The BassZilla Diamond Edition Open Baffle (OB1)
©Dick Olsher 2006
For the past year I have been intrigued by the idea of translating the BassZilla Diamond Edition to a full open baffle (OB) design, with the entire driver complement located on a single large open baffle. Recall that the standard Diamond Edition is a hybrid of sorts, combining a conventional bass cabinet with an OB-mounted full range driver. A couple of DIYer emails, and in particular the enthusiasm of Jonathan Noble for this concept, proved to be the catalyst that led me down this path of investigation toward conversion of the BassZilla’s bass range to dipole operation.
Please note that the Diamond Edition OB series is not intended as a replacement for the standard Diamond Edition, but offers a alternative approach for those looking to simplify the woodworking requirements of the original and are not afraid of bi-amping. The Eminence Definimax 4015LF woofer has been retained, which works well with current-drive amplification (First Watt F2), but is less than optimum with standard voltage-controlled amplifiers, and unfortunately that includes 99.99% of the amplification universe. Thus, consider this design as Chapter 1 in this design quest; hence the designation OB1. There will most likely be an OB2 incorporating a high Qts woofer which should work optimally with conventional amplifiers.
A starting point for this discussion is the VISATON NoBox BB kit (see Tip of the Month #55). Please be sure to check out my review of this kit for a basic overview of the various technical issues relating to dipole bass. One key point is baffle size. It’s easy to get the impression that bigger is better, but that is not necessarily so. A large baffle introduces a significant reflecting surface into the listening environment. It can interfere with the distribution of reflected energy (especially rear-wall reflections) in the room and may adversely impact the soundfield at the listening seat. A large baffle is also more likely to flex and resonate due to the action of the woofer. And there is always the domestic spouse acceptance factor. The largest baffle size one could imagine is an infinite open baffle, which can be approximated by mounting a woofer into a listening room wall so that it vents into another room or the great outdoors. The downside, of course, is that the backwave is totally lost from the listening room. Much of a dipole’s rich ambiance in fact comes from rear-wall reflections, which means that an infinite baffle’s in-room behavior would resemble that of a large closed box. Clearly, too small a baffle is counter productive, unnecessarily giving up bass extension due to front-back cancellation. It occurred to me that the VISATON NoBox baffle meets the Goldilocks criterion: its’ not too large nor too small. Approximating in extent the critical 100 Hz half-wavelength, it’s just right! It makes for an ideal OB launching pad. Therefore, the decision to embrace it with open arms as a "mule" for evaluating various woofer and wide/full range drivers. If you already own a pair of the VISATON baffles, then you face the simple task of mounting the driver complement. Otherwise, these baffles should pose no challenge to even an inexperienced woodworker.
2. DRIVER COMPLEMENT
The woofer is the Eminence Definimax 4015LF, which is also currently recommended for the standard Diamond Edition (see Tip of the Month #54). The cornerstone of the Diamond Edition remains the Lowther DX4. However, I felt that this was a good opportunity to investigate new ribbon tweeters with more output in the uppermost octave (10 kHz – 20 kHz) than offered by the Aurum Cantus G2Si to better match the DX4 above 10 kHz. The tweeter selected (Fountek NeoCd2.0) is available from Madisound. In its stock form this is a good but not spectacular performing tweeter. Enter Tonian Labs (www.tonianlabs.com). They offer two Fountek tweeter modifications. The standard version simply involves changing out the ribbon, but drops the efficiency as much 4 dB. The standard version will work just fine with lower efficiency speakers. The high-efficiency modification requires machining of the enclosure as well as use of a new ribbon. Pricing for the Cd2.0 mods is as follows:
Active crossover networks offer many performance advantages. The following excellent discussion is lifted (with permission) from the Pass Labs Owner’s Manual for the XVR1 electronic crossover, authored by the dynamic Nelson Pass (www.passlabs.com).
