If two differently sized drivers share the same enclosure volume, the larger one will have a tendency to push the smaller one. The larger driver will rarify and compress more air than the smaller one. I'm talking about a sealed enclosure, not as much of a problem with a vented enclosure and even less so with an open baffle.
I believe that it's not 2, 3 and 4 drivers that doubles then triples and then quadruples the power handling, but 2, 4 and 8 drivers that does.
I think the biggest problem is probably the difference in SPL between a large and a small driver. Maybe it's possible to attenuate the larger driver with a resistor or an lpad?
It's not too difficult to find large and small FR drivers with the same impedance (generally 8 ohm) and the cabinet thing is pretty trivial to solve. Just do two in one each of the appropriate size with a airtight spacer.
It's not too difficult to find large and small FR drivers with the same impedance (generally 8 ohm) and the cabinet thing is pretty trivial to solve. Just do two in one each of the appropriate size with a airtight spacer.
Last edited:
There's never a best, just different sets of compromises. Assuming you want to keep the midband with the Alpair alone, then 170Hz should be a decent compromise, so long as you put a fairly high order low-pass on the SDX7.
Couple of points though -if you stick a large resistor in series with the Alpair, you may need to account for that in the response, just as you would for a low DF / high output impedance amp. You may at some point want to consider going active, or doubling up the LF drivers to get the requisite efficiency.
Problem there is that as soon as you stick a resistor in series with a bass unit, you're shifting behaviour toward that of a high output impedance amp, or effectively weakening B*L, raising Qe et al, so you'd need to account for that in the process. Low Q driver would probably be needed if you're going down that route. Either that, or, per Thorsten's box, a very efficient mid-tweet. That said, assuming you design it carefully, some greater efficiency in the LF unit is not automatically a bad thing as you could use it to cancel out step-loss etc. See MJK's OB projects for e.g.
Last edited:
For power dissipation, the ability of the voice coil(s) to handle power without overheating, each additional driver increases the power handling capacity by the same amount. If rated at 10W, three drivers can handle 30W total. This represents the thermal limit.
The relationship between voltage, excursion, power, and SPL is not so simple. Two drivers identically driven will produce 6 dB more SPL. But it will take 4 drivers to gain another 6 dB and 8 drivers to gain the next 6 dB. For amplifier power, at a given load, twice the voltage produces 4x the power, which is the same as a 6 dB increase in SPL.
So, to get a 6 dB increase in SPL from a single driver you need twice the voltage and 4x the power. To get a 12 dB increase you need 4x the voltage and 16x the power.
Alternately, you can wire multiple drivers in series, parallel, or a combination of both. In series such an array will retain the same sensitivity (SPL at a given voltage) but demand less current. In parallel the array will demand more current but will increase sensitivity. Two drivers in parallel will increase sensitivity by 6 dB. Three paralleled will increase sensitivity by 9.5 dB, four will add 12 dB. Four drivers wired 2x2 will draw the same current as a single unit, but be 6 dB more sensitive and able to handle 4x the power of a single driver. The system in your picture has 20 drivers for each channel, which might be arranged several different ways.
The basic argument for a crossed over system is that small drivers are going to be inefficient at low frequencies, so their efficiency at high frequencies must be cut down to match when using them full range, and, even with very long throw cones (which introduce an interesting extra frequency related distortion; you can actually get audible pitch shift on the high frequencies due to the bass frequencies moving the source) they have low power handling before distortion.
Meanwhile, the bigger cone, as well as the beaming problems (which can be largely compensated for by careful membrane design) have severe difficulties maintaining piston response at high frequencies, tending towards break-up, so that only part of the cone is moving at those speeds. However well this is done, there are always some distortion products generated; and, since the bigger loudspeaker is more efficient than its smaller sibling, if both are running full range these are not going to be masked.
I was wondering (without having done any experimentation; I was working on active crossovers in the late sixties, and never really looked back) whether a variation on the acoustic lens theory used on diaphragm horn drivers, using only the radiation from the centre of the cone for high frequencies, and the entire surface for lows, might be thinkable? No, to get a flat frequency response you'd almost certainly have to design the driver from scratch; not really practical for a private individual.
Meanwhile, the bigger cone, as well as the beaming problems (which can be largely compensated for by careful membrane design) have severe difficulties maintaining piston response at high frequencies, tending towards break-up, so that only part of the cone is moving at those speeds. However well this is done, there are always some distortion products generated; and, since the bigger loudspeaker is more efficient than its smaller sibling, if both are running full range these are not going to be masked.
I was wondering (without having done any experimentation; I was working on active crossovers in the late sixties, and never really looked back) whether a variation on the acoustic lens theory used on diaphragm horn drivers, using only the radiation from the centre of the cone for high frequencies, and the entire surface for lows, might be thinkable? No, to get a flat frequency response you'd almost certainly have to design the driver from scratch; not really practical for a private individual.
Which is some of the reasons why you want to backload your driver in some way.
But that goes for all dynamic speakers to some degree. Especially in two or three way designs where the mid has to handle a lot of the same conflicting demands as in a fullrange, only to a lesser degree.
If you make a stiffer membrane you will also almost invariably make a heavier one with lower SPL levels, that you have to drive harder with more excursion an so on.
There is good reason people try to get the cone as light as possible.
You mean like a coaxial driver? That's been done loads of times in various kinds of implementations, pressed rings on then cone, coils and dustcaps that are connected to the driver with flexible suspensions, etc.
A FR with a wizzer could also be viewed as such a device.
No, I wasn't intending to redesign a driver; I might get away with that, but not many here will have built loudspeakers from the magnet up. I was thinking back to the sixties. when phenolic diaphragms and horn loading were the standard way of getting high mids loud, and a thing consisting of perforated metal plates that screwed into a horn mouth. the centre was open, and there were successively more plates to traverse as the sound approached the edges. Thus, the sound got successively more delayed as it approached the walls of the horn, and what was an essentially collimated plane waveform became far more curved, a section of a spherical surface centred far further forward in the horn throat, and thus far less directional. I was idly contemplating a macro version of this, with some HF absorption in the delay plates to reduce distortion products from break-up rings… but if the driver was designed for free field flat frequency response, it'll need equalising.
Indeed. Don't want it flapping about. But this will reduce its output in the low frequency region, so, to be able to sell it as a full range rather than a mid range driver, the manufactures will reduce its output in the mids and highs.
I have no attraction whatsoever toward the "heavier cone, lower efficiency, higher power" philosophy. I am an impulse response speaker person, and got quite unhappy when mass was added to a cone to bring down its resonant frequency (giving better paper specifications) then the tweeter efficiency dropped to match with associated damping problems. But that was long ago.
But I don't like plane arrays, and don't see how the multiple goals of distortion, frequency response, impulse response. phase, directionality and efficiency can be combined in a single driver; too many compromises (yes, I know, and I'm not living with a thousand compromises on a band pass system? Perhaps the true answer lies in a different way of moving air).
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.