Manufacturer: B&C Drivers - Italy
Product one: DE-400TN-8 compression hf driver with matching horn, tweeter system
Product two: DE-35-8 bullet tweeter
Prices:
DE-400TN-8 + matching horn - €MMV, as might ¥ or $ depending on currency fluctuations & local trading conditions
DE35-8 - likewise
Supplier: local distributors

Reviewer: Mark Wheeler - TNT UK
Reviewed: October 2009 - January 2010

For many years I have disagreed with the dominant hegenomy regarding domestic high frequency reproduction.

"Well there's a surprise," the plebs heckle from stage left, "He'd argue big old cone tweeters are best if it disagreed with prevailing wisdom!"

I find most soft-dome tweeters at best innacurate, but mostly unlistenable. To my ears they have consistently sounded like a series of resonances, one after another, that hope to be similar enough in magnitude to sound as if they make a reasonably flat frequency response. The poorest examples, some of them quite expensive, buzz and fizz their way from suggested crossover points often far too close to the fundamental resonance all the way up to a peak mechanical resonance of overwhelming proportions that fool the listener into believing that they are hearing previously inaudible detail from their bandwidth limited cds or vinyl reproduced with matching hf tip resonance.

There have been a handful of fabric (usually silk) soft dome treble translators (from electricity to pressure) but even these would be hard pressed to be described as accurate transducers. Most plastic hard domes merely move the resonances about to fit the designers' prejudices. I embraced the early Elac metal dome (more widely known in its Monitor Audio application, indeed MA took over the old Elac company) as a welcome breath of fresh air and reasonably extended pistonic response, even if its much higher than plastic main break-up mode required a little LCR shgunt to tame it in my system. I have subsequently used, since 1993, one of the prototype sample batch of the Focal TD120 titanium dioxide inverted dome with the removeable aluminium phase plug that was then undergoing field trials, and improves on the Elac with its inverted dome shape and lighter, stiffer material. However, even this better example of the domestic tweeter genre (as used in some serious multi-thousand Euro systems) falls way short of what ought to be possible.

An examination of the air velocity patterns over the nose of a bullet at various velocities (the London Science Museum used to have great aerodynamic display cases with mini wind tunnels to excite young visitors' minds with such ideas) gives us clues why conventional convex domes have such uneven dispersion. The magnificently clear 2d dispersion chart of the B&C DE35-8 at the top of this page demonstrates the even pressure distrubution pattern for this transducer across the passband which I have not seen from a conventional dome direct radiator. Off axis frequency response plots of loudspeaker systems typically show a massive suckout at the top of the bass-mid unit passband above which a sharp boost (rising with the crossover slope) to full output affects all the reflected sounds and power response in the soundfield in the room.

One of the many problems of domestic audio, indeed possibly the biggest problem facing the domestic audio enthusiast is probably the interface between loudspeakers and rooms. One of the toughest calls is the changing dispersion pattern with frequency. Hence in a single driver loudspeaker, the higher the frequency the narrower the beam of sound. During the octave between the baffle width equaling a half-wavelength to the baffle width equalling the wavelength the radiation pattern reduces to mostly forward only. If the speaker measures flat on axis in an anechoic chamber there will be a bass boost of around 6dB in a reflective room, averaged out among speaker and microphone positions (see Room Acoustics - Part II). Furthermore, as the wavelangth of the reproduced frequency approaches the driver circumference, the radiation pattern tends to reduce to a 90° cone on the driver's axis.

This is bad enough, but in multiway systems (2-way bass & treble or 3-way bass & mid & treble for example) this happens upwards in the passband of each driver. So the bass-mid unit might be radiating into less than 90° at 2500Hz when it crosses over to the treble unit that is radiating into 180° assuming a wide flat baffle. If their on axis anechoic response is equalised for amplitude the off axis response will be wildly different at 2000Hz and 3000Hz. This affects the in room frequncy response affecting the perceived timbre of musical instruments. The uneven and unequal delayed reflections at different frequencies affect soundstage almost as significantly as on axis timing errors (phase errors).

"We thought this was a product review" sigh plebs, stage left,"Not another lecture on loudspeaker system design"

Which brings me neatly to public address and sound reinforcement practice. To get a reasonable sound throughout a venue dispersion is the key ingredient, or balance of ingredients if we are being pedantic. To achieve even distribution of all frequencies the sound system designer needs to be able to specify directivity of every driver throughout its passband. Frankly, variations on the horn theme are the cost effective way to go. In the space domain horns hold the aces at all frequencies (depending on frontal area of course) but especially at the higher frequencies. Well designed horns have another advantage, but this lies in the time domain; the controlled wavefront tends to retain the 'shape' formed at the horn mouth so it is more predictable to estimate when each part of the waveform arrives at the listener's lugholes.

