Sub Woofers (Part 2)
A last look at how to extend the low bass response of your studio monitors.
Ben Duncan concludes his look at extending the bass response of your studio's monitoring system.
For the sub-woofer's output to integrate easily and naturally with that of the existing monitor, the speaker drivers concerned need to exhibit a sensible overlap in their capabilities. At one extreme partnering a 6" bass/mid driver with a 30" sub-woofer produces a hole in the overall response: the 30" driver's output begins to fall off circa 150Hz, well before the 6" driver gets going properly. Why? Well for a 30" driver, 150Hz is quite a high frequency. At the other extreme, there'd usually be little advantage in partnering a monitor with an 8" driver together with another 8", or even a 10" sub-woofer (SW) - unless the SW driver possessed superb long-throw capabilities, and thus 6 to 10dB more PHC at the bottom end.
In order to make a reasonable match, especially at the crossover point, we need some idea of the maximum frequency (fmax) the sub-woofer can reproduce clearly, and conversely the minimum frequency (fmin) that the existing monitors can comfortably cope with. Fmin is NOT to be confused with the low frequency response (or fL, the -3dB point): it involves a juxtaposition of this and the low frequency displacement limit, Xmax. As a rule of thumb though, fmin can be taken as being typically one octave above the fL, where the bottom end response drops 3dB (Figure 1). In case you didn't know, going one octave up means doubling the frequency. For example, looking at the frequency response of the Visonik David 6000, fL = 150Hz, so fmin = 2 x 150Hz = 300Hz. Simple!
The table in Figure 2 gives typical fmin and fmax figures for bass/mid and SW drivers respectively. This is only a generalised guide however - specialised drivers will differ, and the most comfortable fmin for micro-monitors such as the Visonik involves other factors, such as intermodulation distortion.
In Part One (November 1984) we nominated 100Hz as a good, general crossover frequency. Pushing up the crossover point to 200Hz, say, will greatly reduce the degree to which hard bass can muddy up the midrange (ie. lessens intermodulation distortion), but then the integration will need to be good, because with the frequency being higher, our hearing will also be more alert to the sub-woofer's physical displacement from the main monitors. Also, of course, the sub-woofer's response may 'drop out' if we go up too high.
To cut a long story short, fmin and fmax are best derived (at least initially) from the manufacturer's response curves, as described above, allowing generous margins. You'll then be able to tweak the crossover point empirically, for best results. By way of example, to extend the response of the Visonik David 6000, we saw that fmin equals 300Hz. This implies that any SW drivers up to the 15" size (fmax = 300Hz) will make a good match.
By comparison, a main monitor system with an 8" driver might have fL = 45Hz, so fmin = 90Hz, and so the larger 18" and 24" subwoofer drivers could be considered. Just to put a spanner in the works though, a sealed-box monitor like the Celestion SL6 has a relatively large mid/bass driver, yet fL is quite high, at 130Hz. As a result, its sub-woofer requirements, perhaps surprisingly, will be similar to the Visonik 6000. You'll therefore need to peer closely at the maker's specification and response curves.
In the last year, special drivers designed explicitly for sub-woofer work have become available. These feature high 'Xmax' figures for a healthy PHC at the bottom end; in plain language, this means the cone can move further before it first distorts the sound badly and, second, hits the end stops. In the new Electro-Voice models, the magnet's pole piece is also covered with PROTEF, a Teflon coating, which helps soften the impact if the drivers' voice-coil bottoms ie. hits the top-plate.
The graph in Figure 3 outlines the extra bass output that specialised drivers can offer us; the curves depict maximum PHC. The EVM15B is a typical 15" driver, whereas the DL15X/W are the special sub-woofer models, all from Electro-Voice. Clearly, the DL models have extra PHC, but how much? And is it significant?
