Thursday, 19 December 2013

A Tannoy 15'' Enclosure for the Enthusiast – Part IV

In the last couple of month I have tried to find a solution for the incomparable 15'' Tannoy driver to be used in a sort of universal enclosure. I concentrated on two nested main questions, the best possible response coming from a living room friendly sized cabinet. My main aim in this process was, independant from any technical design principle (i.e. reflex, open baffle, sealed, horn development), to archive a foremost natural tonal response. It is best described as a transformation of the electrical signal into a acoustic representation of music, lacking any typical electrical significance or sort of common compressions arising. My ideal response is orientated as close as possible to live performances of small scale chamber music or small jazz ensembles in limited scaled room conditions. Here the acoustical resonant corpus of the performing instruments (i.e. cello, brass horns, piano, etc.) is quite similiar to the resonant corpus of the performing transducers (loudspeakers). My definition and further undertakings can be understood by the most of yours within this model in mind.

Typical Tannoy Lancaster cabinet (sealed version for the 12'' driver),
these generation cabinets are well made from lumber core plywood, excellent material.
Some month ago at the beginning of the design process I did not know, with what sort of cabinet I will come out at the end. I did play through in my mind many possibilties, a combined back/front loaded horn design, a vented port cabinet with 200 liters volume or even a bigger sealed enclosure, working as a resonant chamber. I definetly wanted to leave behind the limitations of the common known vintage cabinets. Instead I was very energetic to find new ways to the naturalness arising from a newer understanding of the loudspeaker to be a resonant musical "instrument". In the last decades my understanding of "true fidelity" in conjunction with single audio components has been widened by a general respect of each components resonant properties, where the speakers show the most suggestible influence of all of them as to be the final transducers.

There are the well known designs of similar speaker cabinets made in the past: some Altec designs like the VotT-line, the japanese Onkens, JBL, EV, Jensen and some other professional small scale pa-addressed designs for 15'' diaphragm sizes. They all have one thing in common, to be originally designed for a different speaker type in mind than the Tannoy Dual Concentric. The uncomparable design of the Tannoy offers opportunities other speakers don't show and these should be respected to find a perfect matching solution. And finally I did want to share my experiences with other Tannoy addicts, because those don't have a lot of choices of cabinets in our culture (may be in Asia, foremostly in Japan there are options for after market enclosures).

Tannoy Amesbury with 15'' Monitor Golds
I think nearly 90% of all vintage Tannoy 15'' Dual Concentric speakers are installed in the typical 120 liter volume ported box. In history it all started with corner cabinets for monophonic reproduction, the huge ones were horn designs the smaller acted like vented port enclosures. Both types used the backward radiated energy of the cone to be conducted as a supplementary part to the air. The horn used the flanking walls to extend  its development. Bot design principles are made to reduce the inherent limitations for the lower frequency response of a boxes loudspeaker. With the need for stereo pairs, the range of enclosures was expanded to rectangular versions like the GRF, York and the Chatsworth. These could be better placed in the room for a improved stereo illusion. Additionally Tannoy started in the 1950ties a small universal cabinet line to fulfill the needs of their customers to smaller cabinet sizes, used for the 12'' and 15" units simultaneously. This design got widely sold for decades as Lancaster and ended up in the mid 1970ties as strengthened Berkeley of the same size. A vented port in the beginning and a true reflex after Thiele&Small at the end was the most well sold cabinet design for Tannoy of all times. With a size of 85cm height to 52cm width and and a depth of 28cm, the success of this enclosures represents for lots of people already a huge speaker. This size seem to be as well a maximum dimension accepted as a piece of furniture in living rooms from its users. This enclosure limits the performance of a 15'' chassis a lot, in particular with the hard edged versions (Gold, Red and Silver) the backwarded compressed air cushion show a extremely limitation of possible dynamics. A few of us have got luck, their partners do even accept the little bit bigger versions, the Rectangular York, i.e. Amesbury which ended in the late Arden enclosure. With a useful volume of 180 liters, this cabinets support the technical conditions of the chassis much better, but are still more a compromise than a solution for listening pleasure. The hard edged drivers have a theoretical Qt which will lead to a ideal volume of 350 liter for a 30 Hz response with -3 db fall. With this parameters it starts that the speakers are getting almost invisible in acoustic terms and show a good dynamic performance down to a low end with fast transients. The most users cannot accept two wardrobe sized cabinets in their living room in order to match these ideal conditions for the drivers.