"But surely acoustic lenses on constant directivity horns introduce their own problems that are far worse in the nearfield?" argue well informed plebs, stage left.

Very true, but my favourite tweeter of all time, regardless of its obvious flaws, has to be the Stanley Kelly designed Decca Special Products London Ribbon Tweeter, the one with the 1kHz 90° horn on the front, not the earlier DK30 with the narrower 3kHz horn+lens on the same motor system. Yes the ribbon exhibits high levels of even order distortion (a little like many moving coil cartridges) heard as a tinkly quality on some material, reminiscent of pre-ignition in high-compression internal combustion petrol engines. Yes the ribbon self destructs with regular monotony especially with passive crossovers less than 18dB/8ve at 2500Hz or 24dB/8ve at 1500. I found 24dB/8ve at 1800Hz allowed a seamless transition to a 250mm bass-mid driver, which was so seamless because at 1800Hz that driver's dispersion charecteristics were an excellent match for the London horn.

That 1000 words brings me neatly to why I asked a pro-audio manufacturer to risk their products reputations in an audiophile setting. The ability to predict accurately the dispersion characteristics of a high frequency transducer is fundamental to system interface in a reflective environment like a domestic room. High frequecy drivers designed for professional sound reinforcement applications are among the most predictable, narrow flat panel dipoles being their only serious competitors in this parameter.

Furthermore, most drivers designed for sound reinforcement tend at domestic listening levels to have distortion of a tiny fraction of that shrieked at us from most domestic audio tweeters. The published measured distortion (on those extremely rare occasions that they are revealed) tends to be at the lower end of the designed passband below the first break-up mode of the driver surface. Public address high frequency units tend to have larger radiating areas requiring less movement for the same air volume to be moved, coupled to horn throats designed for optimum loading, to which a horn of the desired directivity and cut off may be bolted.

Pro-sector drivers also tend to be heavier and more solid to withstand the rigours of being heaved in & out of the back of trucks. This makes them less susceptible to airborne and structure-borne vibration from bass.

"So what's the downside? If they're that good why isn't everyone using them already?" demand cynical plebs, stage left, sounding more like New Yorkers than ancient Greeks.
Good ones are very expensive. Very, very expensive compared to the 6-30€ OEM tweeters in most domestic commercial products.

B&C Speakers

B&C Speakers products are still built in Europe, rather than designed here and manufacturing outsourced to cheaper labour markets. Indeed B&C hail from Florence in Italy, so their employees are no doubt surrounded by great art and architecture, which would imply capacity for great thinking as the analytical psychologist James Hillman suggests. B&C speakers are, according to their website, made in a swish new 7000m² building since early 2009.

The B&C tweeter range features edge wound aluminium voice coils, which have two advantages: the material and the winding technique. Low mass is the obvious one for aluninium compared to copper, and more than one cable maker has stated that copper sounds inferior to aluminium (but regular readers know my views about wire hype). The edge wound flat wire technique has a long history in low distortion bass and bass-mid motors because, simply expressed, there is less air and more conductor in the magnet gap and the conductor also conducts heat away more more efficiently.

B&C DE-400TN-8 with matching horn

The B&C DE-400TN-8 is a newly designed compression driver of typical pro-sector design manufactured with high grade materials that excite audiophiles and massive sensitivity beyond the dreams of innefficiency avarice. Coupled to its recommended horn it raises 106dB for the first watt. Who could argue that you would ever need more than a 2.5W single ended 2A3 for this baby in an active system. Let's face it, the crossover distortion and noise from a push-pull class AB solid state amp might amount to 60dB of signal spuriae clanking along with the music through something this sensitive.

The '8' suffix of the nomenclature denotes guess what? You guessed it; these are 8Ω versions. Best of all they exhibit R_min of 7.7ω which combined with that 106dB sensitivity makes them ideal candidates for valve driven high sensitivity systems. Dear reader, you can even use the 8Ω tap on the output transformer with impunity, regardless of how close to the wind sailed the Tx designer and amp builders.

I know my samples are early ones but I found the compression driver magnet front faceplate of one driver had not been correctly tapped with threads. Fortunately I had the correct sized taps in stock so I tapped them myself. I hope that would not be the case in any further examples from production. In a pro setting I would hate to be perched at the top of a stack or clambered down from a gantry to undertake a field replacement of a driver that's been biked over specially, only to find the #*!+#ing thing was unusable.