Reading out a 100Hz, where much of the energy in music is concentrated, the DL15X's capabilities are 6dB greater than the EVM15B. This means we can safely uprate the relative amp power by 3 to 4 times. The DL15W, meanwhile, adds a modest 3dB extra PHC at 100Hz, but scores 6dB or more at the frequencies below 100Hz. Both drivers will take more punishment, and retain a clean sound up to higher levels, the difference versus the EVM15B being particularly evident on bass-end percussion. But the DL15W is better suited to accurate full-range monitoring, whereas the 15X stresses efficiency, at the cost of losing out below 60Hz - notice how its curve crosses over below this point.
So far, we've assumed you're building a sub-woofer from scratch. Off-the-shelf subwoofers are available for a few monitors - in particular the Visonik family. More often though, sub-woofers are sufficiently bulky for the customised approach to be worthwhile. This means building the box into the room. At this stage, and assuming flush mounting in the studio wall is out, then the opportunity to 'save' space still arises. For even if we can't shrink the enclosure, it's often possible to fit the sub-woofer into space that's wasted anyway, like in an awkward corner, which would otherwise only accumulate dead beer cans and Rizla packets.
Another short cut, perhaps to help you check out the validity of the idea - and also if your budget is tight - is to utilise or convert an existing bass cabinet. Assuming largish monitors, the best bet is to use a 'roots' bass instrument speaker - say 1 x 15" or 1 x 18", etc, while for boosting micro-monitor bass, look out for any cabs with 4x12", or better, 4x10" drivers. Although their output is slugged for the guitar's fundamental, (around 100Hz) a lead cab can still boost the low frequency PHC on a micro-monitor.
Returning to the 'from scratch' approach, it's time to choose the loading method. Horn-loaded sub-woofers offer high efficiency, and generate unsurpassed 'bass space' in large rooms, but are relatively difficult to construct and don't operate coherently in the nearfield (ie. when only 1 or 2 metres distant). Direct radiators in sealed boxes are easy to construct, but their low efficiency at the bottom end begs BIG power amps, whereas the relatively weedy PHC forbids too much extravagance in this direction.
The simple answer is to go for a Thiele-loaded enclosure, one with a carefully tuned vent. Today, computer programs are adept at throwing out Thiele cabinet data for every conceivable idiosyncratic requirement, and manufacturers across the spectrum - from Electro-Voice to WEM - will provide plans for accurately tuned cabinets on request. Naturally, this info relates solely to the maker's own drive units! The EV 'TL606' cabinet for a 15" subwoofer is perhaps the best known design of this type. Like the direct radiator in a sealed box, Thiele enclosures sound good close up and are scarcely any harder to build, yet they routinely exhibit a much higher efficiency (for the same low frequency response) or an LF response which goes much lower, for the same efficiency.
The DIY literature available on sub-woofers is limited. Out of 60 or so speaker manufacturers, only three came up with DIY info. Celestion's handbook shows two useful designs: a '2x15" vented cab' and a 'horn-loaded reflex bin'. Sadly, there are no response curves given - but Celestion may be able to supply them on request. Fane's cabinet handbook is much more detailed, and included accurate response curves. Designs to check out include the '1x15" rear-loaded horn', '1x15" front-loaded horn' and the '2 way reflex system'. All these designs are substantially flat to below 50Hz, and the purely vented ('reflex') cabinets can be shaped to best suit the available space - just be sure to keep the overall volume within 2% of the original, and avoid significantly altering the vent's distance from the rear panel and the drive unit.
On a more refined level, Electro-Voice publish exhaustive data sheets on their supercharged DL series, and other bass units. Any of these can be 'dropped into' one of their computer-aided cabinet designs, which are all vented, and finely tuned to boot. For example the TL606, which we discussed earlier. Real enthusiasts may like to design their own subwoofer enclosures from scratch, using the special Thiele data presented in the speaker's specification, together with easy-to-use tables, published in a US specialist speaker journal (address at end). This approach is unlikely to save you any money if you value your time, though...
Ordinary passive crossovers aren't feasible for a central sub-woofer driven from a summed, Left and Right signal: but they can be used for stereo sub-woofer schemes. Few off-the-shelf designs exist, but speaker customisers like Kord Audio can produce low frequency crossovers to order.
