Technical White Papers
ESL Amp Bias White Paper
The purpose of bias is to reduce amplifier distortion. Bias has no effect on frequency response, and it is frequency response that defines "warmth", "richness", and "fullness" in sound. In short, changing the bias will not add any "warmth" to the sound; it will only add warmth to the amplifier chassis.
Different amplifying devices have different transconductance curves, whether they be tubes or transistors. Transconductance is where a certain amount of input voltage results in a certain amount of output current. There is a ratio between the input voltage and output current. This should be a constant at all power levels. Of course, no amplifying device is perfect, so the ratio is NOT constant. Distortion will develop to the degree that a device deviates from a constant transconductance ratio.
Typically, very small input voltages will not produce the same transconductance ratio as input voltages that put the device in the middle of its transconductance curve. In fact, the ratio will be zero at very small voltages, because the device won't conduct at all until a certain threshold is reached -- this can be seen as "crossover notch distortion" on an oscilloscope. Likewise, when a device nears its maximum power, the ratio changes as it can no longer deliver more current as the input voltage increases ("clipping").
The purpose of the bias current is to put the amplifying device into its linear operating range, and this will vary greatly depending on the type of device and its particular operating parameters. Typically, tubes are very non-linear and require quite a lot of bias to reach reasonably low distortion levels. Power MOSFETS are only slightly more linear than tubes, and the best bi-polar transistors are far better than either tubes or MOSFETS.
Before the discovery of negative feedback, some engineers even biased the tubes to the center of their transconductance curve (Class A operation). This was the only way to get distortion down to reasonable (around 3%) levels, but had the penalty of extreme heat generation and rapid tube deterioration.
Transistors vary widely, with power MOSFETs being the less linear than bi-polar types. So generally, MOSFET amps will need more bias to reach low distortion than bipolar types. And there are very different transconductance behavior with different transistor designs in the same class, so bias may vary widely.
Keep in mind that all modern, well-engineered amplifiers use negative feedback (NFB) to reduce distortion to levels far lower than can be achieved with bias alone. An amplifier's distortion is a combination of both bias and negative feedback. Keep in mind that for most humans, the threshold of distortion detection is around 3%. If very special test tones are used, some people can hear about 1% distortion. No test has ever shown that a human can hear distortion levels below 1% under any circumstances.
Sanders Sound Systems amplifiers use a very modern, sophisticated, and expensive type of bi-polar transistor made by Motorola that combines very high power capability with an amazingly linear transconductance transfer function. As a result, we are able to reach virtually un-measurable distortion levels with only a trace of bias.
Without bias or NFB, a typical amplifier will have high distortion levels (perhaps 10% or so, depending on the amplifying device). With enough bias applied to move the device into the linear area of its transconductance curve, the distortion will drop to around 3%, or even lower with very linear devices. The bipolar transistors Sanders Sound Systems uses can reach distortion levels of 0.08% without the use of NFB. Then very little NFB is required to reduce the distortion to non-detectible levels (less than 0.01%).
Some audiophiles remember the days when huge amounts of global NFB was used in amplifiers and this caused problems with TIM (transient intermodulation distortion). A few audiophiles still believe that NFB is undesirable because of TIM. But this is no longer true. Engineers now use NFB around the local circuits, use only a small amount of global feedback, and compensate it properly. The result is that NFB is now free of problems and has no adverse affects on an amplifier's performance. This is a good thing because it is impossible to make a low-distortion amplifier without the use of NFB.
Now let's look at the Sanders Sound Systems ESL amplifier without any NFB and see what effect bias alone has. With no bias at all, crossover notch distortion is present and the distortion is unacceptably high. With just enough bias to turn on the transistors, the distortion suddenly drops to around 0.2%. If we keep turning up the bias, we can reach a minimum distortion level of .06%. This is remarkably low distortion for an amplifier without any NFB. But this much bias requires a continuous power dissipation of more than 70 watts of power.
But why should we waste all that power? If we turn the bias down to just 3 watts, the distortion only climbs to about 0.2% -- still well under the 1% human distortion detection threshold. But we prefer lower distortion levels. So by adding just a little NFB, we can cause the distortion level to become un-measurable (less than 0.01%), and there is only 3 watts being wasted as heat.
In summary, there is no magic. Sanders Sound Systems amplifiers sound completely clean because they have very low distortion and lots of power. How this is achieved has nothing to do with its frequency response. We use a combination of very good transistors, low bias, and low NFB to achieve this performance, while keeping the amplifier running cool. Adding more bias will not change the sound of the amplifier in any way.