The Third Thooper Dooper Whizzer


This third unit was the result of trying to find the cause of some differences between the two different sized cones as well as the material used.  As it turned out, each has their faults.  Perhaps a sound absorbing material inside the cone and fixed to the beam in the center would help.  The wool inserted in an earlier design did help in smoothing the upper register but keeping it away from the cone wasn't achieved. That stuffing may have been responsible for the lack of a resonance peak shown in fig 1. Also, the coil gap would have to have been screened.  In the end, the latex paint helped.  Perhaps both would be better. A method of making a conical shaped screen to slide over the beam is being considered.

Consider any chamber housing a midrange or tweeter. The internal dimensions will be comparable to many frequency wavelengths causing serious peaks and dips in the response. Sound absorbing material is usually stuffed inside the chamber.  I use washed but otherwise untreated wool imported from down under.

The interior of these cones is circularly symmetrical at any point along its length and the cone is radiating from both sides which is tantamount to creating an acoustical nightmare.  Also, the circumference at any point has to be considered.  At the bottom, it's about 5.5" and at the top it's 12.5" corresponding to 2476hz and 1083hz, respectively. Considering these frequencies' fractional and whole number multiples, acoustical chaos can ensue.

The "twang" in Duane eddy's guitar was achieved by placing the speaker and mic into an empty water tank, which would have been circular and made of steel which has a high acoustic reflectivity.  The reflected waves would bounce back and forth and always meet in the center.  This is similar to the acoustics of the Oval Office which is an ellipse, having two focal points instead of one as in a circular room (tank).  One can whisper at one focal point and be heard at the other while many close to the one whispering would not necessarily hear it.  In an elliptical swimming pool, if one were to drop a stone at one focal point, a splash of water would appear at the other focal point.  This effect would be enhanced if the ellipse was a ellipse of revolution, like a symmetrical egg cut in half along its major axis.  Sorry for the digression.  I had to check the veracity of the above as until now, it was considered questionable.  See this link.    
5th paragraph, the line in parentheses




Photo 1 shows the first narrow cone after having been removed and photo 2 shows the second narrow cone installed.  The first is 65# paper and latex coated.  The second and third are thinner paper by 0.003' which is the same paper used for the wider (grey) cone.








The third narrow cone of thinner paper which will replace the wider cone.  It has been sprayed with clear lacquer on both sides.




Photo 4 shows the larger grey cone removed with its replacement alongside.  Photo 5 shows a close-up of the coil and spider that was salvaged from the damaged Super 12 still on the magnet.  The cone was held to the coil with ridge filler nail polish, which is thicker.  This was used to allow easy removal with pure acetone.  Nail polish remover would have taken longer.   I realize it doesn't look very nice but it's an experiment and it works.








The dowel glued to the pole piece.  Once the glue had set, the cone was applied followed by the stabilizing ring on top.





The pair in my living room awaiting audition. The woofers are Eminence BassLite S2012





Photos 8 and 9 show the left and right, resp.  The notch filter's coil and variable resistor is barely seen to the right in photo 8.  The other unit has nom notch filter, probably due to the slightly longer voice coil.  the shorter coil on the left unit is underhung as it has 6 less turns.








The tekky stuff 



Two impedance curves.  The flat one is that of the underhung.  It was caused by the voice coil slightly touching the pole piece.  This couldn't be heard because of the extremely low excursions.  A rub/buzz test was performed using a 10hz signal and driving the coil to an excursion of about 1mm which made the rubbing quite audible.  Loosening the three Philips screws and repositioning the coil and spider fixed that.

This coil shift probably happened during the wool stuffing experiment.



The corrected curves.  Note the extremely high impedances at resonance.  CLIO automatically scaled the Y ordinate so an expansion of that is shown in figure 3 which facilitates the determination of nominal impedance



Despite the irregularities in the responses, the two systems are equal at 500hz, 16W





These are the total system impedances, woofer and crossover.  The red is the right channel without the notch filter.  The black is that of the left channel, the one with the notch filter, the effect of which is seen by the hump between 3khz and 4khz.

The hump centered around 700hz is the result of the crossover point.





The black trace is that of the 65# cardstock cone.  Note the 10dB plus rise between 2.5khz and 6khz.  The parametric equalizer made short order of that and the one at 600hz.