Notwithstanding the application problems of a 106dB tweeter in a 30m² room, via an active crossover and single ended amplification this big tweeter really sang (after 2 weeks running in with the horns taped together face-to-face playing out-of-phase signals as loud as tolerable elsewhere in the house - it's not easy with drivers this sensitive). They really are very sensitive, although I suggest their average is nearer 104dB subjectively, for the top two octaves, but this subjective effect might be due to their consistent dispersion. The DE-400TN have a beast of a response peak around the region constructors might be considering to cross over, for example a second order crossover at 2kHZ would actually be acoustically flat down to the recommended 1500Hz minimum. For example, I found setting my active crossover at 5kHz high pass matched (by ear) low-pass 4kHz due to the responses of each driver. I but settled on 24dB/8ve at 2400Hz with mid sized bass-mid drivers 160-200mm. I ran all my favourite killer treble albums through them but soon forgot the DE400TN-8's were there.

The B&C DE-400TN-8 really are a pleasure to listen to vinyl and cd. They sacrifice the extended treble of the latest supertweeters, which reduces the illusion of "air around the performers" according to many reviewers. I didn't notice any performers suffocating due to lack of oxygen, but I did notice that the soundstage seems particularly stable with these horns. I suspect the even in-room power response, facilitated by consistent dispersion from 2.4kHz upwards (my chosen crossover point in this excercise), aided this illusion.

Timing is pretty good up until perhaps 15kHz, and certainly much better than soft-domes (of any material) and most hard domes (indeed the Focal TDX phase plug has a dramatic effect in this area too). Above 15kHz the DE-400TN-8 loses timing focus but does not drastically time shift as many domes do as they break up. Stick noise on cymbals is almost timed to state of the art standards, but there are some dipoles that set the pace for this. Off axis the B&C DE-400TN-8 comes into its own as the sound remains consistent from each speaker as the listenner moves either laterally or vertically. However, the old trick of crossing axes in front of the listenners to broaden the stereo seat by the proximity of the nearer speaker being offset by being further off-axis than the farther speaker does not work. With even dispersion there is no axial hot-spot for treble. Position the speakers for accurate bass and direct them for smooth transition at the crossover overlap and the horn directivity does the rest. These are not quite as ideal a nearfield wavefront as the London Ribbon horns but they are designed primarily for far-field use on stage where the symmetrical horizontal and vertical dispersion patterns would be ideal. In a domestic listening room at 3m they are fine, but wider horizontal and shallower vertical dispersion might be more useful in some listening environments; I suspect the horn shape is designed to fit inton multiple arrays. Purchasors could try different horns (25mm or 1" being the commonest throat size on the market and I know of some beautiful carved wooden horns for domestic use with 1" throats), of course domes offer none of this control.

Determined to find more widespread application than in active systems with flea powered amplifiers, I padded them down with a resistor network made of various combinations of values and brands until I had an almost resistive 8Ω load and about 8dB insertion loss. I attached the DE-400TN-8 plus network to the passive crossover of the modified Hammer Dynamics system I use with SE amplifiers. The Hammers use a 300mm bass-mid unit with a 125mm slave or whizzer cone to extend the high frequency response and dispersion. There is quite a complicated passive crossover to flatten the response of the bass-mid driver but a simpler high-pass to the tweeter. The bass-mid driver does not have the dispersion of a 125mm driver at higher frequencies because the front edge of the whizzer lies 30mm behind the driver frame. The crossover frequency is an unusually high 8kHz to keep it above our ears' most sensitive octave, where the Hammers midrange response is most raggedly beamed. Here the superior even dispersion of the B&C DE-400TN-8 drew attention to the obvious gap off-axis (and therefore the in-room power response) at the top of the Hammers' 300mm main driver. The sound was infinitely superior to the dreadful soft dome Audax that come with the Hammer standard kit and much better than a couple of HF drivers recommended for Hammers on various internet forums. But this is one place where the dispersion needs to be more tightly focused than 90° to maintain the off-axis beam to the hot-seat set up by the bass-mid Hammer driver.

Distortion is noticably absent with the DE-400TN-8 and matching horn. These tweeters have that hear-thru quality that enables us to detect minor problems earlier in the reproduction chain. I noticed within the first track that I had slightly misaligned the azimuth on my Cartridge Man MusicMaker when I'd taken the arm off to adjust something on the turntable. The limitations of some review equipment was laid bare but not exaggerated. System set up becomes easy with this driver because it never seems a case of tweeking two wrongs to make a right. This tweeter is right within its limitations; the 18kHz cut-of being the most obvious one, even to my 50 year old ears. However, I prefer acurate information up to 18kHz than a succession of resonances up to 45kHz (a tweeter that spent a short time in the system recently before being returned). The wide acceptible listening area in the listening room comes initially as a shock to those of us used to the hot seat of domestic hifi, but is in reality much more comfortable.

