By Julian D. Hirsch
Fuse Distortion?
THIS month's column was inspired by a newsletter (published by a manufacturer of high-quality amplifiers) in which concern was expressed about distortions introduced by inserting a protective fuse in the line from amplifier to loudspeaker. Such fuses are often built into speakers, or in some cases an amplifier, and many speaker manufacturers recommend that their products be protected by user-installed fuses.
Distortion from a fuse? How could this possibly come about? It is normal for the electrical resistance of most metals to increase with temperature; thus, a hot fuse has a higher resistance than a cold one. In itself this cannot produce distortion-at worst, it would slightly reduce the power available to the speakers.
However, if the fuse wire were very thin, with a low thermal mass, the fuse element could heat and cool very rapidly, possibly even during a single cycle of a low-frequency audio waveform. Consequently, at a very low audio frequency the fuse resistance might thermally "track" the signal waveform. In such a case, there would be a resistance in series with the load that would vary with the change in current flow through the speaker during each cycle of the signal. The result would be a form of nonlinear distortion, since the current waveform would not vary linearly with the signal.
The amplifiers made by the manufacturer whose newsletter raised this subject are noted for their very low harmonic-distortion ratings.
Being concerned about this source of distortion, he avoided it by placing speaker fuses within his amplifiers' feedback loops rather than in the speaker lines. Finding my interest aroused by a "problem" whose theoretical existence I recognized but whose practical importance seemed questionable, I decided to find out for myself just how serious the non-linearity of a speaker fuse could be.
I had at hand an amplifier with extremely low distortion and a new Hewlett-Packard Model 239A low-distortion oscillator, both with intrinsic distortions of less than 0.001 percent. I connected the output of one channel of the amplifier to a precision 8-ohm non-inductive resistor load with provision for inserting a fuse in series with the speaker line. The amplifier was driven at 20 Hz from the low-distortion oscillator. The output signal (across the load) was first passed through our Radford distortion analyzer to reduce the level of the fundamental frequency so the harmonics could be measured on our Hewlett-Packard 3580A spectrum analyzer. This extends the system's measurement range from 90 to as much as 130 dB (corresponding to 0.00003 percent distortion!), although the 239A oscillator's distortion typically falls within 0.001 to 0.0007 percent, depending on the frequency.
WITH no fuse in the speaker line, the output current was set at 2 amperes (32 watts). The distortion was 0.003 percent, equally di vided between second and third harmonics. After a 2-ampere (3AG) fuse was inserted in the speaker line, the second harmonic was un changed but the third harmonic increased to 0.0063 percent, for a total harmonic-distortion reading of 0.0067 percent. The test was repeated with a 1.5-ampere fuse, at its rated current, with essentially identical results.
Larger fuses have heavier internal elements and could therefore be expected to have less of a distorting effect (they are also less likely to protect a speaker from anything short of catastrophic amplifier failure). We tested a 5-ampere fuse with the amplifier delivering 100 watts to a 4-ohm load (necessary to develop the fuse's rated current). The distortion was a mere 0.0022 percent with or without the fuse.
The conclusion (to me, at least) is inescapable. A typical speaker-protecting fuse will introduce some measurable distortion, but only at the lowest frequencies and at current levels close to the fuse rating. This distortion may gel as "high" as 0.01 percent, although our reading never approached that figure. Not being one of those people who believe that everything that is measurable can be heard. I would judge the importance of this effect on the sound of an audio system as about the same as that produced by a conjunction of Venus and Mars.
But there is another potential problem with speaker fuses. A fuse's internal resistance can be appreciable, and the effective damping factor of an amplifier can be radically reduced by such added series resistance. A couple of years ago I wrote here about what happens to the effective damping factor when a small resistance is added to the speaker-line circuit.
The answer is, essentially nothing. since the internal d.c. resistance of a speaker (typically several ohms) makes worrying about an extra tenth of an ohm or so seem rather silly. I measured the dynamic resistance (at rated current) of the 2-ampere fuses I used and found it to be 0.1 ohm.
How does this 0.1 ohm of fuse series resistance relate to the resistances in other sections of the speaker circuit? Well, if you use #18 lamp cord for your speaker wiring (admittedly a smaller gauge than desirable), a 16-foot length will introduce a total resistance of about 0.4 ohm. The special (and expensive) "low-inductance" speaker cables that are popular in some audio circles typically have a resistance of 0.15 to 0.2 ohm for the same length (they also have some rather undesirable side effects, but space does not permit going into that here). If the amplifier is a good-quality unit with a damping-factor rating of, say, 400, its internal "source" resistance will be about 0.02 ohm. With a typical speaker resistance of 2 ohms, speaker cables having a 0.2-ohm resistance, and the amplifier's 0.02 ohm, the total is 2.22 ohms. The true damping factor is then 3.6, a far cry from the amplifier's inherent factor of 400! Adding a fuse with a 0.1-ohm resistance will reduce this to 3.45, which is certainly no more serious than the 0.01 percent, or less, distortion that the fuse will also add! My advice, then, is to forget imaginary problems and fuse your speakers if the amplifier you are using is capable of delivering significantly more power than the speakers are rated to handle-and if the speakers are not already protected by their own fuses. I doubt that anyone will be able to hear the 0.01 percent added distortion at 20 Hz, but I can attest that the silence when a speaker voice coil burns out is deafening!
AUDIO BASICS: From Infrasonic to Ultrasonic, RALPH HODGES
Source: Stereo Review (USA magazine)