The biggest enclosure of all vintage Tannoy back loaded horn designs, the Tannoy Autograph Professional with two 15'' drivers in parallel, very poor material choice: chipboard
Mr. Ronnie H. Rackham followed another theoretical ideal for optimized sensitivity and improvement in  response of loudspeaker cabinets. Horn designs use the backside radiated energy from the cone in order to transmit it through its development to an opening as addition to the complementary other half from the front side. It acts like a mechanical amplifier for the conducted frequencies and adds on the second half of the wave, so the sensitivity will be increased. The very early variants have been designed for the corner position in the room in order to use the flanking walls for the horn extension. Only the GRF Rectangular and the Autograph Professional were made for the plain wall position. The Rectangular was a quite small sized cabinet. Here the ideal horn development (around a lenght of 3 m with a continous opening to 2 x 2 m for 40 Hz) has been cut to be anymore unrecognizable sizes. With a length of 1,50 cm and a opening of 40 x 60 cm the horn acts like a lower middle frequency horn, with a cut off end around 400 Hz. This was my first 15'' Tannoy speaker and I never liked its response. With its hefty frequency belly around 400-500 Hz, this cabinet has set a fundamental dubiousness about horn designs in my very early experience with 15'' Tannoy speakers.

Inside the Tannoy Autograph, the complex development of a back loaded corner horn shown at a set of japanese after market copies (plywood).
The today operation of the less compromised horn designs (Westminster, Autograph, GRF Professional) makes a dramatically increased acceptance necessary, as to be the speaker a room dominant piece of furniture. Beside the Westminster I do know almost all vintage Tannoy enclosures very well, since I have or had the most of them myself, or some of them are installed at close friends audio set ups. To be honest, not one of the bespoken designs matches my current expectations about a natural sounding music transducer. They all show a unacceptable part in the lower frequencies were folded horn brings a quite uneven response. The positve benefit of all horn designs is the undistorted way the driver can act like in free air resonance, without limiting air cushion and its fast transient response.
Ideally I want a performance with realistic and fast dynamic transients, a huge and holographic soundstage, with refined and detailed smooth Hf-response (without any phase shifts and resonant knots), a similiar performance like a natural wooden instrument of comparable size. A horn design at it best comes very close but will be extremely huge and room dominant. The other possibilty to leave the limitations off a boxed speaker is the folded baffle principle. To make it work, it needs a lot of acceptance as new approach of understanding of the role of cabinet (new was meant to be connected with Tannoy speakers – Western Electric and Klangfilm did incorporate such design principles about 70 to 80 years ago).

Chinese copies of the Westminster for 12'' and 15" drivers. They do copy everything.

I definitely search a solution to use with the superior hard edged chassis from the Monitor Gold, Red or Silver line in conjunction with low power tube amplification. These hard edged drivers need naturally a bigger enclosure to work superior to their later derivates from the HPD line. The limitations of the early cones depend hardly to the physical parameters shown at their low Qt of 0,26, but their sonic advantages are a close mirror of these facts. The nominal 15 ohm (at 1000 Hz) voice coil (Red, Silver and Black) was the preferred VC-impedance when tube amplification was the standard. This impedance matches the plate resistance preference of a pair of tubes and even more of a DHT-Triode better, the different coild at the output transformer to match are only a little help, since impedance shifts over the total frequency spectrum of the speaker can not be covered here. In particular in the lower frequencies the hard edge surrounds have helped tube amps a lot to control the cone movement. All this results is a increase of fine dynamic details and a better resolution of finest overtone harmonies. The light weight paper cones have a much lower mass to be driven and can therefor much better controlled by a given tube power amp, resulting in a improved sensitivity. The heavy HPD cone (double weight through the stiffing back support brackets and the three layer high power voice coil 85W) does not only make the double power output possible, it needs it it to be adequately driven.When the Monitor Gold was changed already for the establishing transistor amplifiers, the highly maneouvrable HPD cones are a clear reminiscence of the adaption to the high power output transistor amplifiers of their time. When used with tube power amplification, in particular with low power DHT tubes, the HPD drivers show mirrored characterics of the former chassis. All the live, color and speed in lower mid frequencies is gone, to me it sounds quite uninspiring and retarded comparing it with the earlier drivers. Within this drivers the front loaded funnel cannot add any extra quality to the soundstage.

