Loudspeaker Analysis

Home Contact Us
Examined some 1970s vintage speakers, and this is a short article on various aspects of their evaluation. Speaker design is more science now than it was then. Having CAD tools helps a lot too. 
They have a brand name on the box, and they look impressive, does that mean they are any good? 

Last Updated: January 7 2007

What are they really?

View of Sansui Tower speakers

The units are badged Sansui, but I have my doubts. From the amount of damage to the grills and cabinets, they had seen a few parties, and not surprisingly, the tweeters were open circuit. The speakers stand about 1 meter tall, are constructed from particle board, not MDF, have little internal damping material, contain two 10" paper coned woofers and a horn tweeter. One woofers cone was coming away from the surround in a small section.

Dismantling the units showed the woofers to be marked Plessey/Foster C250L24 7ohm  and the tweeters to be Plessey/Foster H016N41 8 ohm and the enclosure to be a 70 liter bass reflex  with a round vent of 98mm diameter, 100 mm long. The crossover network was:- 

The center woofer is actually connected via a high pass filter, and note that no inductors were used. The upper woofer is actually being used as a midrange, yet it isn't isolated from the woofer chamber.

The 10" woofers

The damaged driver cone surround join was repaired with a contact cement and all cleaned up. The woofer's Thiele-Small parrameters were then measured to see how suitable the enclosures really were. This was done using the sine generator in TrueRTA, an 8 ohm 100watt series resistor, a power amplifier and a Digital Multimeter. The values were put in an Excel spreadsheet. Vas was determined using the added weight method, by sticking 15gms of Blu.tack on the cones, and determining the Resonance Frequency again. 

The linear plots of the Resonance Frequencies are:-

The power amplifer used in taking the measurements has a low frequency rolloff from 20Hz, and that accounts for the low end dip in the above plots.

These values and others determined with the DMM are then fed into:-

The woofers end up having :

Fs = 44Hz
Qms = 3.51
Qes = 0.75
Qts= 0.62
Vas = 96 liters

That isn't particularly impressive for a 10" driver. They don't have much magnet and are lighter than a comparable unit today. It does mean they have more top end, making the 2 way system feasable.

Theoretical Low End Response

Putting these numbers into Speaker Workshop results in a low end plot of the box to be like:-

This doesn't take into account that the upper woofer isn't actually being driven at low frequencies, and may physically be in series with the other woofer, but electrically it isn't. An undriven driver will act something like a passive radiator, in addition to the port. Note that the roll off is quite steep. Doesn't look like you will get much bass from this.

If we use the box as is and use a single 10" driver, the result looks something like:

This has less of a peak, a gentle slope and an extra half an octave, at least, of bass extension. 

Measured Low End Response

If we measure the woofer response, in the original 2 woofer system, with pink noise and a microphone very close to the dust cap we get the following.

This shows a very steep roll off from 90 Hz, and a big dip at 60 Hz. This doesn't take into account the effect of the port. The undriven middle woofer has put a big hole in the low end, but it comes up again at 30 Hz. 

If we measure the single woofer system, by removing the midrange/woofer and replacing it with a 18mm MDF cover,  

we get:

This unsmoothed plot isn't taking into account the effect of the port. The mic is only flat to 32 Hz, so ignore the plot under that. The RadioShack Analog Sound Level Meter on the C setting is used for signal level and low end and midrange plots. A RTA mic calibration file has been used. 

With a slightly different mic and speaker set up, the following shows the close mic'ed port and single woofer system responses. 

The result is a usable response to under 40 Hz, and no funky dip like the original. The cabinet needs some internal bracing, and we will add more damping material. BoxNotes shows that some peaks and dips correspond to cabinet resonances.

There is more too it than just frequency response thought. Transient response and distortion are important too. May be later.. 

The low end of the system sounds better with a single woofer in each box, and that is how we will renovate them. The originals were designed by marketing... 

Myth Busted.

The High End

Dismantling a tweeter showed that the voice coil/dome assembly was open circuit. Foster is now called Fostex, but I couldn't find a reference for this tweeter, even though other part numbered units looked similar and have replacement parts available. 

Replacing them was the simplest solution. All that requires is physically mounting them, designing a suitable cross over network, and doing all that testing and listening.... 

The RadioShack SPM has no usable response over 10 kHz, so a custom preamp and a electret mic insert is pressed into service. The mic insert used only has a response to 15 kHz, and it isn't flat, but it will do for these investigations. 

The close mic'ed pink noise response of a Jaycar CT2010 8 ohm soft dome tweeter, when driven by the existing 3.3uF bipolar electrolytic is:

This isn't a bad match for the single woofer, and results in just a slight dip at about 2.5k Hz. We need to do a bit more work yet. The tweeter has a resonance at around 1kHz that must be avoided. Higher order cross-over networks and woofer impedance compensation is required.