ABX Double Blind Test Results: Capacitors

ABX Double Blind Test Results: Capacitors

SMWTMS ABX Test Results

Capacitors

Capacitors	Result	Correct	p less than	Listeners
Polypropelene vs. Ceramic		28 / 32 = 88%	0.005	2

The conditions of this test were extreme and do not represent normal operation. The experimenters could not confirm any audible difference between polypropelene and ceramic in many ABX tests under normal conditions. The condition that produced the above result of a difference was applying a 3.6 Watt heat source next to the capacitors. Even though the capacitor test circuit was not enclosed, with the added heat the ceramics lost capacitance to such an extent that the system's low frequence response was rolled off by 3 dB at 40 Hz.

This experimental result should have been no surprise to any engineer and points only to the well known caution against using temperature sensitive ceramics in environments with varying temperature.

The specific capacitors had been provided by a person who claimed these specific capacitors would sound differnt under normal audio use. That claim was not be confirmed by the experiments.

Prior to the subjective testing, the capacitors were measured for value and dissipation factor on a bridge. The thermal sensitivity of the ceramic capacitors was noted at that point. Distortion was measured also and it was found that the ceramic capacitors could be made to show as much as 0.26% THD under realtively heavy load conditions. The polypropelene capacitors with the same loading did not exceed the measurement residual of 0.009%. The circuit used for the subjective test used the capacitors for coupling between a low impedance output buffer and a 100 kilohm input stage. In this circuit only minor distortion differences were measurable at around the 0.01% THD level.

Send questions or comments on this capacitor test to David Carlstrom.

PSACS ABX Test Results
Capacitors

Capacitors	Result	Correct	p less than	Listeners
WonderCap vs. Radio Shack		38 / 94 = 40%	-	7
400 Hz. Allpass Filter vs. Straight Wire		21 / 42 = 50%	-	4

The WonderCap was 8 mfd. To get the same value three Radio Shack non-polar electrolytics were used, 4.7 + 2.2 + 1.0 = 7.9 mfd.

The 400 Hz. allpass filter was made of carbon resistors and ceramic capacitors. This passive filter had an 11 dB loss which was matched in the the straight wire side with an attenuator.

In each test, the capacitor pairs or filters and attenuators coupled the analog output of a Marantz CD-63 CD player to the input of a Bryston 2B direct coupled power amplifier. Individual listeners made the comparisons through AKG 280 headphones connected to the amplifer's headphone jack.

Send questions or comments on the PSACS tests to Tom Nousaine.

	The different symbol means it is unlikely the percent correct score occurred by chance and thus the null hypothesis was disproven, which substantiates a real difference in sound quality.
	The same symbol means that it is likely the correct score occurred by chance, the null hypothesis is not rejected, which means no difference was found. It does remain possible a difference may be substantianted with further testing.

Send questions or comments on the ABX Web Page to David Carlstrom.