Crossover distortion should only show in the centre of the waveform around the zero crossing point. Make sure you have a suitable load attached as distortion will not show with no load. 4 ohms is good as a test.
Power should be kept low as the distortion shows much more at lower levels.
The flat tops on the sine could be down to the actual source. If it is from a function generator (so a synthesised sine using diode clipping) then it is likely normal. See if the input waveform looks the same.
This shows crossover distortion (yes, it's only a simulation but it is accurate)
And zoomed in to one of those points:
Power should be kept low as the distortion shows much more at lower levels.
The flat tops on the sine could be down to the actual source. If it is from a function generator (so a synthesised sine using diode clipping) then it is likely normal. See if the input waveform looks the same.
This shows crossover distortion (yes, it's only a simulation but it is accurate)
And zoomed in to one of those points:
Oh yes, sorry I know what crossover distortion is, at the lowest bias point that’s what I thought I was showing here
You can see the slight squiggly line there, but maybe it’s not enough. I have the amp driven into a pair of 8 ohm dummy loads. I should get another pair of 8 ohm so I can drive into two 4 ohm loads, but can’t at the moment. Should I be seeing more crossover distortion? The distortion on the peaks I knew wasn’t crossover distortion, but was curious if it could be caused by the bias voltage.
Dan
8 ohms is fine, 4 ohms shows the effect more (easier to see).
Not necessarily, the output stage will be wrapped up in the global feedback loop and that compensates and is able to remove most of the distortion, even at zero bias.
If you have a pure sine source such as a test CD (not a function generator) then you can listen at very low volume to a 1kHz sine and crossover distortion is very audible as a harshness to the pure tone. Turn the bias up even a little and it goes completely. It's a very good test, you can really hear it and its very noticeable.
An MP3 file should work for that test.
Not necessarily, the output stage will be wrapped up in the global feedback loop and that compensates and is able to remove most of the distortion, even at zero bias.
If you have a pure sine source such as a test CD (not a function generator) then you can listen at very low volume to a 1kHz sine and crossover distortion is very audible as a harshness to the pure tone. Turn the bias up even a little and it goes completely. It's a very good test, you can really hear it and its very noticeable.
An MP3 file should work for that test.
Bias setting can be a bit of an argument point.
I used to design my own amps and obviously they came with no bias setting.
So I set bias to zero, applied a small sine wave, monitored output on a scope and increased bias until crossover distortion went.
On a single mosfet pair it worked out to be 10mA.
Amp sounded great.
I used to design my own amps and obviously they came with no bias setting.
So I set bias to zero, applied a small sine wave, monitored output on a scope and increased bias until crossover distortion went.
On a single mosfet pair it worked out to be 10mA.
Amp sounded great.