Funny story: About ten years ago I had the idea to make my own "mini-beast" that would use many J309s in a head amp (not for ES headphones, though). I never got real happy with my simulations, though, and the thought of testing and grading dozens of transistors didn't sound fun. The J309 was probably a poor choice anyway, and they have pretty much disappeared. It was sort of the go-to JFET when I was in college, probably just to suit the comfort level of my professors who had used them.
Discussion of Wrenchone's mod has been given it's own thread, here - https://www.diyaudio.com/community/threads/wrenchones-mosfet-headphone-design.412500/Hello People. Sorry for being out of pocket. I have moved 4x over the last 3 years as well as being sequestered in the work version of virtual jury duty. All of my DIY audio stuff fell by the wayside. I have dusted off some cobwebs in the brain to figure out where I was. It feels good to look over the last version of the complementary JFET buffer mini Beast. I will post some pictures and discussion where this project will go. Please be patient ( do not jump in) and watch until I am finished updates. I will let you know when I am finished with the updates.
This latest PCB is a 2-stage mini beast. The first stage is the preamp/driver. The second stage is the power output section. The preamp section is 8 parallel buffer cells. You can populate from 1 to 8 buffers as desired. The second section is 72 parallel buffer cells. On the PCB pictures, the the top row is the preamp/driver and the lower 72 cells are the power output section.
On post #301, I shared the schematics to the two different cells for the two stages.
The pictures of the PCB will be next.......
This latest PCB is a 2-stage mini beast. The first stage is the preamp/driver. The second stage is the power output section. The preamp section is 8 parallel buffer cells. You can populate from 1 to 8 buffers as desired. The second section is 72 parallel buffer cells. On the PCB pictures, the the top row is the preamp/driver and the lower 72 cells are the power output section.
On post #301, I shared the schematics to the two different cells for the two stages.
The pictures of the PCB will be next.......
Last edited:
The "Intermediate Output" labels are where you can connect a scope probe to the preamp output.
Last edited:
Close up of some of the bottom of the PCB.......
My intent for a build is to use some Punkydawgs 2SK170/2SJ74 pairs for the preamp section. I have some and this will be a good use for them. I will start with 2 pairs to match what Mr. Pass seems to have done for what has been posted here.
Since Mr Pass has mentioned Vpinch matching as a way to perform matching, I will work on a rig and do some matching before any building of the power amp section.
The PCB is 6-layers. There are exactly 6 nodes between the preamp buffer cells and the power amp buffer cells. Every schematic node gets its own unique layer. Zero routing. It makes me grin to have created a PCB that requires so much soldering and zero visible streets anywhere and instead just 6 mud puddles.
I am working on a version using 12 parallel JFETS, sufficient to drive a headphone of 30-35 ohms impedance. I have some Audio-Technica headphones in my sights. Selecting 12 JFETs for a particular Vgs and Id is going to be a chore. I'll probably use some TIS75s for the chore, as I made a huge buy of them from BG Micro before they went belly-up.
Please keep us updated on your endeavours. It's been a great read and an approach that has a lot of charm.
I'd like to do a small desktop amp capable of say 3-5W at voltages low enough so as not to require an expensive case. Spreading the heat over many devices means no heatsinks. Once you remove the cost of the case and the heatsinks the idea of several hundred J113s becomes quite acceptable.
I'd like to do a small desktop amp capable of say 3-5W at voltages low enough so as not to require an expensive case. Spreading the heat over many devices means no heatsinks. Once you remove the cost of the case and the heatsinks the idea of several hundred J113s becomes quite acceptable.
Ok. Updates completed. My next task is to create a Vp rig and figure out what JFETs to use in the power amp section. Feel free to comment away....throw stones....ignore......etc.....
You can adjust each parallel JFET individually and not bother trying to match JFETs. I have thought of using a POT in each JFET cell instead of hard-wired resistors.
@BeardyWan Do you want every JFET to be J113? That would be the single-ended version. I do have a 1-stage PCB that you could make a 3-5W amp from all J113. It would require a driver stage. The input capacitance of these massively parallel Beasts is quite high.
I have a 9-cell cascode driver PCB but it is as yet untested.
I have a 9-cell cascode driver PCB but it is as yet untested.
Woofertester -
Running the jfets at different drain currents means the would each have a different transconductance, probably with bad effects on the distortion profile. So, I prefer to make he effort and match all the jfets. Current THD for the amp as designed is about 0.2%, predominantly second harmonic.
Running the jfets at different drain currents means the would each have a different transconductance, probably with bad effects on the distortion profile. So, I prefer to make he effort and match all the jfets. Current THD for the amp as designed is about 0.2%, predominantly second harmonic.
Last edited:
@woofertester I think Single ended J113 buffer would be the ticket. I am also curious whether/when Nelson will release his Zenductor II version with choke loaded massively paralleled jfets. I need to reread and study this thread a bit more before making any decisions.
will PM you with any questions,.
will PM you with any questions,.