This thread has been split from here - https://www.diyaudio.com/community/threads/beast-with-1000-jfets-redux.364186/I played with the idea in sim, and ended up with a design using all mosfets, not a jfet in sight. It uses a "fake SIT " pair at the output, using an IRF510, , a ZVP3306A, and a couple of resistors to mimic the triode-ish behavior of a SIT. I drive the SIT with a ring-of two current source using a small signal PNP and an IRF9510. There's my favorite depletion mode mosfet at the input, the Infineon BSS159, and I close the loop from the output to the BSS159 source. Distortion sims at about 0.05%, mostly 2nd harmonic, and gain is about 10X., about the right gain for a headphone amp. I may grab a pair of Audio Technica studio headphones, as they look to sit more comfortably on the ear than the Grados I currently own.
Last edited by a moderator:
The gain phase analyzer will tell the tale. Despite your quibbles, I'm building this thing.
The BSS159 is also serving as the summing node for the global feedback,, so restricting its bandwidth will have an affect on the global feedback. Again, I'll be checking the finished product with a bona-fide gain-phase analyzer, so I will find out one way or the other.
The BSS159 is also serving as the summing node for the global feedback,, so restricting its bandwidth will have an affect on the global feedback. Again, I'll be checking the finished product with a bona-fide gain-phase analyzer, so I will find out one way or the other.
Last edited:
'Beast' is based on how many precioust toshiba jfets were wasted for few watts
NP has pallets full of those.
Having learned from sad experience, I wanted to tie the drain of the P-Channel mosfet to a negative source for some reliable turn-off drive. I will generate the negative supply using some sort of voltage multiplier scheme and a zener diode, tapped off the main XFMR winding supplying the +24V.
The output voltage at the center point is governed by the ratio of R6 and R7. The combination of the N-Channel mosfet, the P-channel small-signal mosfet and R6 and R7 act like a rough approximation of a triode/SIT.
The output voltage at the center point is governed by the ratio of R6 and R7. The combination of the N-Channel mosfet, the P-channel small-signal mosfet and R6 and R7 act like a rough approximation of a triode/SIT.
Last edited:
R10 and R11 are used to set the bias current in depletion mode mosfet M4. I split the resistor and bypassed part of it for more open loop gain. D1 is hooked to the -6V voltage source for gate protection. I will generate the -6V supply using some sort of voltage multiplier scheme off the main winding that generates the 24V, with a smallish input cap to strangle the current capability of the multiplier circuit.
Last edited: