Thiel SCS4 Measurements and Analysis
Impedance Magnitude Measurement of the SCS4 (40 db Graph)
System impedance drops to about 3 ohms in perhaps the worst place it can, the region between 200 and 300 Hz. The reason this is worrisome is because this is the range of frequencies where the greatest long term power demanded in the musical spectrum is often found. Direct measurement of the system DCR shows a VC impedance of less than 3 ohms. (I measured 2.8 ohms on my sample).
The blip in the curve at about 1.6 kHz may be the result of the resonant frequency of the tweeter interacting with the crossover. It might also be a poorly controlled resonance in the woofer cone. This is the same place where the midrange THD grows to be highest during a single frequency sine sweep. (It makes sense because this is often the low frequency limit for the tweeter, where the voice coil will have to move the farthest, putting the greatest strain on the motor of the tweeter. According to Thiel, the actual crossover frequency of the system is 2500 Hz. If that is the case, then this seems to be a resonance not ideally controlled.
The tweeter impedance remains very flat at the highest frequencies. This is either the result of network compensation (the crossover) or a shorting ring (copper usually) on the tweeter motor. The extension of the tweeter on axis is good out to 20 kHz as advertised.
(with 1/24th octave resolution - 40 db Graph)
As you may be able to tell from the curve above, this is a live room measurement. For those of you who do not make a living with measurement microphones, you are likely wondering either why I might try to do this without an anechoic (echo free) chamber. Besides the million dollar price tag, to build one which is good down to very low frequencies requires a huge amount of space. That said, those of us who do (make their living with a measurement mike) can often look at these curves and see which aberrations are due to the room, and which are due to the speaker. (I call it reading through the room). As a rule of thumb, I take measurements in live rooms near-field on the ground plane (the floor) which works best for subwoofers or to see the low frequency performance of a speaker alone. I have found there is often too much error in this method to use it for full range measurements. The other technique is to place the speaker midway between the floor and ceiling and the mike at a distance around three times the size of the loudspeaker unit. In this case the curve above is taken at 3 times the driver diameter or in the case of a small 6.5 inch coax, 16.5 inches away. (A compromise between sources of measurement error). This is how the SCS4 was measured. If you are asking yourself "why is Paul telling me this?" it is because in a different laboratory, we may see a slightly different curve, and I am mindful not to ascribe ALL of the numerous peaks and dips in the curve above to the speaker alone. Taking a curve of the mate (other in the pair) of this speaker at a larger distance in another laboratory, I did not measure the large hole at 5 kHz. That said, 5 kHz is not the frequency range where rooms typically misbehave.
response at 0, 15, 45 and 90 degrees (40 db graph)
Blue – on axis, Red – 15 deg, Pink – 45 deg, Green – 90 degrees
As you can see, the off axis response is smooth and rolls off in a predictable and well behaved way. At any angle within 45 degrees, there is a small loss only and it is confined to frequencies above 2k Hz, the range of the tweeter. The severe loss at 90 degrees is to a great extent a result of the shadowing of the tweeter by the woofer cone.
For the distortion measurement, I used a signal generator (Spectra RTA) to run a slow 20 second sweep from 20 to 22 kHz. The distortion meter was also running, and was displayed in % THD. What I was looking for was a range over which the distortion would spike. It was already clear from listening, what I would find, and the numbers bear it out.
|Power Used||Below 50 Hz||50 - 100 Hz||100 - 1kHz||1.6 kHz|
THD DISTORTION MEASUREMENTS OF THIEL SCS4
As the excursion of the small woofer peaked, so did the distortion. Once the excursion was minimized, THD fell considerably. At the blip in the Impedance curve, right around 1.6 kHz where the response dipped (see curve), there was a noticeable rise in the distortion; tolerable, but still noticeable. The voltage for the first sweep was set at 2.83 volts (This is by the way, 1 watt into an 8 ohm speaker but is actually 2 watts into a 4 ohm speaker load, because Power = V squared/Ohms). When manufacturers use 2.83 volts to specify sensitivity to a 4 ohm speaker, instead of 2.0 volts, I tend to believe they are trying to get one over on the consumer, by inflating the sensitivity figure by 3 db, or 100%. Thiel specifies 87 db for 2.83 Volts RMS for the SCS4, which means its 1 watt sensitivity is actually 84 db! After the 2 watt sweep was run, the voltage was then increased to 9 volts, or 20 watts into 4 ohms. The signal power was then raised to 11 volts (30 watts), and the test is repeated for a final time. It was thought by me a sweep of any more than 30 watts RMS would result in a tweeter failure.
This must be an audiophile review! lol