The Search for the Ultimate Approach to Amplification

The purpose of this article is to educate the consumer what they need to look for when selecting a good amplifier.

Sometimes it seems like every minute aspect of audio has been addressed. We have special formulas to wash CDs. We have very expensive cables. It would seem that no stone has been left unturned to achieve the best possible reproduction possible. Ironically, the most obvious, inexpensive and effective way to improve audio reproduction is almost always ignored.

No part of the audio chain causes more destructive artifacts to the music than the passive crossover. The best designed and executed passive crossover can zap 30% to 50% of your amplifiers power and introduce gross distortions.

There are only two ways to work around this problem. One is to use full range drivers and the other is to use active crossovers with multiple amplifiers.

Active crossovers and bi-amplification have been considered exotic and expensive. nOrh has produced the first amplifier that makes it easy and inexpensive to take advantage of bi-amplification. I highly encourage you to read the excellent article on bi-amplification written by Rod Elliott.

I remember back about 20 years ago in Stereo Review, they published a review on the Soundcraftsman 5002 amplifier. The review made the statement that all amplifiers sound the same. I remember how this article was attacked over and over by the underground press. Audiophiles believed there was a big difference in the sound from one amplifier to another.

Some audiophiles believe that the difference in one amplifier is a result of the output devices. Some argue the that tubes are better than transistors. Others argue for a particular type of transformer.

When we use a tape player, CD, FM tuner or phonograph to play music, we are actually generating small voltage signals. This voltage signal is a representation of the reproduced sound. This signal is not able to drive speakers directly because speaker require current. (An electrical voltage will send a current into a load, and voltage is a requirement for a current into a load.) From the source device (CD player, tape, etc), through the preamplifier, the signal is still a voltage signal. The output stage of an amplifier must convert the voltage signal into a current signal. The problem is that the current signal is not always a representation of the voltage signal. The load created by a loudspeaker and crossover are complex. They shift (delay) the phase relationship between the voltage and current.

The picture on the left shows a situation where the current signal is 90 degrees out of phase with the voltage signal. In other words, when the voltage is at its peak, the current is zero. A loudspeaker presents a complex load where the phase will change according to the frequency generated.

If you used a scope to compare the voltage signals, they would look perfect. However, using a current probe, you would see the difference.

There are other factors besides the speaker that cause this phase shift. The capacitors in the power supply has its own inductance. An engineer might decide to build a power supply with a single 50,000 micro farad capacitor. However, it is more likely that using five 10,000 micro farad capacitors will have less inductance and give faster current response. Five 10,000 micro farad capacitors is usually more expensive than one 50,000 micro farad capacitors.

Electrolytic capacitor will always delay current delivery to the amplifier. The best electrolytic capacitors are much more expensive than low grade capacitors. Philips 056/057 caps are probably the best electrolytic capacitors. Eight of these would work better than four but four of these would work better than eight low grade caps.

A power supply that appears to be perfect can modulate when it is connected to an amplifier.

Current response is very sensitive to changes in the supply voltage. Most audio grade transformers have a difference from zero to maximum current of 5% to 15%.

Transformers that produce 3% or less must be specially ordered. Most transformer manufacturers will protest because the transformer has to be wound extra thick or with double the number or wires. This increases the complexity and costs of manufacturing the transformer.

The picture on the left shows the MultiAmps two specially built R-Core transformers. Each has only 1 ohm impedance and each transformer is powerful enough to drive both channels. The MultiAmp is a dual mono design with no common grounds or circuits per channel.

An engineer has to decide on what sort of output devices to use. Bi-polar power transistors require relatively high drive currents into the base. Bi-polar devices require a driver stage that is already a voltage to current converter.

MOS-FETS are generally better voltage to current converters than Bi-polar devices. Most commercial amplifiers use an "Emitter Follower" output stage. There are many advantages to this configuration but it is well known that an "emitter follower" is also a voltage to current phase shifter.

It is easy to test an amplifier to see what its THD, TIM and SNA ratings are. However, these measurements do not provide much useful information about the sonic qualities of the amplifier. It is much more important to improve the current response of the amplifier.

Suppose we took an amplifier and substituted batteries for the power supply. The sound would definitely improve because batteries have faster current response and no ripples. However, while the amplifier will sound better, the THD and TIM measurement will not change. (A 100 WPC stereo amplifier can be powered with six car batteries). Stressing components can also impact their sound. A transistor's saturation voltage will change as the transistor becomes warmer. An amplifier will sound different whether it is cold or it is hot. Stressing a semiconductor for long periods of time can cause a change in the sonic quality of the amplifier.

The very best designers in the world must take all these factors and other factors into consideration. Where cost is no object, fewer compromises are required. The best amplifiers from Krell, Plinius, Mark Levenson, and Pass Labs will always take most of these factors into consideration. Where cost is an important factor, compromises must be made and most of these will have some audible impact.

Voltage / Current at 90 degree out of phase