BBC LS3/5a loudspeaker 1989 Follow-Up

A Follow-Up appeared in March 1989 (Vol.12 No.3):

In last month's review of the venerable LS3/5a loudspeaker (Vol.11 No.2, p.115), you will remember that I had intended to compare an 11-year-old pair of these BBC-designed monitor speakers against a brand-new pair. As explained in the review, this turned out not to be possible before the review deadline, due to the new pair getting lost in shipping. (As Murphy's Law would have it, they finally arrived just after the February issue had been put to bed.) The new version of the LS3/5a differs in detail from those manufactured before 1987, the supplier of the drive-units, KEF, having carried out a research program to render the design more consistent in manufacture and nearer the target performance originally specified by the BBC.

In particular, the increased uniformity in tweeter production meant that the auto-transformer in the appropriate leg of the crossover used to match the tweeter level to the woofer could be replaced by a simple constant-impedance resistive divider. This results in a slightly lower overall impedance, to an 11 ohm characteristic rather than the original's 16 ohms. In addition, the original Neoprene surround of the B110 bass/midrange unit has been replaced by a PVC formulation to reduce the amplitude of a persistent slight peak around 1kHz. As the target of the redesign was to produce a sound identical to that of the original, the LS3/5a designation remains unchanged.

To conclude my review, therefore, I carried out listening tests comparing old against new, both pairs being auditioned on 24" Chicago stands positioned well out in the room and driven by VTL 100W monoblocks. The differences were minor in degree but nonetheless noticeable. In particular, a slight nasality characteristic of the older pair was significantly lower in level with the new pair, resulting in a slightly warmer balance overall. Dynamics also seemed very slightly less compressed and the treble slightly sweeter, with less of a "fizz" apparent. To put these comments into perspective, when I tried auditioning one of the decade-old '3/5as with one of the '88 samples as a pair, what differences there were between the two did not prevent the speakers from performing quite well as a stereo pair. I have auditioned a number of speakers where there was more of a sonic difference between channels.

Fig.1 shows the impedance of the new version, which can be compared with that of the old version, fig.2. As can be seen, the latest version does have an overall lower impedance, with minima just below 8 ohms, but the shape is identical apart from in the three octaves above 7kHz.

Fig.1 Rogers LS3/5a, 1978 sample, electrical impedance (5 ohms/vertical div.).

Fig.2 Rogers LS3/5a, 1989 sample, electrical impedance (5 ohms/vertical div.).

Fig.3 shows the averaged, 1/3-octave frequency response of the two pairs of speakers, measured on-axis in-room at a distance of 1m. (The upper trace is the 1978 LS3/5a; the lower trace is the 1989 speaker.) Differences can be seen primarily in the treble, where the latest version lacks the slight prominence in the 1kHz region. The top two octaves are also hinged down a couple of dB compared with the older sample. The response in the lower frequencies seems pretty much identical, though the design's intrinsic 160Hz hump is perhaps a smidgen better-controlled. (Note that the dip in the 250Hz region on both traces is due to destructive interference between the direct sound from the speaker and that reflected from the floor between the speaker and the measuring microphone.)

Fig.3 Rogers LS3/5a, spatially averaged, 1/3-octave response in JA's room (upper trace is the 1978 LS3/5a; lower trace is the 1989 speaker).

[In 1996 I measured the response of KEF's final, bi-wirable version of the LS3/5a loudspeaker, intended to be identical-performing to the Rogers 1989 sample reviewed. The acoustic crossover of that 1996 version and its on-axis response, averaged across a 30-degree window on the tweeter axis, are shown in figs.4 & 5, respectively. The individual step responses of the two drive-units are shown in fig.6, revealing that the tweeter is connected in inverted acoustic polarity, the woofer in positive, the opposite of the Harbeth version I reviewed in 1993.—John Atkinson]

Fig.4 KEF LS3/5a, 1996 sample, acoustic crossover on tweeter axis at 50", corrected for microphone response, with nearfield woofer response plotted below 300Hz.

Fig.5 KEF LS3/5a, 1996 sample, anechoic response on tweeter axis at 50", averaged across 30 degrees horizontal window and corrected for microphone response, with nearfield woofer response plotted below 300Hz.

Fig.6 KEF LS3/5a, step response of tweeter (red) and of woofer (blue) on tweeter axis at 45" (5ms time window, 30kHz bandwidth).

To conclude, my experience of the latest version of the LS3/5a confirms last month's tentative recommendation. Still somewhat compromised concerning overall dynamics and HF smoothness and clarity when compared with such modern miniatures as the considerably more expensive Acoustic Energy AE1 and Celestion SL600Si, and having a distinctly tubby midbass, the latest version of the LS3/5a still has one of the least colored midbands around, throws a deep, beautifully defined soundstage, and has a slightly sweeter top end than it used to. At $650/pair, the LS3/5a (manufactured by Spendor, Harbeth, and Goodmans, as well as by Rogers) is worthy of a Class C recommendation in Stereophile's "Recommended Components." But if you have an old pair and are happy with the sound, I wouldn't bother changing to the new version.—John Atkinson

edgrr's picture

Is it worth trying to use my amazing old 15 ohm monitors with a new Panasonic SC-PMX70 (3 ohm) or should I just stick to the speakers provided?
I moved to Peru and will hold off on buying a "proper" audio setup til I go back to the First World...

David Rapalyea's picture

Someone should to put these components into a bigger box with a passive radiator and recreate the KEF 104.

I have both the 104 and the 104ab. I prefer the 104 on my solid state Marantz 1060 and the single ended Reisong A10 on the 104ab.