Kolinummi concept for reducing crossover distortion implemented on Blameless

Met on the Internet


Thanks for these resources, there is a lifetime of material to go through. The interview is interesting because Hawksford seems to say that his concept is a useful design philosophy, but doesn't seem particular about one feedback design approach being superior or substantially different.

mazurek ! Ideas are in the air. You might be interested.
Thank you for the cross-reference. I used feedback to a second additional VBE multiplier in series with both output transistors, but this other reference uses feedback to a single bias spreading VBE multiplier.

I tried simulating that arrangement with the feedback to the bias spreader tapped to either the center of the drivers or the output. For bias spreader feedback tapped to the output, in a blameless style amplifier the bias spreader mainly affects the top transistor so it linearized only half the waveform. For feedback tapped to the center of the drivers, it caused the voltage to swing faster through the zero crossing region, but it did not show a linearization benefit in simulation. I wonder if this feedback provided some additional rejection to current source variation.

It does appear to show there are a number of realizable circuits with feedback to the VBE multiplier, but for the case of a single ended VAS it appears that running feedback to the existing VBE spreader only linearizes half the output, versus running feedback to an additional VBE spreader in series with both linearizes both output polarities.