I'm curious - what would you consider an acceptable lower limit for a LF crossover point in a passive three- or four-way speaker? What goes wrong when you go below this limit?
I said I would answer you later. So here is my reply. It is of necessity lengthy and somewhat complex To understand this you also have to understand the formidable problem of creating a three way speaker with passive crossovers.
Now the mid driver is also known as the band pass driver. Because passive crossovers have gentle slopes there is a lot of driver overlap. This results in what is called band pass gain. Now this comes about from the addition of the LF, band pass and HF drivers. In order to be able to cope with this in a reasonable way you need to keep the two crossover frequencies about three octaves apart. Now because there is significant out of band response either side of the crossover frequencies the band pass driver has to have a controllable response an octave either side of the crossover frequencies. So the mid, band pass, driver needs to have a response covering five octaves.
Now you know why I'm so interested about full range drivers. More research in this area, is bound to also result in better band pass drivers.
So let us look more closely at the problem confronting the designer of a three way speaker.
First take a look at the problem of constructing a low pass filter in the 120 Hz range.
Now there is a huge problem here. The reason being that the values of the inductor and capacitors involved are huge. Now I leaned years ago not to do this, and I have read many authorities that put the lower acceptable limit for a passive crossover at around 350 Hz at the lowest.
I doodled with an 8 ohm nominal second order three way filters with the low pass to band pass at 125 Hz. This circuit had an inductor of 16 mH and a cap of 103 mfd!
Now those values are huge. That inductor in series with the woofer will have a series resistance of about 1.5 ohms. This will play havoc with the bass tuning. The other solution is to give it an iron core with all the saturation and hysteresis problems that entails. The cap being in the 100 mfd range is going to call for a non polarizing electrolytic type, that are not the best components for a speaker crossover. The other alternative is to parallel expensive polypropylene types.
In addition to that, the crossover is operating close to the impedance peaks of the bass drivers. The interaction of those huge inductors and the speakers within an octave of their double hump impedance peaks from the reflex cabinet tuning, has the potential to create a situation of complex phase angles between voltage and current. This is likely to present a difficult load, and crossovers such as these are well know to actually create amplifier ringing in some amps.
The next issue, is that it is virtually impossible to create a passive low pass filter with a slope greater than 12 db per octave in this range. Now the mid driver operating that low will be starting to roll off second order, for a combined electrical and mechanical roll off somewhere between 18 and 24 db per octave. Now it is unlikely the woofer would start to roll off at 120 Hz, so the slopes will not likely sum to a flat frequency response.
The next problem is that 12 db second order filters are inclined to put the drivers 180 degrees out of phase at crossover, therefore creating a deep null at the crossover point. The same would be true between the mid and tweeters with second order filters. The time honored solution is to reverse the phase of the band pass driver. However doing that at 120 Hz, will put a good portion of the bass out of phase with other speakers in the system.
Now with an active electronic crossover and amps to each driver, the interaction of crossover inductors and driver is a non issue. Any order of slope to the desired driver is easily possible. There is no insertion loss from these huge passive components. The insertion loss of a passive crossover at 120 Hz is huge.
So how did this all happen. I think the first answer is marketing. There is a perception that a three way speaker is superior to a two way. I will show you that is not necessarily so.
Let us consider a three way with crossover frequencies at 350 Hz and 4 KHz then the band pass driver needs to have a good response from 200 Hz to 8 KHz.
Now developing a driver to cover that range is a huge R & D undertaking. There is a permanent dearth of good band pass drivers with these characteristics. That is one of the reasons a pair of B & W 200Ds cost $23,000 a pair.
So what if you only have a standard 6.5 inch bass mid driver to press into service as your band pass driver. No way is that going to have a suitable response to 8 KHz. There will be peaks from cone break up long before that. Few would make it to 4KHz.
So crossover points at 120Hz and 1.8 KHz give you your three octave band pass spread, but leave you with a passive crossover in a bad place, unless you design with active filters.
So to summarize this is the dilemma of the three way passive solution.
1). Develop a driver to cover 200 Hz to 8 K Hz like B & W. A formidable undertaking to say the least.
2). Try and smooth the response of the bass mids they selected to well above 4 KHz. That would be a next to impossible assignment.
3). Narrow the band pass range to less than three octaves. That would be a terrible option.
So you can see why two and two and a half way speakers are so popular. With a passive solution it puts the crossover generally in the 1.8 KHz to 3.5KHz range. This is a viable solution. There are now plenty of small bass/mids that can reach below 50 Hz, and some to 30 Hz. This makes for an easy splice to a sub. This explains why the speaker industry is so geared this way.
In order to break the mold and go to extended range integrated systems, we need drivers able to cover much larger frequency ranges and make the powered solution with active crossovers more common.