The well known front loaded funnels from the Autograph and Westminster cabinets are a perfect extension for the 15'' dual concentric driver. Such a design helps fundamentally to conduct the cone radiation to the air of the listening room. In theory it works like a pantograph with a adjusted transmission ratio. The finer physical resolution of the stiff light weight cones completes the full potential of this design feature. A clearly listenable quality advantage of wider spread tonality and much improved micro dynamic response are the obvious benefits. In a horn design principle the cone almost works at its physical optimum, since it is not retarded by the air cushion of a closed cabinet. A further fine adjustment of the crossover design will complete this positive effects.
My personal experience was once initialised by the classic 12'' chassis (Monitor Red and Silver), where Ronnie H. Rackham set the Lf-cone as a fullrange unit. Later with the Monitor Gold line (1966) Tannoy limited the 12'' inch chassis performance with a 12 dB filter cutting above 1500 Hz. This intervention brought a flat measured response in the 1000 Hz region, but it can be easily approved, that it takes out a lot of micro dynamic detail. To my taste it sounds slower and a bit uninspiring, so I do like the more lively presentation in full range mode a lot better, even if there might be some little uneven inaccuracies in the measured response. Starting from this experiences and accustomed over the years with every day use, I missed the benefits of the full range driver when I restarted to listen to the 15'' chassis (Gold, Red and Silver) some years ago. Together with my first test of the front loaded funnel I changed here the crossover into full range mode. This intervention is a basic principle for the following incorporation of the front loaded funnel with greatly improved success. The Westminster and Autograph have as well verified (lower) crossover points in conjunction to the front horn opening. I tested both filters as well in my set up and decided to stay with the unfiltered full range mode for my 15'' Monitor Red drivers as to be a superior solution.

With the rejection of the filter for the Lf-cone even with the bigger 15'' units, two important advantages come together. First the sensitivity gets a better ratio. I am sure that the cone gets into trouble at the upper end of its physical ability. It will show a good portion of deformation in the measured response curve, this starts around 4000 Hz and peaks around 8000 hertz with already 6 db or even more loss, so it does not have a fundamental effect to the overall sound. On the other hand the uncontrolled cone plays so much better in terms of dynamics and speed in the important 1000-2000 Hz area, that I can accept the effect as almost completely positive.
As already said in other articles, the concentric high frequency horn is cut at 1000 hz with a 6db filter in the ordinary crossovers. Listening with the front loaded horn funnel shows a dramatic better distribution and dispersion of fine detail and micro dynamics in this most important tonal range around 1000 Hz, due to its acoustic coupling from 400 Hz on to the higher frequencies. It gives a 100% better front radiation of the cone and helps the high frequencies to be supported into the lower registers. Here almost every instrument or any voice will be covered with finer micro dynamic abilities. The overall soundstage is dramatically improved in its illusion of width and depts and the impression of air is noticeably extended, so this change has a fundamental impact with extremely positive effects for my new design.
Another word about measurements. They can help to find some problems, they can shorten evaluation proceeds, but their graphics doesn't say anything fundamental about sound qualities and their harmonic structure. If you look to graphs of measured natural instruments, they show a tremendous deviation from the idealized flat response. It is almost a opposite fact, the most colorful tonal presentation with the best harmonics in overtones, show the most degraded graphs. I don't think that the publishing of graphs shows any content about musical response, so I don't show them.

"Copy and Paste" in China. I do not know what these copies will cost, nor do I know where to get them. I do not know if they will even match my high expectations for tonality and natural musical response...but they are made from plywood.  I wish to have a chance to listen to a pair of these?!
The original two versions don't differ in size but in material choice of their boards and used drivers. The Royal (with HPD derivate driver) is made from plywood, since the TW (Tulip Waveguide, speakers with horn exposed dome tweeter and with ferrite magnet) is poorly made from chipboard. 
I would state, that without the support of the front loaded horn the Tannoy does only do half the job and is a different speaker. After two years of implantation in my set up, I don't find back to the plain ordinary response with typical turning point filter at 1000 hz in the classic crossover. I have tried several times to switch back (with different cabinet designs), for me it is very easy to do, I have built a switch into my network to be operated from the outside of the enclosure. I tried again and again with different equipment and in different rooms – I always tried with and without the funnel – no way, I cannot accept anymore the classic approach.
It is now a different loudspeaker to go. Try yourself, the intervention into the network, its easy to be done, just bridge the capacitor and the coil at the Lf-rail with a crocodile clip wire, – done. For the funnel to test, I can recommend a simple version made from 2mm grey card board as a first stage. This will show almost all the benefits with your existing speaker enclosure. If you have Lancaster cabinets, remove additionally the back panels and position the speakers almost 1 m (4 ft) as minimum distance in front of the backing wall. I swear, you never have heard such a good sound from your Tannoy speakers, if your chain is able to supports it. Don't fear this step, it will be reversible in 5 minutes... And if you do, please ... let me know what you think about!

Read on soon about constructing the new open baffle enclosures in the Christmas holidays, Volker

All photographs shown from Google with respect to their  unknown original resource.