To: Audio safety aficionados Reference: EN60065:1993 and similar IEC, UL, and CSA standards
Just a few thinks to chew on. I encourage a response from everyone, especially those involved or associated with European or North American safety agencies. 1. Clause 4.1.5 includes a description of a band pass circuit for the input signal for audio amplifier temperature testing. Assuming the RC time constant (tau) is defined (I think) as 1/(3db frequency corner), clause 4.1.5 describes a band pass with a low frequency corner at 5 microseconds, or 200 kHz, and a high frequency corner at 250 microseconds, or 4 kHz. Not much of a band pass, and clearly (If one refers to the calculation shown in fig. 2a) there is a typo and the clause 4.1.5 text should read 5 milliseconds rather than 5 microseconds. Keeping this in mind, clause 4.1.5 text describes a 12dB per octave (double pass) band pass with 3dB points at 200 Hz and 4 kHz. This is not what is shown in fig. 2b of the standard. First of all, the response curve shown in fig. 2b is centered around approx. 180 Hz. Second, the 3dB points are at approx. 40Hz (low) and 550 Hz (high). In our lab, for the sake of simplicity, we perform audio testing by injecting white noise from a CD player into a graphic equalizer, adjusting the EQ bands at flat response for frequencies 200Hz - 4kHz, and -12dB for other frequencies. Comments? 2. Clause 4.2.6 describes loading the amplifier to 1/8 maximum power, using the signal described in 4.1.5. Assuming max power is determined by finding the point where there is an increase in THD at clipping. How does one accurately determine this point with a white-noise signal input? Are there THD meters that can do this, and if so, how much do they cost? (sorry for my ignorance here). Also, since the THD of white noise would fluctuate (since white noise is random), at what point would one determine that clipping had been reached? In our lab, the clipping level is determined by amplifying a 1kHz sinusoidal signal and looking for clipping on an oscilloscope. The input voltage to the amplifier is set at RATED (not 1.06 times rated). If it is possible to measure THD of white noise, I am guessing that clipping (say, defined as THD>10%) would come at a lower output power level than clipping of a sinusoid, due to the random nature of the white noise - causing peaks that would clip as the output signal approaches the DC bus rails. Comments? 3. Clause 4.3.1 states that as a fault condition, one shorts across creepage and clearance that is less than basic insulation (curve A). Should that not say to short across creepage and clearance that is less than double/reinforced insulation (curve B)? Otherwise, one could have basic insulation creepage (say 3.0 mm) plus 1 micron, and since only the basic insulation is shorted, there remains 1 micron of insulation. While the 1 micron distance would not provide supplementary insulation, I think that it is appropriate that any insulation that is less than double/reinforced can be considered shorted as a fault condition. Looking forward to your responses, Jeffrey Gray Product safety lab manager, Compliance West