Frankly the B&C DE-400TN-8 has risen to the front rank of my favourite HF drivers. It is probably my favourite current production high frequency system; I could sacrifice 6dB of its amazing sensitivity for an extra half-octave at the top and many SET tube-heads would sacrifice power handling for an extra octave at the bottom to replace those Altec VOTs if B&C wish to dominate that market, but as it stands the B&C DE-400TN-8 is an amazing new design that operates in a class of its own.

B&C DE35-8

The application requirements of the Hammer Dynamics Super 12 brings us neatly to the B&C DE35-8 bullet tweeter.

The splendid multicloured contour map of dispersion at the top of this page is that from the DE35-8. This is much better than those B&K circular radiation distribution charts of 30 years ago that we used to pore over and have to guess what the charts might have looked like for frequencies other than the 2 or 3 spot frequencies offered. However, those circular charts are, in themselves, infinitely superior to anechoic frequecy response graphs with a second trace for 15° off horizontal axis and a third for 30° off axis. No designer can ever predict, or even make an educated guess, at what a speaker system will sound like in a real room unless he or she is equipped with good quality data (data quality, not sound quality), which is presumably why B&C provide a proper dispersion plotted against frequency chart for the DE35-8. Every drive unit manufacturer should provide these, IEC baffle responses and T/S data as an absolute minimum; they should also provide transfer characteristic charts too but I haven't seen these since Richard Allen in the 1970s. This is what B&C do tell us about the DE35-8 (a much more comprehensive set of data than we get offered by most PA driver manufacturers so B&C are ahead in my book already).

Once again the '8' suffix refers to the 8Ω voice coil of this version, with a valve friendly R_min of 7Ω. Bullet radiators are usually the supertweeters of the sound reinforcement world. They are designed to be hot and add sparkle to the wall of sound emanating from large frontal area stacks. They have VERY tightly controlled dispersion and are usually arranged in curved arrays to provied the coverage needed for a particular venue. When you next attend a big stadium or arena scale gig, look up at the glinting row of aluminium bullets on each stack. That's where all the cymbal shimmer and stick noise comes from.

The B&C DE35-8 are specified at 108dB/W. On paper that's just 2dB hotter than the DE-400TN-8, or the latter would need almost 2W to make a similar noise as the bullet makies from a Watt. But these babies seem even hotter than that, straight after the DE-400TN it feels like 3 or 4 more dB fr4om that first Watt. Knowing the Hammer bass-mid units are supposedly rated at 98dB/W, and my experience in the Hammer system with the Hammer crossover (Litz inductor version) I reckon 95-96dB/W is a more accurate description of overall passband sensitivity (more useful to know than spot frequencies at the top of the inductive upslope of big magnet drivers) so I soldered a -12dB voltage divider that kept the resistance at crossover frequency the same as the Hammer Audax tweeter. This still sounded too hot once the tweeter had 50 hours on it, so much so that I found my son stretched socks over them in a desperate attempt to make them tolerable; rather like the old kleenex studio trick.

Subjectively selecting different resistors in the network until it sounded right, the network was dropping another 2 dB. Every home experimenter should keep a box of four each of 1.8Ω, 2.0Ω, 2.2Ω, 4.7Ω and 5.6Ω in 2, 3 or 5W for this kind of fine tuning (keep all the old ones you take out when upgrading commercial crossovers). The DE35 is recommended to add sparkle to guitar amp cabinets, although I prefer more smaller drive units driven hard for a trad sound than big drivers with hot tweeters, which is just far too 80s for me. Even while tweeter level was elevated less than 2dB the sound was still totally unintegrated, like two separate drummers, one playing skins and the other playing cymbals standing a couple of metres in front.

Thus it proved much more difficult to integrate the B&C DE35-8 from a pressure response persepctive. However, as soon as it was run in (an even longer process than the DE-400TN-8) it offered a much more even power response in the Hammer system, through the Hammers crossover region, accross the width of the room. With the B&C DE35-8 there is a definite, well controlled, hot spot on axis from crossover to roll-off that matches big directional drivers well.

On paper the DE35-8 ought to be more ecxtended at the top end (both are specified at only 18kHz) but the DE-400TN-8 seems to go higher (yes at 50 I can still hear beyond 20kHz even if the level is subjectively now well below 10kHz at the same SPL). I suspect the downslope of the DE35-8's response is responsible for this effect, lowering the highest frequencies by a subjectively greater amount because in terms of response at crossover the highest frequencies are an extra 2dB down.

