What's Inside?

Introduction The following pictures are of a 15 ohm Chartwell Project Symphony LS3/5a. Other manufacturers' versions, including ones made in the Chartwell factory are similar but somewhat tidier than this example. These pictures can also be viewed with reference to BBC Research Department Technical Report RD 1976/29 and A Little Legend, The BBC LS3/5A. Which was written by Trevor Butler and originally published in HiFi News and Record Review in January 1989. The LS3/5a was manufactured by several licensees during its lifetime including Rogers, Spendor, KEF, RAM, Chartwell, Goodmans and Harbeth. By 1987 too many B110 SP1003 drive units were falling outside specification and so KEF reworked the drive unit. This required a redesign of the crossover circuit. These two versions are usually known as the 15 ohm LS3/5a and the 11 ohm LS3/5a.

A Chartwell "Project Symphony" LS3/5a. The design of the LS3/5a became very popular with BBC staff and many of them wanted pairs for use at home. The speaker was never cheap, tight manufacturing tolerances demanding that all the components be selected. Arrangements were made with Chartwell to supply cabinets, crossovers and drive units in kit form allowing staff to assemble their own speakers. Project Symphony kits were never intended to be available to the general public although many pairs were bought by staff on behalf of their friends. The price was £100 plus purchase tax per pair. Provided that they were built carefully Project Symphonys should sound the same as their factory made counterparts.

The original Chartwell instructions for building the Project Symphony. Click on the image for a larger view. (With thanks to Dave Plowman).

The drive units chosen for the LS3/5a were the KEF T27 SP1032 tweeter and KEF B110 bass/midrange driver.

The 15 ohm LS3/5a used B110 version SP1003 and the redesigned LS3/5a, usually known as the 11 ohm version, used the B110 version SP1228,  All drive units were preselected selected at the KEF factory.

Examples of off-centre B110 drive units are not uncommon and may possibly be "cured" by running the speaker upside down or turning the B110 on the baffle. It is well worth checking for this particularly if buying a used 15 ohm pair. The BBC selected SP1003 drive unit used in the 15 ohm version is virtually unobtainable. So there are many examples of 15 ohm LS3/5as where the drivers have been replaced either with un-selected B110 SP1003s pulled from other KEF speakers, or with the new version B110. In either case the balance of the LS3/5a is affected and it will fall outside original specification.

LS3/5as are balanced for use with the Tygan grill in place. The T27 has a square felt surround to prevent interference effects associated with the cabinet edges and the tweeter dome is covered with a metal grill for protection sourced from a Celestion HF2000. This grill has a small effect on the frequency response which was considered beneficial as it raised the output a high frequencies.

The Cabinet

The cabinet external dimensions are 31 x 19 x 16 CMS giving a internal volume of close to 5 litres. The cabinet walls are made from 12mm (0.5 inch) selected birch plywood with internal fillets from beech. Because of the critical nature of the design it was found that the substitution of other wood for the walls or fillets caused clearly audible changes. The use of any hardwood for the beech fillets produced a audible colouration caused by the resonance of the B110 on its chassis. The inside of the cabinet walls are lined with car body damping panels, a single layer is used on the side walls and two layers on the top and bottom. All internal surfaces except the front panel are lined with polyurethane foam; 25mm thick on the top & bottom and 16mm thick on the sides. The cabinet, including the screw holes, is carefully sealed

The 15 Ohm Crossover FL6/23
Because the LS3/5a is a passive design, bass extension can only be achieved by "throwing away" sensitivity at mid and high frequencies. This accounts for the very poor sensitivity of the design.

Derek Hughes kindly supplied the following information about the value of the tapped inductor, L3:

Overall inductance for the autotransformer = 0.64mH (ground to tap 7)
The other tap values are approximately:
Tap 6 0.5mH
Tap 5 0.38mH
Tap 4 0.3mH (this is also the feed to the output)
Tap 3 0.22mH
Tap 2 0.165mH

Derek also told me that Spendor 15 ohm LS3/5as did not use a tapped inductor because they tightly selected the drive units. I presume then that the Spendor inductor was 0.3mH to ground with the input and output across the top of it.

The crossover shown is model FL6/23 made by Chartwell which was in use in 15 ohm models until 1987. At this point in manufacture all the licensees made up their own crossovers and so variation in printed circuit board layout is seen, although all 15 ohm LS3/5as use the same circuit.

The response of the B110 is equalised by the 1.53mH inductor and 82R resistor. The 6.2uF capacitor, 2.67mH inductor and 33R resistor compensate for the hump in this characteristic. The crossover frequency is 3kHz. L3 operates as a shunt inductor for the T27 tweeter and is adjustable to allow matching of different sensitivity T27 and B110 units.

C2 is adjusted on test to keep the crossover frequency constant. Chartwell suggest a value of 2.0uF for tap position 5 in the Project Symphony instructions but the original BBC designs varied the value depending on the position of the L3 tap:

Tap 2 C2a - 2.2uF (Type 1) C2b - 0.47uF (Type 3) Tap 3 C2a - 1.0uF (Type 1) C2b - 1.00uF (Type 1) Tap 4 C2a - 1.5uF (Type 1) C2b - ------ Tap 5 C2a - 1.0uF (Type 1) C2b - 0.22uF (Type 4) Tap 6 C2a - 0.47uF (Type 3) C2b - 0.47uF (Type 3) Tap 7 C2a - 0.68uF (Type 2) C2b - ------

C2 to be made up of the above combinations of C2a and C2b in parallel.

Type 1 - Polycarbonate, 63Volt +/-5%
Type 2 - Advance Filmcap 160 Volt +/-5% CTD 15
Type 3 - Advance Filmcap 160 Volt +/-5% CTD 10
Type 4 - Advance Filmcap 160 Volt +/-5% CTR 6

The crossover PCB is mounted behind the tweeter with a piece of felt between them to damp board resonances. Note the crossover tap (L3) which was adjusted on test.

R4, value 8r2 runs very hot. Examples are common where the polyurethane foam on the adjacent cabinet wall and the nearby crossover capacitor have melted. Changes in the value of R4 due to heat ageing are very likely and may be expected to affect the treble. In extreme cases total meltdown of the HF crossover components can occur if a LS3/5a has been fed with prolonged high level HF energy. The crossover shown was removed from a 15 ohm LS3/5a used in a professional studio where muting was not applied during fast tape spooling. The tweeter had gone open circuit and the heat generated had melted all the capacitors and burnt out all the resistors in the HF section of the crossover. Extensive scorching of the PCB is evident.

The 11 ohm Crossover FL6/38 SP2128

After the design was revised around 1987 all manufacturing of the LS3/5a crossovers was undertaken by KEF. This was because when a speaker fell out side specification there had been disagreements between KEF and the licensees each blaming the other for the problems! With KEF supplying both drive units and crossover this source of conflict was resolved. Greater consistency of B110 manufacture and a smoother response meant that the new crossover could be simplified, eliminating the costly autotransformer used in the older version. Matching of relative sensitivity between the B110 and T27 was now achieved with a resistive ladder. The new crossover made bi-wiring possible and although LS3/5as were always used by the BBC with the terminals strapped, this made the speaker more attractive to consumers.

Back to contents page