Wharfedale Super 8 - Wood Cone

 

This project started and ended with the idea of using a wood cone in the Super 8.  The purpose was NOT to improve the Super 8; the intention was to study the effect of a wood cone versus paper.  The Super 8 units available are easy to disassemble without damage to the cone and spider. 

In the beginning, it was noticed that a pair of papyrus cone 6.5" midrange speakers recently purchased had a noticeably different sound.  Tests were performed and frequency response was negated as the whole reason. A few other similar speakers exhibited much the same frequency response within less than 2dB variation.  However, a spectrum analysis told a much different story.  Impulse response tests are shown at the end of this diatribe.

A Tektronix harmonic distortion analyzer was used but stability of the instrument was difficult due to reflections in the room and any motion that would upset the standing wave pattern/reverberant field above about 1khz. The speaker was then compared with another of closely similar parameters using a spectrum analyzer and the result was astounding.  The spectrum analyzer was then applied to the Super 8, actually a few of them.  Two are completely original, one being a CS, cloth surround (annulus) and the other an FS, foam surround (annulus).  These two units also have the older phenolic spider which made disassembly easy without the need of solvents; the adhesive around the annulus is over 60 years old.  The foam surround speaker was used only as a reference as disassembly would certainly damage that 6 decades old foam.

After running tests on both the FS and CS, the CS was disassembled and a new paper cone installed.  This cone has a sealant treated corrugated cloth annulus, shown in Photo 1. The next 8 photos briefly show that process with short comments.

 

PHOTO 1

 

PHOTO 2:  Double spider speaker

The long thing is probably a wood chip but the reason for this piccy is the little fella who was on the dust cap.  It may be a baby brown recluse whose bite is worse than that of the black widow.  It's about the size of a grain of salt.  After taking the photo, the speaker was taken outside and a puff of breath sent the little fella into its natural habitat.

Actually, the part that centers the voice coil was and may still be called a centering device.  The term spider came from the appearance of the early centering devices, many being more elaborate than the one shown on this speaker.  Photo 3.

 

 

Left: The cone and coil assembly removed without the use of solvents.  Right: The label in remarkable condition.  This and the other original unit were probably never used as the paint around the mounting holes isn't scratched.  Way back then, these units were installed behind the baffle.
PHOTO 3

PHOTO 4

 

 

 

PHOTO 5

The spider was held in with duct tape as this is a temporary installation.  This presented no problem but securing the annulus did, as is seen and explained in the following piccies.

 

 

 

 

PHOTO 6:   A Slight Digression

Ahhh, duct tape, aka duck tape, the latter probably a misnomer until the Frog tape was seen.  I'm sure there's a Gorilla tape.  Since there are two tapes on the right named after animals, can we expect lizard tape, alligator tape, dinosaur tape or pterodactyl tape?

While the above may sound funny or even ludicrous, such tape, regardless of the name, can be used to temporarily fix just about anything.  Remember, that stuff saved the Apollo 13 crew.

Q.E.D.

 

 

PHOTO 7

The first setup didn't work well for the spectrum analysis, which was peppered with harmonics.  It was traced to the outer edge of the annulus which was vibrating against the basket rim.  More tape made no difference.

 

 

PHOTO 8

This quarter inch thick gasket clamped to the speaker helped but wasn't sufficient.  The taped annulus had no effect when running impedance as the voltage applied is very low and the speaker was facing up.  The weight of the cone was sufficient to keep the annulus edge from vibrating.  Had it done so, that would have showed in the impedance curve.

 

 

PHOTO 9

Problem solved.  The clamps were used because there wasn't enough room for the screws.

 

 

PHOTO 10

Heat treated with a hair dryer and manually flexing the cone at two diametrically opposite locations where the grain is perpendicular to the circumference.  This had to be done a few times to get the result seen in photo 12 (right)

 

 

PHOTO 11

 

 

 

PHOTO 12

Left: the cone just after forming.  Right: After heat treating. (photo 10)  The grain of the two attached pieces is perpendicular to the grain in the cone.  After purchasing two wood cone speakers, it was noticed that the cone is made up of two separate sections which may eliminate the effect shown on the left.

 

 

PHOTO 13

The almost finished product, unfinished due to the dust cap's not being glued to the cone.  At this point, the project is experimental; the speaker can easily be restored to its original configuration.

 

 

 

 

 

Description of the speakers involved

Figs 1 & 2 below show the impedance and response curves, resp.  The colours for each speaker are the same in both figs.

EB2 yellow;  EB3 green;  EB4 blue;  EB4 red (with new cone & coil);  EB4 black (with wood cone)

The EB2 has a cloth annulus and a canvas spider so it was used as a reference.  The cone and coil (copper) are original. This unit was not spectrum analyzed.

The EB3 is a Super 8 FS/AL with aluminum voice coil and the original foam annulus. It has a phenolic spider. Also used as a reference.