“Higher power. Typically two amplifiers driving a speaker with an electronic crossover will deliver peak wattages almost 4 times the peak (8 times the rms) wattage rating of a single amplifier of the same power rating. This is because the voltage waveforms coming of the high and low pass filters are not mixed, and the high pass voltage does not ride on top of the low pass voltages. When the waveform of a low frequency approaches the limit of the amplifier, there is no voltage left over for the high frequency. With an electronic crossover, the available voltage is nearly doubled, and power is quadrupled.”
“Better use of power. Passive crossover networks divide down the power output of the amplifiers, both in adjusting the efficiency of a driver against the other drivers used, but also by dividing down the output signal as the frequency leaves the bandpass frequency range. Very often this means that only a fraction of the power available from the amplifier can get to the loudspeaker driver. It also means that the least efficient driver sets the efficiency standard for all the drivers. Active crossovers do not suffer from this effect.”
“Higher damping factor. Many loudspeaker drivers depend on a low source impedance from the amplifier to deliver their best performance. Even moderately good amplifiers have a quite low source impedance, also known as high damping factor, and this quality helps the loudspeaker driver to start and stop its motion more quickly and also evens out the frequency response. Any passive crossover introduces significant impedance between the amplifier and the amplifier, and can reduce damping factor from 1,000 to as low as 1.”
“Lower distortion. In addition to lowering the distortion of individual drivers and eliminating the linear and nonlinear distortions of high power passive filter components, active filters dramatically lower the intermodulation distortion of the amplifiers. Since the high frequencies share the same amplifier as the low frequencies, the low frequencies can modulate, or change the amplitude, of the high frequencies, particularly under high power conditions.
“More control and flexibility. It is difficult and time consuming to adjust the characteristics of a passive crossover filter. It is easy with an electronic crossover to adjust the crossover frequencies, the filter cutoff rate (slope), the “Q” or sharpness of the cutoff knee, and the volume level of the loudspeaker driver. Unlike a passive crossover, these characteristics do not depend on the driver’s impedance which varies considerably at different frequencies.”
“Tailoring amplifiers to speakers. Some amplifiers are more appropriate for powering different frequencies and loudspeaker drivers. As an example, you might prefer to drive your woofer with a big powerful solid state amplifier which will deliver the maximum damping and control and drive your midrange and tweeter with a tube amplifier chosen for its ambience and the sweet character if its high frequencies.”
The OB1 is designed to be bi-amplified, specifically using the First Watt F2 power amplifier for driving the Definimax 4015LF woofer. Why that is an important prerequisite should be evident from the following Figure which shows the difference in performance between the First Watt F2 (Green curve) and F3 (Red curve).
Both measurements were performed near field with the 4015LF mounted in the NoBox baffle. With the F3, a voltage-controlled amplifier, the response starts rolling off at 80 Hz, while the F2 keeps going and going.
It should be pointed out that although the Definimax is matched in efficiency to that of the Lowther DX4, in a bi-amplified system it is not necessary to match woofer sensitivity to that of a high-efficiency wide range driver, as the relative driver levels may be adjusted at the electronic crossover using the High-Pass (HP) and Low-pass (LP) volume controls. In fact, the speaker’s tonal balance in terms of bass weight versus midrange and treble emphasis is totally under user control.
I was fortunate to have on hand the Pass Labs XVR1 solid-state electronic crossover. This is an amazingly flexible device that allows user selection of crossover slopes, from 1st order all the way up to 4th order. In addition, a wide range of crossover frequencies may be selected using a jumper system. There is literally a gazillion number of possible settings, which makes it feasible to conduct an in-depth investigation of driver integration. And this was exactly what I embarked on over a period of many weeks, examining every available slope and several crossover points to determine the optimum settings for the Definimax 4015LF woofer and Lowther DX4. The basic idea was to keep an open mind and give both symmetric and asymmetric combinations of HP/LP slopes a chance to perform. Settings that appeared promising on the basis of acoustic measurements, were then selected for listening test evaluations. In the end, the best sounding crossover parameters, and by a large margin, turned out to be 250 Hz/6 dB per octave for the HP channel feeding the upper range amplifier and 250 Hz/18 dB octave for the LP channel feeding the bass amplifier.
