No info on the schematic I have. What is the best way to do this. I replaced some transistors and don't see any crossover distortion on the output at full power. Do I lower bias until I see crossover notch and then increase it until it goes away?
Try 26mV across each emitter resistor, it was good enough for Barney Oliver.
http://hifisonix.com/wordpress/wp-c...oss-Over-Distortion-in-Class-B-Amplifiers.pdf
Last edited:
I have found with my quasi designs I don't need any bias as there is no crossover distortion even with zero bias.
I will read the whole article and report back. I perused over it quickly and see it's full stuff I need to fill my brain with. Thank you.
Before posting I started low at 10mV across the emitter resistors I had big clean sine waves at full power. I wonder if an FFT would tell us more between a cold and hot bias.
The bias servo should generate a voltage across it, which is equal to each base-emitter voltage drop introduced to signal: In a quasi-complementary output the Darlington pair drops approximately 2 x 600 mV and the Szklai pair 1 x 600 mV. That's 3 x 600 mV in total so there should be around 1.8VDC potential difference between bases of the driver transistors. You may use that as a rough guide. Figures are naturally approximations. We are dealing with "diodes" with dynamic impedance, not with steady voltage drop.
Crossover notch should be more apparent at LOW signal levels than at rated output power level, because the notch cuts a proportionally bigger amount of signal that has lower amplitude.
Additionally, since there is global negative feedback involved it will play its part in correcting distortion, including crossover distortion. In open loop mode the crossover notch would probably be more distinct.
Crossover notch should be more apparent at LOW signal levels than at rated output power level, because the notch cuts a proportionally bigger amount of signal that has lower amplitude.
Additionally, since there is global negative feedback involved it will play its part in correcting distortion, including crossover distortion. In open loop mode the crossover notch would probably be more distinct.
Last edited:
The crossover distortion is heard at very low levels. Adjust the pot until, with no signal and no loudspeaker connected, there is 2 - 5mV drop across R29.
Thank you all for helping out. I did input a signal and watched the output at low signal output and at cold bias there was noticeable crossover distortion. It was only noticeable are really cold to no bias conditions, e.g. 10mA and under. I did settle on 30mA and it sounds fantastic.
SWR amps are usually biased with 2 Ohm load. 2VRMS into 2 Ohms @ 1KHz, adjust until the crossover goes away, no more. Use the return jack for your input point, efx control in the full "wet" position. 120mv input and then raise the master until you get 2VRMS output.
Craig
Craig
Last edited:
- Status
- Not open for further replies.