The red trace is that of the original thin paper cone, the one with the larger top, the grey one.  It didn't sound all that bad with exception of the nasal sound caused by the 10dB peak centered around 600hz.  That was easily subdued by an equalizer.

The dip between 1khz and 3khz looks worse due to its being between the 600hz peak and the slight rises in the upper register.  It later proved to be beneficial as will be seen later.

The drop in low frequency response of the red trace is due to turning the preamp bass control down to 9 o'clock.



These curves are those of the last thin paper cone.  The peak between 2.5khz and 6khz is replaced by a smaller one. The red trace shows the effect of the notch, but now there's an obnoxious sounding peak between 200hz and 600hz. The parametric did wonders with the obnoxious sound but that wasn't the best way to go.





The orange trace is that of both units with only the notch on one channel.  The black trace is an attempt at smoothing the response using the parametric equalizer.

The 1/3rd octave graphic equalizer was used below 100hz.  This allows bypassing its use.






A Common Sense Approach

It came as a surprise that these fellas didn't sound pleasing when the response was flat within about 8dB above 500hz.  Something else was in the mix and it was discovered to be distortion.  Since CLIO will measure THD, 2nd and 3rd harmonics, lets have a look.  Figures 9 and 10



Room response

The light grey trace is without equalization and the green trace is with equalization. While the green trace sounded ok at levels below 80dB, it became irritating at higher levels.  So, various music was played and the eq adjusted for a suitable sound.  Each set of eq settings was saved and noted.  The DEQ2496 has 64 memories.  After a couple of hours of fiddling with it, the orange trace was obtained, which is a compromise for just about all my music preferences with the exception of classical.






The red trace is that of the flat response obtained with the parametric.  The lower traces are orange=THD, GREEN=2nd harmonic and PURPLE=3rd harmonic

At the peak around 1khz, THD=31%, mostly 3rd harmonic.  The band between 700hz and 2khz ranges from 6% to 31%.  While 6% isn't bad in the bass below about 150hz,  it's incredibly high in the above mentioned band.

The scaling of THD to dB is in table 1 below

For more on that topic, use this link






This red trace was acquired by using my ears with various music.  After getting one type of music satisfactory, the settings were saved to memory and recalled for another type of music.  If satisfactory, a third type of music was played and the above repeated.  If changes were necessary, that setting was then saved to memory. After several saved settings, these were looked at on the device's screen and a compromise merge was made. This was done manually.

The system does have an automatic mode using pink noise and the ECM8000 microphone recommended by Behringer.  However, this will attempt a flat response.  That method may have been easier than using the CLIO but I have a tendency to avoid automatic modes of operation in such cases. After having done this manually for decades, one gets a feel for where to make the adjustments and by how much.

The surprising thing here is that THD has dropped significantly.  That may be due to attenuating the peaks and NOT boosting the nulls.  If a null is caused by room reflections, boosting ain't gonna work. It's kind of like algebraic addition where +N and -N = 0, regardless of the value of N.

There are equalizers that only attenuate.




dB %
-10 31
-15 18
-20 10
-25 6
-30 3
-35 2
-40 1





After a couple of weeks, an idea came to mind which was to stuff the inside of the cone with wool.  This was previously done but the wool was in contact with the cone which did have an effect, albeit unreliable. So, keep the wool from making contact with the cone.

This weird contraption did keep the wool from making contact with the cone but other than that, it made no appreciable effect.  See fig 11 below

There is 21g of wool stuffed into the screen cone and it's quite densely packed.  It could barely be compressed into a sphere the size of a grapefruit.

The idea was scuttled; it didn't work any better than it looks.






Left channel only.

RED-without the wool

BLACK-with the wool

These curves are obtained at 1w8ft and without the crossover and without the notch filter





Another cone is in the offing.  Behind a door was found a brown paper wrapped 36" by 24" sheet of water colour paper that was purchased before '74 when I lived in N.Y.  It's highly textured and is 0.020" thick.  I was recently told by a professor of art that it's likely to be made of or with cotton.

So far, these two cones produce a satisfactory sound, with a little help of a parametric equalizer.  It would be nice to find a material for the cone that doesn't need electronic correction.


Back to the loudspeaker main page

Table of Contents