The EB4 is a Super 8 CS/AL with aluminum voice coil and the original cloth annulus. It has a phenolic spider. The one converted.

The EB4 is the one used for the new paper cone (photos 7&8) and the wood cone replacements.

The EB4 has a hand wound copper coil with a dcr of 11.11 ohms and a Z of 13 ohms.

 

 

FIGURE 1

The lopsided black curve (14) is that of the EB4 unit with the wood cone and hand cut cardboard spider.  That spider will have to be replaced with a more stable one as this one seems too weak to hold the cone and coil assembly in  a vertically centered position.  In short, the cone and coil assembly drops about 1mm to 2mm.  While this has little effect on low level frequency response as the cone barely moves enough to displace the windings in the magnetic field, it quite possibly can when the cone is driven to an excursion of >1mm.  A spectrum analysis should show that.  However, a speaker playing music reproduces a multiplicity of frequencies simultaneously.  the lower frequencies can easily drive the cone to >1mm.  The uneven position of the coil in the field can then cause non-linear distortion at any frequencies being reproduced at that moment.

The cone in the Super 8 has an effective piston area of 0.01887 square meters.  Taking a frequency of 500hz which is at the lower practical limit for this speaker (Wharfedale's  common crossover point was usually 800hz), at 0.1mm excursion, this will produce 96dB at 1m.  At half that excursion, the SPL is 90dB.  But, at 0.25mm and 0.5mm, the SPL becomes 104dB and 110dB, resp.  Taking this further, at 1mm excursion, which is just about Xmax, the SPL goes to 116dB.

The conclusion is that anyone listening to music at 104dB or higher isn't concerned about non-linear distortion in a loudspeaker.

What's odd here is that the EB3 & EB4 are labeled as 10 ohms and 15 ohms, resp., yet both measure as slightly less that 8 ohms.  Probably the effect of 60 years.

 

 

FIGURE 2

The green and blue traces are those of units EB3 and EB4, resp. These are both all original units, meaning that the cones are around 60 years old and brittle.  The red trace is the EB4 with a new cone.  The probable reason for this radical difference will be shown in the spectrum analyses of these units.

The prpple trace will be explained later.

 

 

 

 

Spectrum Analyses

This can get lengthy so it'll be taken in sections.  First, the all original EB3 and EB4 units will be compared.  Then the EB4 original will be compared to the EB4 with the new paper cone.  Following that, the EB4 with the new paper cone will be compared with the EB4 with the wood cone. Then, the EB4 with the wood cone will be compared to a recently purchased new speaker with a wood cone. (The prpple trace of fig.2)

In these images, the fundamental is the tall trace on the left.  The harmonics are the shorter traces above the noise floor.  The thin traces below the noise floor are part of that noise.  The noise floor is 80dB below below the top of the scale which is auto adjusted. These traces are measured with about 1W applied to the speaker.

The sum of all these spikes other than the fundamental is referred to as THD+N (Total harmonic Distortion plus Noise)  The top left image would be about 1% or less THD+N while the last two on the left can be as high as 10% THD+N.

     

TABLE 1

EB3  All original Super 8 FS/AL  

EB4  All original Super 8 CS/AL

 

Figure 3  400hz

 

Figure 4   400hz

 

Figure 5   700hz

 

Figure 6   700hz

 

Figure 7   1200hz

 

Figure 8   1200hz

 

Figure 9   2000hz

 

Figure 10   2000hz

 

Figure 11   3000hz

 

Figure 12   3000hz

 

Figure 13   4000hz

 

Figure 14   4000hz

 

Figure 15   5000hz

 

Figure 16   5000hz

 

 

The spectra of both the EB3 & EB4 are quite similar up to 4khz, where the EB3 goes to hell in a handcart.  This may be due to the possibility of the EB4 being a much later manufactured speaker.  This assumption is based on its having a dome paper dustcap with a hole in the center.  Also, the label on the back of the magnet assembly is white, whereas that on the EB3 is the older gold label.  Also, as unscientific as the following may seem, the cone of the EB3 feels stiffer to the touch than that in the EB4, possibly due to becoming brittle with age.  This would result in serious cone breakup at higher frequencies.

Looking art the green(EB3) and blue(EB4) traces in fig.2, a bump of about 4dB can be seen in the green trace between about 3800hz and 4400hz.  That may be the result of the plethora of harmonics above and below 4khz as seen in fig.13.  The LMS does have a 4th order bandpass filter which will remove some but not all these nearby harmonics.  However, that logic doesn't explain the difference between between figs 15 & 16, despite the response traces in fig.2 being identical between 4500hz and 5500hz.

 

 

 

The comparison between the original EB4 and the EB4 with the new paper cone.

