#subkick

The Thiele/Small parameters that matter for making a subkick are: Fs (free-air resonant frequency in Hz), Vas (compliance equivalent volume in liters, or how loose the suspension is), and Qts (unitless total Q, or the inverse of the speaker's ability to damp its own movement). These parameters are interrelated -- lower Fs tends to correlate loosely with higher Vas, which in turn correlates with higher Qts. If you've ever designed a speaker cabinet, this is all familiar.

The speaker’s rated impedance isn't that important either, since it's measured at 1kHz, outside our range of interest for this application. Don't worry about matching impedance with a mic preamp -- an 8-ohm speaker might have an impedance of 50-100 ohms at its resonant frequency, which is not that different from a dynamic mic. Most any input impedance typically found on a mic preamp or line-in will appear to the speaker transducer as a nearly open load -- and since the resonance is what we want, there's no reason to try to damp it by presenting a super-low impedance to the speaker. That said, if you have a variable-impedance preamp, you may find that setting it below 300 ohms slightly chills out the speaker's fundamental resonance while retaining the initial attack, which can be useful in some circumstances (but then, why are you using a subkick?). The point is, don't waste money on transformers for a subkick.

The cone's diameter doesn't really matter, except to the extent that larger drivers tend to have lower Fs and higher Vas and Qts -- but they also have exponentially larger cone mass, which means they respond much more slowly to transients. Since subkicks as a class generally have poor transient response already, I think it's best to stick to smaller (5-8") woofers that have the right specs. Besides, no one wants to store or lug around a 15-inch, 23-pound "microphone." Likewise, lighter-weight paper cones will perform better than heavier poly, kevlar, or aluminum cones (metal cones may ring at upper frequencies too).

Power rating (watts) is a thermal limit and is utterly irrelevant for subkicks. Xmax, as a mechanical excursion limit, could be relevant to SPL handling, if Vas is very high and Xmax is very low. The cone could bottom-out on big hits and you'll get nasty noise and distortion. This is unlikely to be a problem in the vast majority of cases, because a woofer designed that way would be terrible for its intended purpose of actually reproducing audio.

The original Yamaha NS10 (JA1801) woofer sported these specs: 54Hz Fs, 28L Vas, 0.57 Qts, 8 ohms, 7" diameter paper cone.

What this means practically: the JA1801 will ring at 54Hz, regardless of what frequencies are actually exciting it. The 28L Vas indicates a pretty relaxed suspension, which means that the cone is more easily excited, and with the relatively high Q the resonant fundamental will sustain quite a bit longer than the initial microphonic attack. Sometimes the perceived differences in frequency response between different subkicks are actually time-domain differences -- one might sound bassier than another even though their size & Fs are the same, simply because one rings longer than the other.

You can choose a woofer with different specs to get the performance you want. If you want a higher or lower fundamental, choose a woofer with a higher or lower Fs (I like a little bit lower). Keep in mind that there will be plenty of output above and below the peak resonance, so don't choose a Fs so low that you have to highpass or lose headroom on mostly inaudible subsonics. If you want a tighter decay and less ringing, go for a lower Vas/Qts -- though below ~0.25 Qts, the woofer becomes just a bad microphone and not much of a sub synth anymore. Anything above 0.8 will likely be unusable without heavy gating.

The pictured DIY subkick is made with a cheap Peerless SDS-160F25PR01. It's a 6.5" 8-ohm woofer with 42Hz Fs, 29L Vas, and 0.49 Qts. It's very similar to the JA1801, but it rings a little deeper, has comparable sensitivity and SPL handling, and has a slightly shorter decay. I like it a lot.

Above, you can see a frequency response plot of the subkick versus a GT55 condenser mic (see previous post) on a high-gain guitar cabinet (CV75 speakers, 5150 lead channel, re-amped downtuned chuggy rhythm part). The amp/cabinet rolls off steeply below 100Hz, which is captured well by the condenser (it has fairly linear response down to 20Hz). By comparison, notice that the subkick has +12dB response at 50Hz, despite the fact that the condenser already has its bass exaggerated by proximity effect! That's the woofer resonating -- it isn't a part of the actual sound field. At 25Hz and 100Hz (one octave above and below), the amplitude falls back to pretty close to the condenser. You can also see a little kink/bump right at 45Hz in the subkick plot that isn't in the condenser plot, which is the tip of the woofer's resonant peak poking out (note: it would poke out a lot more on an impulsive/high-transient source like a kick drum -- a guitar cab is basically the complete opposite, more like a steady pink-noise source, great for evaluating microphone differences). You can also see the subkick's cone breakup modes happening from 1-2kHz, and then it rolls off a lot above that -- in most cases, it helps to lowpass the subkick around 300Hz to keep the garbage out of the signal.

Enclosures: Yamaha made an actual branded Subkick built inside a drumshell (and on a drum stand) for a while, but it didn't sound like the original open-woofer DIY setup (and also didn't use a JA1801 woofer). Plenty of DIY subkicks are mounted in old tom shells or other homemade boxes, which can help to protect the cone and the wiring. The problem here is that any speaker enclosure -- especially a drum shell! -- adds its own secondary resonance and alters the woofer's impedance and Qts to be much more complex. The easily-predictable free air resonance of the woofer is what we're after with a subkick, not the complex interaction of two separate mechanical resonators. The other problem is that an enclosure that's not totally open in the back will change the pickup pattern from figure-8 (pure velocity or "pressure-gradient") to a weird sort of cardioid (combo pressure & velocity) with unpredictable off-axis response. That could potentially reduce the subkick's excellent rejection of the rest of the kit, by widening its front pickup, and moving its nulls from 90 degrees off-axis toward the rear. Better to mount the speaker to a ring, or just bolt the frame directly to the bottom half of a standard mic clip (as pictured). If you get really lucky, an enclosure might add something useful to the sound, but it's incredibly difficult to model that in advance. If you're worried about protecting the woofer cone, save any appropriate packing material (if you bought it new, see picture below), or bolt some small screws through the empty mounting holes and attach a cover that way.

Wiring: Take a balanced output directly from the voice coil (it's exactly like a transformerless dynamic mic). Getting the output balanced from the get-go will basically eliminate any noise picked up in the wiring, although the voice coil itself will still be sensitive to hum/buzz (FAR below its output level, though). There's no reason to unbalance the output with noise-susceptible TS wiring, and then have to waste a DI box converting it to balanced.

The thing to remember is polarity. Speakers are marked with a positive and negative terminal -- the positive terminal is where you apply a positive voltage to get the cone to move OUT (away from the magnet). Since we're using the speaker in reverse, we want a positive voltage to appear when the cone is pushed IN (toward the magnet) by a blast of air. That positive voltage will appear at the negative terminal relative to the positive terminal. So to follow standard microphone convention, you should wire the negative speaker terminal to pin 2 of the XLR connector (or the tip of a TRS plug). Wire the positive speaker terminal to pin 3 of the XLR (or the ring of a TRS plug). You can verify the polarity is correct by looking at a time-aligned recorded waveform that has lots of upper harmonics (like a distorted guitar), i.e. lots of asymmetry. Don’t try to verify by flipping the phase and listening -- frequency response is so different from a normal mic that it’ll sound sort of ok both ways. Look at the waveform!

Pin 1 of the XLR is chassis ground, and you can either leave it unconnected, or use it to ground the speaker's frame/chassis/magnet (as I did with the subkick pictured). Either way, pin 1 should be totally isolated from the voice coil and from pins 2 and 3.