THESIS BITCHES!
Holy fucking shit I actually have a shot at finishing this thing!! I spent a lot of time at it both today and yesterday trying to design a Vacuum tube replacement circuit. Yes, I realize that technically that means I spent 2 days on what basically amounts to the climax of my thesis. I think there is a sex joke there... ANYWAY! I had a major breakthrough today and I am totally pumped about it. I've been trying to design a Class AB FET amplifier and not been having a lot of luck. Mainly I have had 2 problems: finding matched FETs and biasing attenuation.
The first problem is reasonably easy to solve in concept. Technically all I need to do is get on DigiKey and surf until I find 2 FETs that have the specs I want, one p-channel and one n-channel. Well... it doesn't really work like that. First I need to figure out what I want. I chose to go with bigger being better and got 80W FETs. OK... well then I need to make sure they have PSpice models. OK... PSpice Pro has a huge library but so does Digi.... Anyway... long story short (too late) I finally found 2 that work just great.
My second problem is that I was losing about half of my power to the biasing resistors on the back of the FETs. (The FETs need a voltage divider on the back to keep them "on" so the signal doesn't have a "hiccup" near 0v. That being said, I was losing about 50% of my signal to those resistors. This is totally unacceptable. This is where it gets interesting. I tried adding an amp or two and after failing at that I had my major breakthrough. What do caps look like in DC? Well they look like open circuits. DC hates caps and thats why you use them to filter out DC in this type of circuit. What do large caps look like in series in AC? Shorts! AC doesn't give a shit about caps in series provided the capacitance is sufficiently large. The FET biasing is DC so why not just bypass the biasing resistors using caps. That way the signal gets through unobstructed and the FETs can have HUGE biasing resistors to reduce power waste. I nearly shat myself when I tried this. It work! Not only that but it works WELL! The signal is basically at unity gain!
So I no longer have attenuation of the input signal... how does the output look? Does it look like a tube amp? No. It looks BETTER. The curves are actually better than some of the tube amps I played with. *glee* If I could run across a beach into my own arms right now I would.
Also... I spent about 5 hours dicking around trying to figure out why I had a DC shift to my output. Turns out I put my FET in backwards. Fuck old circuit symbols!
I might post pictures later. Also I am thinking about naming the Design. Ivanova perhaps?
The first problem is reasonably easy to solve in concept. Technically all I need to do is get on DigiKey and surf until I find 2 FETs that have the specs I want, one p-channel and one n-channel. Well... it doesn't really work like that. First I need to figure out what I want. I chose to go with bigger being better and got 80W FETs. OK... well then I need to make sure they have PSpice models. OK... PSpice Pro has a huge library but so does Digi.... Anyway... long story short (too late) I finally found 2 that work just great.
My second problem is that I was losing about half of my power to the biasing resistors on the back of the FETs. (The FETs need a voltage divider on the back to keep them "on" so the signal doesn't have a "hiccup" near 0v. That being said, I was losing about 50% of my signal to those resistors. This is totally unacceptable. This is where it gets interesting. I tried adding an amp or two and after failing at that I had my major breakthrough. What do caps look like in DC? Well they look like open circuits. DC hates caps and thats why you use them to filter out DC in this type of circuit. What do large caps look like in series in AC? Shorts! AC doesn't give a shit about caps in series provided the capacitance is sufficiently large. The FET biasing is DC so why not just bypass the biasing resistors using caps. That way the signal gets through unobstructed and the FETs can have HUGE biasing resistors to reduce power waste. I nearly shat myself when I tried this. It work! Not only that but it works WELL! The signal is basically at unity gain!
So I no longer have attenuation of the input signal... how does the output look? Does it look like a tube amp? No. It looks BETTER. The curves are actually better than some of the tube amps I played with. *glee* If I could run across a beach into my own arms right now I would.
Also... I spent about 5 hours dicking around trying to figure out why I had a DC shift to my output. Turns out I put my FET in backwards. Fuck old circuit symbols!
I might post pictures later. Also I am thinking about naming the Design. Ivanova perhaps?