At first the DE35-8 seems somewhat rougher than the DE-400TN-8 implying higher levels of distortion. However, it is a function of this horn shape that there will be a jagged frequency response towards the cutoff frequency and extended listening with the correct attenuation padding bore this out. For example, distortion did not seem to increase with volume, remaining constantly in proportion to the signal even when it became uncomfortably loud. This reduced a little over time as the units broke in, but the DE35-8's retained their 'hot' character thoughout the duration of the review. This is much less obvious in comparison with something like the original Hammer Audax soft dome than compared with the Focal titanium inverted dome or the supremely low distortion DE-400TN-8.

The obvious two suspects contributing to this problem are the 5dB response peak at 4kHz and the 3dB peak at 5kHz which will produce significant output with anything less than a 24dB/8ve crossover at the manufacturer's suggested 5kHz crossover. Indeed, and LCR shunt accross the driver terminals might be in order here (although I did not try this option myself). Secondly, the mylar diaphragm may contribute, being substantially less pistonic than the titanium of the DE-400TN-8 (and I prefer titanium over mylar equipped direct radiators too) and I would love to hear a version of the B&C DE35-8 equipped with a titanium diaphragm.

With a big driver like the twin cone Hammer 30cm bass-mid unit and a crossover point anywhere above 1500Hz, for a 250mm driver 2200hz, for a 200mm driver 3500Hz a controlled dispersion tweeter is likely to have a more even room integrated power response in the critical midband than the sudden lurch to near 180° tweeter dispersion from half that cone size from the bass-mid unit. This is the experience with the Hammer twin cone units whose wizzer cone enables response to struggle up to a claimed 8kHz and the DE35-8 integrates beautifully above this.

The B&C DE35-8 is a niche product in the domestic audio world, being less universal than the DE-400TN-8. Where either might work, the DE-400TN-8 is more accurate and transparent, but where a higher crossover frequency of above 5kHz is possible and tight dispersion is demanded the DE35-8 will come as a pleasant surprise to domestic users. The B&C DE35-8 is an excellent controlled dispersion ultra-high sensitivity tweeter.

Conclusion

The B&C DE-400TN-8 has quickly established itself as my favourite high frequency driver currently in production. The low distortion and accurately controlled dispersion easily overcome the hassle of a driver that doesn't fit in the same hole as your old soft dome. The Decca London Ribbon horn has finally met a joint favourite. The Decca's tinkly distortion and greater extension remain seductive but B&C DE-400TN-8 has lower distortion, is more robust, 6dB more sensitive and occupies half the baffle area. The B&C DE-400TN-8 is a clear success as a low distortion, low colouration, domestic nearfield audiophile tweeter and should therefore have no problems fulfilling its design role in the pro-sector providing it is sufficiently indestructible.

The B&C DE-400TN-8 even holds its own against the best direct radiators like the Focal inverted dome series

The B&C DE35-8 is of much more limited appeal to the domestic market. It is ideal to add some zest to a PA rig and would be ideal for top-end tizz in a nightclub specialising in drum&bass where accurate treble is a must. In domestic audiophile settings its main role will be supplementing BIG directional drivers like the Hammer Dynamics (where it proves a drop-in replacement (with resistors) superior to the presently recommended options and now that I have found a tweeter that actually works well in the Hammer Dynamics system I will finally publish my Hammer evolution experiments in these pages. It is not a drop-in Hammer tweeter as it is too big to sit in front of the main driver, so it must be a drop-out replacement. Yeah, tune in, turn down ('tis 108dB) and drop-out if you have a Hammer system or similar. The B&C DE35-8 would also be ideal powered by pre-amp type valves in a multi-way active all-valve system; above 5kHz 1W should be enough in anyone's sitting room.

These B&C drivers demonstrate once again that the low distortion design goals of the pro-sector are valid in audiophile settings where obsessive pursuit of flat steady state frequency response holds sway. The dynamic range of the DE-400TN-8 reinforces (pun induced groan) my criticisms of most direct radiator tweeters, the powerful 1.8T magnet, copper shorting cap and titanium diaphragm all justify their cost here. The 1500Hz cut-off would enable 2-way designs with big bass-mid drivers and 106dB sensitivity would allow this with flea powered valve amplifiers; how about active crossovers feeding a treble amp with an output-transformer-less array of big cheap old triodes like 6080 or 6N7 anyone?

The B&C DE35-8 is my recommendation for high sensitivity tight dispersion control applications like the Hammer Dynamics system and the B&C DE-400TN-8 is now my favourite current production tweeter. Now I understand why some enthusiasts choose to build those immense multiway horn systems so beloved by the Japanese SET enthusiasts.