Another Look At The Wharfedale 5 incher


A question was asked of me as to what I used for4 a replacement annulus for the Super 3 and the 5 inch units.  Actually, a few materials have been used.  Open cell foam, much like the original, a much denser foam, almost like rubber but porous and felt.  All three have a thickness of about  1/16th inch, 1.5mm.  

By themselves, the dense rubber type foam proved to have the advantage of preventing doublet action if the unit was mounted in a sealed chamber.  Doublet action occurs when the wavelength is greater than the diameter of the cone, (actually, it may be 1/2 wavelength as seen in the red trace of Fig 2. This might also account for the peak centered around 2600 hz) after which the front wave being 180 degrees out of phase with the rear wave will begin to cancel each other.  This is the reason why a woofer can't produce bass in free air.  For a 12 inch woofer, doublet action will begin around 1300 hz.  The rolloff is about 6 dB/octave, so at 40hz, the output is about -30 dB below the output at 1300 hz.

For the Super 3 and Super 5, the rolloffs begin at about 2722 hz and 1756 hz, respectively with an original open cell foam surround.

In the W4 and the later W90, there are two 5 inch units in an isolated chamber.  The units in the W4 are, if memory serves me well, mounted to the baffle while in the W90, all 4 units are in the same isolated chamber but are free floating, so to speak.  A better description is in the open.  If the peak centered around 2500 hz is present, (red curve, figure 2) it may explain why so many like the presence of the W4 and the venerable W90.

In the W90, the3 chamber housing the two pairs of units measures about 28" by 6" by 6".  The 6 inch dimension corresponds to about a 2300 hz wavelength and the 28 inch dimension 486 hz wavelength.  The height and depth being close to each other will also exacerbate the response around 2300 hz; those 6 inch dimensions being approximate.  That peak, if it exists in the W90 could account for the claims of its sonic presence.

In another experiment, The Wharfedale 5 inch unit is compared to the much later (1990's) French Focal 5K013L, of which I have 12, eight being in use in two systems.  The result was quite astonishing, enough to build a pair of cabinets and do an A-B listening test.






The test setup.  The enclosure for the speaker is lightly stuffed with pure wool.  This setup was used for all gated response measurements at 1w1m.





The rear of the W5 unit.  The white coating on the felt annulus is RTV, an uncured rubber supplied in a tube at many hardware stores.

The 10.83 ohms marked above the label is the voice coil DC resistance.





The front of the same speaker with the felt annulus.

The felt comes in many colors of which I have black, green and purple.

The purple felt sounds better.








RED:  foam annulus

GREEN: felt annulus (no RTV)

BLUE: felt annulus with RTV (the resonance has increased by about 30 hz)  Very little RTV was applied; just enough to seal the felt. 








GREEN 56: dense foam surround (in box)

PURPLE 57: untreated felt (in box)

BLACK 53: felt sealed with RTV (in box)

RED 64: free air (with the felt sealed, the acoustic doublet will begin when the half wavelength is about 5.25 inches, about 1300 hz)




The following is a comparison between the stiff foam (left) and the RTV sealed felt (right) at the frequencies indicated above each photo.  They are spectrograms taken on a PicoScope Model 2205A using a Behringer ECM-8000 microphone at a distance of 1 meter from the speaker.  The microphone output is amplified by an ART Tube MP Studio amplifier. The scope output is fed into an ADCOM GFA-535L power amplifier which drives the speaker.

The speaker is mounted in a sealed box of 348 in^3 and is driven with 1 watt (ref. 1khz).  The SPL at 1 meter is 88 dB.

The sum of all harmonic distortion introduced by the test equipment is of the order of 0.4%, far lower that that of most, if not any loudspeaker and is common to all spectrograms.  The purpose here is not to measure the THD of the speaker but to compare the effect of different annuli.

The fundamental is the longer spike at the left.  The first (left) vertical grid is 5 khz.

Overall, considering all the spectrograms, it appears that the much stiffer foam annulus does better at the higher frequencies, above 800 hz.  Looking at the first two pairs of graphs, there seems to be a lot of action between the 400/500 hz fundamental and 5 khz, probably due to the elasticity of the annulus and the fact that it's in the resonance band.  At 500 hz it's just at the edge iof that resonance band, above which it is gone. See Figure 1.  The felt on the right does not show this.  My best SWAG (Scientific Wild Ass Guess) is that the felt annulus has little effect on the cone motion.  The purpose of the annulus is to prevent the cone from sagging; it keeps its axis perpendicular to the magnet assembly and should not have any other effect.

It should be noted that this experiment was not intended to determine the better annulus but only to show the difference.  The determination of which is better is left to the listener.  Many like and even prefer the sound of vintage loudspeakers and for many reasons, even nostalgic reasons.





All the action up to 5 khz may be due to the elasticity of the dense foam triggered at resonance.


The felt annulus wouldn't react at resonance as does the dense foam, hence no continuous cone movement due to the elasticity of the dense foam.



The same logic applies here as was postulated previously






800 hz is well above the resonance band between 100 hz and 400 hz hence the lack of uncontrolled activity at the left.



The left here is slightly cleaner then with the dense foam.  There is a lot of harmonics above about 5 khz probably due to the lack of damping of the felt annulua.



From this point on, there isn't much difference between the foam and felt.



It's somewhat of a trade-off.  Then, few will take a 5 inch unit much below 1000 hz, typically 800 hz for the lower end of a mid-range.




















Back to the Wharfedale Index

Back to the loudspeaker main page