Behind the wood. Now we get into some Vox 3 tech.

"The two 5" woofers are the usual high-standard Galm Audio drivers. Ours were made to Chris' specs with carbon-fibre membrane, 37mm 4-layer voice coil and neodymium motor. These voice coils are complex beasts that require very tight tolerances. Armin wasn't that keen on them mainly due to their low 84dB sensitivity. Chris insisted on going down the force-canceling back-to-back route. 'Has to be a proper job' were his words. By doing this, we also benefit from the 6dB lift of paralleling two drivers. This brings overall sensitivity up to around 89dB.

"These woofers are on our usual 2ndorder filter and come in around 200Hz in a sealed compartment with two of the 10cm² ceramic pads for some velocity damping. As you know, I'm not a fan of that soft fluffy stuffing which seems to go all woolly. The port is tuned to 35Hz.

"The TangBand is off the shelf and has a bamboo membrane. We run it fully open at the bottom but it does need a few notch filters and a gentle roll-off at around 10kHz where it hands over to the small Raal 64-10 ribbon. This again is just on a 2nd-order with a single cap and I had Raal adjust their transformer windings so we don’t need to use any shelving resistors. It also has a custom 2mm bronze face plate."

"So all in all, we managed bandwidth of 35Hz – 35kHz which is pretty amazing for such a small speaker. The front-facing ribbon version sounded very good but with two side-firing woofers, the spectral and spatial image was LF-centric. I felt that it needed a similar halo effect in the HF region, hence the upfiring position. It also cleaned up our cosmetics where two drivers on the front looked a bit cramped. As mentioned, we're still working on the custom widebander. I'm just missing that bit of extra magic an AlNiCo motor can bring to the presentation. The prototype with bronze basket in these photos is something Armin cobbled from parts he had. It was never intended to be the final version but simply a development platform. However, sonically it proved to be exactly what I was looking for.

"The main issue we still have is the small 4" diameter-to-cone-depth proportion. It seems to create nasty dips in the 3-4kHz region. Initially I had a Chinese-made wood cone driver which I was quite fond of but which, due to the same issue, we finally had to ditch. We're now experimenting with a shallower cone profile so finger crossed. I am also looking at various papers for the cone. I have some high-quality Japanese Washi and papyrus papers on my scope. Ive been through the real wood veneers which are notoriously difficult to control on such a small cone. The irregular fiber structure has forces pull in all sorts of directions. Other people have been there already but hey, if you really want to know, you have to have a go at it yourself." Behind that casual last sentence so easy to pen hide time, money, many dead ends aka abandoned experiments and a renaissance man's curiosity over the big picture which wants to cover all the bases.

Asked what he views as AlNiCo's advantages, Martin replied "in short, all AL I've come across seemed to have a smoother more singing quality than Ferrite and Neodymium. Like the guy below says, 'scooped sound, sort of fuller, richer and more flowing'. One track I always play is Gianmaria Testa’s "Milla giorni" from his Testa at work album. It has this guitar intro that really brings out the Alnico twang as I call it. To me it just has a magical feel. The following comment 'the voice coil moves less' certainly seems to be true with the Enviée driver which has an Xmax of ~3mm. I of course wondered why an AlNiCo magnet might have that edge and found this. Pretty interesting stuff which confirms some of the comments we've all heard about AlNiCo's better frequency extremes and being able to play louder without distortion.