due to cable capacitance. However, if these benefits are to be realized in practice, it is necessary to set up the complete system (including both the splitter and the console) with the correct gain structure. Failure to do this may result in the system actually performing worse when compared with a simple passive splitter, so it is well worth spending a few minutes to get familiar with the concepts involved. The reason that it matters at all is that amplifiers are not perfect. All active electronics add a small amount of noise to the signal - for example a typical well-designed amplifier will have residual noise at around -100dBu on its output, irrespective of any input signal.
Figure 1 shows a microphone connected directly to a console. In this simple case we bring the microphone signal into the console, and immediately amplify it in the first active stage - the microphone amplifier. To take a practical example, a common dynamic vocal microphone subjected to an SPL of 110dB will produce an output of approximately -33dBu. In order to bring this up to a usable level in the console, we will set the microphone amplifier to +33 dB of gain, resulting in a 0dBu signal leaving the amplifier. To this will be added the noise of the amplifier, but since this is at around -100dBu on the amplifier output, we still have a signal-to-noise ratio of around 100dB.
Figure 2 shows the same signal connected using an active splitter system. The splitter contains a variable-gain microphone amplifier, which then feeds a number of independent output amplifiers. One of these is then connected to the console input, which itself has a variable-gain microphone amplifier. The crux of the matter is how best to set the gain of the two microphone amplifiers. It is tempting to simply set the splitter to unity gain, and insert it in the signal path expecting nothing to change - after all, this is what we would do with a passive splitter. However, we can immediately see a problem with this approach. We bring our microphone signal at the same level of -33dBu into the splitter, but now instead of it immediately hitting an amplifier with gain, it is simply passed at the same level through the splitter. The splitter's microphone amplifier and the line driver will each add noise at about -100dBu to this signal, just as the console's microphone amplifier did. Note that because the signal is still at -33dBu, the signal-to-noise ratio at point A is now only 67dB. This signal arrives at the microphone amplifier in the console, and we boost the whole thing by +33dB. This restores the signal to 0dBu as desired, but also brings up the splitter's output noise by 33dB - so we still have a signal-to-noise of only 67dB. The additional noise at -100 from the console's microphone amplifier is of no real consequence in this case.
In order to restore the performance of our system and to actually benefit from the improved line driving ability of the splitter, what we should have done is to use the microphone amplifier on the splitter. If we set the splitter's microphone amplifier to +33dB of gain, then the noise contribution of that amplifier (at -100dBu) will now be added to a signal with a level of 0dBu, instead of -33dBu. This will preserve our 100dB signal-to-noise ratio in the splitter, instead of reducing it to 67dB. The console input section is now set to 0dB of gain, so there is no increase in the splitter's noise contribution as a result, and we merely add the console noise at -100dBu to our signal.
So, the conclusion that we reach is:
When using any active splitter system, as much gain as possible should be added using the splitter's microphone amplifier, and as little as possible using the console.
Obviously the limit of this approach will be the point at which the splitter's output will clip on loud sounds. It is worth noting, however, that with the popular dynamic vocal microphone used for this example, and +30dB of gain on the splitter's microphone amplifier, that it would require an SPL of 133.3dB to produce an output of +20dBu from the splitter - still within the output capability of most professional equipment.