I'm going to listen to it for a few days at 100ma,then a few days at 30na,see if there is actually any difference. I think sometimes you can almost convince yourself things are better
30 ma is more than enough to remove virtually all distortion. Even just a couple of milliamps does that.
Simulated at 10ma and 100ma.
Simulated at 10ma and 100ma.
They shouldn't be but as with all things some will say they can hear differences. Sometimes a bit of distortion can actually 'sound better' in some cases. Its hard/impossible to quantify, I would say just trust your own ears.
I've simulated the wrong amp...
Here we go, 30ma and 100ma
I've simulated the wrong amp...
Here we go, 30ma and 100ma
It is the peak value of the signal at the frequency shown. So 'Harmonic 1' is the fundamental 1kHz sine at aprox 3.99 volts peak amplitude which is our 2.83 volts rms or 1 watt into 8 ohm.
Harmonic 2 is the 2kHz component. A pure sine would have no harmonics at all but harmonic distortion means there are multiples of the fundamental present. We look at those up the 10th harmonic. So the amplitude of the 2nd is 1.173E-4 volts in one of the examples. Pretty small.
Then we have harmonic 3 and so on.
The normalised value just makes the 1st or fundamental '1' and scales the others to that value.
If anyone wonders why the sine wave output looks like this:
it is just because the sim is calculating 2000ms run time and data collected from 1600ms which helps get a nicer FFT image. This is the 'no drivers' @ 100ma:
it is just because the sim is calculating 2000ms run time and data collected from 1600ms which helps get a nicer FFT image. This is the 'no drivers' @ 100ma:
More, its higher.
Gives 0.005% and 0.004%
And with no drivers:
Gives 0.003% and 0.004%
Higher bias generally pushes distortion down at lowish levels such as our 1 watt into 8 ohm (2.83 vrms into 8 ohm)
Zero bias gave 0.02% which is still very low.
There isn't really in pure numbers, that is why you should try it yourself and give it a fair listen. The problem is you know zero bias raises the distortion and so you are 'wanting' it it to sound worse and if you're not careful that is exactly what you hear. You listen for 30 seconds and think, Nah, this is no good and go back to 100 ma.
A really good test is to listen to a pure sine of say 1khz at a very very low level (and it must be pure such as from an analogue type signal generator and not a typical function generator) and if you turn the bias to zero and hear any change in tone or timbre of the sine then that is crossover distortion.
If you do have an amp that does that then you will also find that a bias current of even 1ma will remove that audible distortion completely.
A really good test is to listen to a pure sine of say 1khz at a very very low level (and it must be pure such as from an analogue type signal generator and not a typical function generator) and if you turn the bias to zero and hear any change in tone or timbre of the sine then that is crossover distortion.
If you do have an amp that does that then you will also find that a bias current of even 1ma will remove that audible distortion completely.
There is usually an optimum point that depends on the output stage and its configuration such as an Emitter Follower and the value of the emitter resistors or CFP (complementary feedback pair) but that applies to normal bjt transistors.
While some of that still applies to lateral FET's the laterals also have the unusual property of having a positive temperature coefficient at low values meaning bias current increases with heat and that then changes to a negative coefficient at current over around 100ma. That is where the 100ma figure for laterals comes from as that is the point the two coefficients coincide.
The negative tempco makes the Laterals immune from thermal runaway and they also do not need what would be the emitter resistors used with bjt's (and which are used to help thermal stability).
While some of that still applies to lateral FET's the laterals also have the unusual property of having a positive temperature coefficient at low values meaning bias current increases with heat and that then changes to a negative coefficient at current over around 100ma. That is where the 100ma figure for laterals comes from as that is the point the two coefficients coincide.
The negative tempco makes the Laterals immune from thermal runaway and they also do not need what would be the emitter resistors used with bjt's (and which are used to help thermal stability).
It might only need 30 or 60 mA to get the gm high enough for it not to matter anymore (which keeps crossover distortion low), but require 100 or more mA to be inherently thermally stable (so the bias stays where you put it). If you set it to a lower point that is initially acceptable for distortion it might not STAY that way over temperature. Get it over the hump and it will.