It should be noted that the figures on the left of table 2 are then same as those on the right of table 1

TABLE 2

EB4  All original Super 8 CS/AL EB4  with flexair© type cone

© Jensen Speakers

Figure 17  400hz  1.76%THD+N  

Figure 18  400hz

 

Figure 19  700hz  1.07% THD+N  

Figure 20  700hz

 

 

Figure 21 1200hz   2.24% THD+N

 

Figure 22  1200hz
 

Figure 23  2000hz  2.49% THD+N

 

Figure 24  2000hz
 

Figure 25  3000hz  2.49% THD+N

 

Figure 26  3000hz
 

Figure 27  4000hz  2.67% THD+N

 

Figure 28  4000hz
 

Figure 29  5000hz  3.62% THD+N

 

Figure 30  5000hz

 

Regarding the above, the spectra of these two speakers are remarkably similar.

 

 

The comparison between the EB4 with the new paper cone and the EB4 with the wood cone

It should be noted that the figures on the left of table 3 are then same as those on the right of table 2

TABLE 3

EB4  with flexair© type cone

© Jensen Speakers

EB4  with wood cone
Figure 31  400hz

 

Figure 32  400hz

 

Figure 33  700hz

 

Figure 34  700hz

 

Figure 35  1200hz

 

Figure 36  1200hz

 

Figure 37  2000hz

 

Figure 38  2000hz

 

Figure 39  3000hz

 

Figure 40  3000hz

 

Figure 41  4000hz

 

Figure 42  4000hz

 

Figure 43  5000hz

 

Figure 44  5000hz

 

 

Again, six of one and half a dozen of another.  Very little difference between the new paper cone and a wood one.  BTW, the wood is red oak veneer.  It see3ms worthy to mention that a very high end 7 inch midrange unit costing over $900 has 0.6% second and the same at third harmoni9c between 2khz and 5khz.  This sums to around 1% and that's only two harmonics.  A spectrum analysis or a complete THD+N analysis will undoubtedly yield higher distortion figures.  The point that is being attempted to clarify here is that no speaker is free of distortion, with possible exception of plasma speakers, which have no diaphragm and are totally impractical outside the laboratory.

Someday.  Keep in mind that in the sixties, the laser existed only jn a lab and was huge requiring lots of power.  Today, one can generate a laser with a AAA battery in a key fob.

 

 

The comparison between the EB4 with the wood cone and the newly purchased wood cone 6" drivers.

It should be noted that the figures on the left of table 4 are then same as those on the right of table 3

TABLE 4

EB4  with wood cone New 6" wood cone speakers
Figure 45  400hz

 

Figure 46  400hz

 

Figure 47  700hz

 

Figure 48  700hz

 

Figure 49  1200hz

 

Figure 50  1200hz

 

Figure 51  2000hz

 

Figure 52  2000hz

 

Figure 53  3000hz

 

Figure 54  3000hz

 

Figure 55  4000hz

 

Figure 56  4000hz

 

Figure 57  5000hz

 

Figure 58  5000hz

 

 

Again, very similar with small exception of the pairs beginning at fig.51-52; the difference is very subtle and probably inaudible.

 

 

 

Impulse response comparison between the EB4 with the wood cone and the newly purchased wood cone 6" drivers.

TABLE 4

 

These impulses were generated with the CLIO as it will allow a half cycle to be generated at any frequency.  The signal was captured by the PicoScope as this system makes it easier to capture the image.  The 6" wood cone claims a response to about 75hz which renders the 50hz samples kind of meaningless.  At 100hz and 500hz, the EB4 seems to have a better damping.  However, the 6" unit does better above 2khz.

This driver is better used as a midrange above 400hz if used in a 3-way system.  It may also do quite well with or without a tweeter for low level easy listening as it does well up to about 7khz.  This unit will find it's way into a 3-way system as a midrange covering the band from 400hz to 5khz

The generator is a constant voltage output.  The power figures associated with each figure are calculated at the impedance of the speaker at the generated frequency.

 

EB4 with the Wood Cone Wood Cone 6"
Fig.58  50hz  12w

 

Fig.59  50hz  3.6w

 

Fig.60  100hz  6.4w

 

Fig.61  100hz  17w

 

Fig.62  500hz  16w

 

Fig.63  500hz  23w

 

Fig.64  2000hz  13w

 

Fig.65  2000hz  18w

 

Fig.66  3000hz  11w

 

Fig.67  3000hz  17w

 

Fig.68  4000hz10w

 

Fig.69  4000hz  16.5w

 

 

 

Test Equipment.

Linear-X Loudspeaker Management System (LMS)

Pico Scope 2205A

Tektronix AA501 Distortion Analyzer

CLIO Pocket CP-01

Behringer ECM-8000 microphone

ART Tube MP (Mic Preamp)

ADCOM GFP565 preamp

ADCOM GFA535L power amp

Two desktop computers and one laptop.

 

 

Back to the Wharfedale page

Back to the loudspeaker main page