|
发表于 2009-12-14 06:10:10
|
显示全部楼层
pretty good work.
here is a small suggestion.
a fatal flaw in your design is that a) it is not thermally stable. and b) you cannot adjust idle current. you kind of addressed a) with large R1/R2. the real solution is to use a "Vbe multiplier" in place of D1/D2. But that's just go with your simpler design but replace D1/D2 with two LEDs so your amp doesn't run in Class C.
you can solve the 2nd problem by adding two resistors, Re, to Q3 and Q4's emitter.
let's say that the diodes' forward voltage drop is 1.8v. you will need to figure out how much idle current you want for Q3/Q4. I would say that 20-30ma is fairly good for modern power transistors. That means 20ma*15v=0.3w, you can run to220 transistors there without using heatsinks in the winter but can be tough in the summer. so let's say that you want to idle at 20ma.
That means the voltage drop on Re+R5+Re=1.8v*2-0.6v*2=2.4v. so Re+R5+Re=2.4v/20ma=120ohm.
you want the voltage drop over R5 to be about 0.6v*2=1.2v so that Q1/Q2 are about open at idle. that means on the two Re resistors, you need to drop 2.4v-1.2v=1.2v. At 20ma, Re+Re=1.2v/20ma=60ohm, so Re=33ohm.
you want to select R5's value so that when it is half open, it drops 1.2v. that means you should use a 67ohm pot.
obviously, you recalculate it if you want the drivers to idle at different levels but that though process is the same.
Also, regardless of how much idle current, you should drop about 25mv on R1/R2 each. engineers from HP years ago decided that is the best idle current. so for your amp, since R1=R2=1ohm, Q1/Q2 should idle at 25mv/1ohm = 25ma.
(原文件名:ourdev_512171 KL.JPG) |
|