Active schemes can take several forms. For sub-woofers, we need a two-way or 'Bi-amp' crossover. The low pass, sub-woofer side needs to roll-off above 100Hz (say), while for the main monitor the converse applies: ALL frequencies should pass, other than those below 100Hz (say). With sophisticated units, the crossover points are readily reprogrammed with plug-in cards or modules, the 'bass' output going to the sub-woofer, the 'treble' going to the main monitors. Because the term treble is something of a misnomer here, we'll call it the high pass output.
Budget crossovers generally have fixed, or switchable, yet specific midband crossover frequencies, say no lower than 250Hz on the low pass ('bass') output, and 2kHz on the high pass output. But don't fret: the crossover point can be taken down to 100Hz by increasing the appropriate capacitor values - ask the manufacturer for advice, or consult a boffin! You may also need to link-out any sub-bass filtration at 60Hz, it's frequently included to protect PA speakers from dropped microphones, but won't be needed here.
Crossover slopes of -24dB/octave are ideal, or second best, -18dB/octave. Remember: the -24dB slope has the most seamless integration, providing both the response curves are level around the 100Hz region.
It's much easier to set the balance on crossovers with ±6 or ±10dB attenuators (say), rather than the usual 0 to -70dB 'volume control' arrangements. Gain setting is critical to within 2dB or less, yet easily set by ear, given a control with good resolution. Alternatively, forget the attenuators on the crossover, and set them at 10: provided the sub-woofer-cum-amp goes louder than the main monitors (this is often the case), then the subwoofer level can be wound down to balance its amplifier's own gain control. These, and other factors are covered in detail in our survey of Active Monitor techniques (See HSR December & January 83/84), while Figure 4 illustrates the basic hook-up.
Figure 5 gives details of a DIY Summing Amp. This converts the stereo feed to mono before feeding into the sub-woofer side of the active crossover. Note that the output is typically 2 or 3dB greater than each input, depending on the instantaneous phase difference between channels. This 'gain' is enough to help balance the sub-woofer, using an attenuator alone. Naturally, a DIY active crossover project is outside the scope of this text, but the four part series 'A Versatile Active Crossover' which appeared in Hi-Fi News magazine in 1981 includes extensive design information and tables, enabling the keen constructor to amend the article's 4 way DIY design to sub-woofers.
Lastly comes the choice of power amplifier. Given a single sub-woofer, and the fact that most amps are stereo, the advantages of bridge-mode (mono) operation should be evident. This means we can drive the amp in mono, without wasting any of the power, or more to the point, headroom. Generally, the sub-woofer will be more efficient than the existing monitors, so you may find an amp smaller than that used on the main monitors will suffice. But do bear in mind that the main monitors can be driven harder, having been relieved of that low-bass burden; this suggests larger amps all round. If in doubt, temporarily hooking-up any amplifier of known power rating will help you arrive at just conclusions. And don't fret about small power differences - the output only needs to be in the correct ballpark; so long as it isn't overdriven, the exact output isn't at all critical. One other point: redundant amps, the sort with a grotty top-end sound, can be roped in for subwoofer work with impunity. Perhaps there's an old WEM 100 watt slave lurking in your garage?
Sub-woofers are positively guaranteed to excite the low frequency (LF) acoustic idiosyncrasies of the control room, including ones you didn't know about beforehand! In untreated spaces, monitors with a restricted and gently rolled-off LF response will sound cleaner and more articulate because 'bass-in-the-head' doesn't aggravate the room modes.
In general, these room modes are least troublesome when you site your ears away from the walls and the room's centre. If the control room floor and ceiling are untreated and parallel, adjusting the height of the seating will also greatly influence the bass quality, with the standing waves being worst close to the floor and ceiling. The sure remedy is to followup on HSR's Acoustics series with hammer and nails. Failing that, bass booming is a sure sign that the LF energy input is too much for the room to cope with. Turning down the subwoofer's amp will help here.
If the sub-woofer bass sounds disjointed, this implies that (i) either the crossover point is wrong, or (ii), the driver's responses aren't integrating, leaving a gap or peak in the combined response.
Part 1 | Part 2
Feature by Ben Duncan
Previous article in this issue:
Next article in this issue: