Aux Fed Subs and the consequences
Aux fed subs is a technique employed to try and help manage the low end of a mix at FOH. It's used to try and keep the low end punchy and remove some of the 'mud' you can find as many instruments and vocals accumulate together in a mix.
However, there can be some dire consequences to using the technique that you should be aware of, and if possible - avoid using the method.
What's your end goal?
Bear with this slight side-track, but it will make sense!
When mixing FOH audio in a live context, I like to set myself the challenge of making it sound "as good" as the studio version, even though I'm working with a live band and all the curveballs that brings with it. Sure, it will never sound as good when you're using an SM58 and the vocal track was recorded with a vintage Neumann, and sure, the musicianship may not be as good and you can't do infinite takes to get the sound just right, but the challenge is there. I may never 100% reach my goal of 'studio quality' when mixing, but if it means I get 90% of the way there, then I'm happy. Alternatively, I could just set myself the goal of 'making it sound good', and then just achieve that - and perhaps, get 50% of the way there toward a 'studio sound'. I'd rather set the bar really high and get close, than set it low and easily achieve it.
When listening to a PA system, we use a stereo, well mixed and mastered, uncompressed audio file that we know what it 'should sound like'. This allows us to decide where the PA is doing well, or lacking, and gives us a 'reference point' to work from. This song was mixed in a studio and mastered, and we end up with a 2ch audio file to play back. Notice? They don't give you a 3ch mix (Left, Right and Sub). That would be silly, right? Because we expect the sound engineers to have done their job in the studio to manage the low end, and provide a balanced mix that you can play back in stereo.
Similarly, in a live context, if we're aiming for top quality audio, we should be expected that using the tools available to us, we can produce a 90% 'good sounding' mix that is 2ch output, rather than splitting up our 'blank canvas' of the PA and destroying the hard work the system tech may have put in.
A PA system is made up of lots of elements to provide you with a full range 'canvas' to mix on. A large d&b system could comprise of J-INFRAs (27Hz-60Hz, 3x21"), J-SUB (32Hz-100Hz, 3x18"), and J8/J12 (48Hz-17kHz, 2x12" LF, 1x10" MF, 2x1.4" HF). A well tuned system is likely to have elements in CUT mode, and you can see the overlap in the following diagram:
stream / recording - matrixes off LR
corporate - E3/5/8, need sub to fill in low end