Quote from diyaudio.com
'So by covering the 90Hz though 4,000Hz with one driver we completely eliminate all crossovers in the critical hearing, vocal and fundamental range of most instruments. I am completely annoyed with beaming that goes with most larger wide-range drivers,
I incorporate that technique into my projects. It works well. I will not crossover lower than 3.5k, preferably 4Khz or over. Though THEIR IS signficant audio information still in this range, I find it to be of minimal consequence. Also remember that IM distortion rises significantly when trying to cover too wide of a range with a single transducer.
I do find that their are some drivers that are suited for that range that are not so small as you seek. A 4" driver will have to be crossed signifcantly higher than 90 Hz. You can use a few 5" units that low, but will then have to use a ported enclosure tuned close to that range which IMO compromises that frequency range. Realisticly, you will need to cross with a steep slope at 150-175 with any 4" driver, while allowing for acceptable output levels. I do not see any way to use a driver 5" or smaller under 100 Hz, with excellent results. You will deal with either non linear distortions or poor transient performance.
Focal makes a very coherent W cone woofer and an Audiom midrgane, 6.5" and 7" respectively, thatoffer outstanding off axis performance a little over 4 Khz. Suprisingly, many of the 'full range' drivers have poor off axis response over 3 Khz, even though on axis in flat to around 20 Khz. Pay attention to this little fact when viewing their data sheets.This will result in poor power response, very undesirable IMHO as this will seriously damage imaging"
I found this particularly interesting as I use the doped 5.25" mid bass units full range. Well their output rolls off due to the dope from around 6KHz so the crossover with the tweeter is well out of the most important range
But it is not all good news
Beaming is happening and there is increased intermodulation distortion. I don't know what IM distortion sounds like and don't know how to reduce it without going three or four way and using electrical filters. Filters are not happening so I must accept that everything is a compromise. I don't know if having four mid bass drivers per channel will reduce IM distortion as the work is shared out and the cones hardly moving.