Hello. I have been looking at the circuit on this web page:
http://electronics-diy.com/electronic_schematic.php?id=1048http://electronics-diy.com/schematics/1048/100W-HIFI-MOSFET-Amplifier-Circuit.gifwhich is fairly simple. It has a differential input pair with feedback, a second amplifying pair with a constant current load, and a pair of N and P channel mosfets 2SK1058 and 2SJ152 for driving the output. However, these parts are obsolete. As far as I can tell, these MOSFETs have four attributes that make them useful for audio amplification: low threshold voltage (0.5 to 1.5 V), fairly low gate capacitance (< 1000 pF), 7 A maximum drain current, and 120-160 volt maximum drain-source voltage. I am finding it hard to find any common MOSFETs for linear amplification that can be used for audio. Perhaps a N and P channel pair could be recommended?
Also, I was thinking of if such a pair could not be found, using something like a IRF530 and IRF9530. These have sufficient drain current, low gate capacitance, sufficient maximum drain voltage (about 100 volts), but the threshold voltage is 2 to 4 volts. In the schematic, two forward biased diodes provide about 1.4 volt drop between the two gates of the N and P channel transistors. If the voltage drop of the diodes is too small compared to the sum of the threshold voltages of the transistors, significant crossover distortion will result. So for these transistors, I would think that an increased voltage drop is needed here. If the voltage drop is too high, the quiescent power will be significant in class AB operation. So the voltage needs to be adjusted properly. As a primitive way to adjust this voltage, putting many 1N4148 diodes in series with jumper blocks to short them out would allow this to be adjusted in approximately 0.7 volt increments. If I used a 1N5711 that would add a 0.4 volt drop increment as well. It is an unelegant but simple way to adjust the difference between the two gate voltages to approximately equal the threshold voltage sum.
So does this sound like it might work or do I have this wrong?
Thanks,
Dan