Once the optimum parameters were established, I obtained a Marchand XM26 tube crossover outfitted with these frequency/slope modules (www.marchandelec.com). At my request, the unit was equipped with stepped-attenuator volume controls. Unlike the situation with a potentiometer, this feature allows setting volume levels in a reproducible fashion by simply counting attenuator steps. In subsequent auditions, the XM26 acquitted itself marvelously. Its palette of harmonic colors and dynamic shadings is superb with just a slight hint of tube warmth making its presence known. The XM26 can be confidently recommended for even high-end bi-amp applications.
The complete OB1 crossover layout is given in the crossover schematic which is available free upon request to BassZilla Kit Plan owners. The Kit Plans have been updated with the OB1 details. Please email me at: firstname.lastname@example.org. As with the standard Diamond Edition, a passive EQ network is used with the DX4 and a passive HP network is still used for the super tweeter.
From an engineer’s viewpoint, the term fast bass sounds like an oxymoron as a bass transient’s rise time is actually defined by its upper frequency content which is not reproduced by the woofer in a multi-way system. However, in the context of a musician’s vocabulary, I think that the term makes a lot of sense. It refers to lack of undamped box resonances, which by their nature exhibit a long time signature. A decay time in the hundreds of milliseconds translates to muddy bass when bass lines are obscured by the continued outpouring of sound energy at one or more resonant frequencies. In the extreme, this can lead to one-note bass reproduction, where pitch definition is completely obliterated. The joy of OB bass is its lack of boxy resonances – assuming for the moment that the open baffle is itself well damped. Bass lines are generally detailed and well controlled, given a system Q no higher than about 0.7 and a decent matching bass amplifier.
The separation of bass and upper frequency ranges in a bi-amped system - ahead of the power amplifiers - greatly increased the headroom of the upper channel amp. Relieved of the chore of having to negotiate the bass range that 8 wpc SET amp of mine sang with greater purity and clarity (lower distortion) and enhanced dynamic conviction. In general, and in contrast with a passive crossover, the midrange benefited the most, attaining a sense of ease and nuanced voicing that I never heard with the passive Diamond Edition.
The following figure shows in-room frequency response measurements, with the green curve being on the tweeter axis, while the yellow curve corresponds to the response on the Lowther DX4 axis. I am quite pleased with the resultant driver full –range-tweeter integration, always a difficult task when the super tweeter is fairly directional in the vertical plane. There is, however some change in sonic character when the listening axis is varied from the tweeter level to the full range, and you should therefore experiment to see which axis you prefer. My favorite is with ear level slightly above the full range axis.
I also appreciated the ability to dial in the sort of upper bass (120 to 240 Hz octave) and lower midrange weight that I like by adjusting the volume of the bass channel – quite a bonus relative to a passive system where you are stuck with a predetermined tonal balance.
5. IMPLICATIONS FOR THE STANDARD DIAMOND EDITION
Since some of you may be interested in modifying the standard Diamond Edition to use the Tonian Labs version of the Fountek NeoCd2.0 ribbon tweeter, I would like to specifically address this topic. Although I have not tried this with the standard Edition, I see no reason why this new tweeter should not work in that context as well. If you want to try it, here are the required crossover network changes.
(a) Lowther DX4
(1)Change series coil after EQ network from 0.30 to 0.20 mH.
(2)Change the “Zobel” network from 12 Ohm/2 uF to 10 Ohm/7 uF
(b) Fountek NeoCd2.0
(1) Change the HP filter to that shown in the OB1 schematic: a 3.3 uF series cap followed by a 0.30 mH shunt coil.