DIY Front End 2022

Thanks Cinco! Using the 73 parts tested, it looks like 20 pairs with all 4 transistors matched within .1mA could be created. So the yield is 40/73 = 55% for this particular batch of parts. I have no idea if .1mA is the right tolerance for matching, but it is twice as good as the +/- .1mA the transistors from the diyAudio store are matched to, so I am thinking it is a conservative selection.

Regarding the thermal loading concern would it be better to be using multiple single transistor packages (2SK109) rather than the dual transistor packages (2SK2145)?
 
I have no idea if .1mA is the right tolerance for matching

That is a personal choice, as well as dependent on application.

If you run the JFET at Idss of say 8mA (as those from the store), 0.1mA is 1.25%.
If you run them at 1.5mA bias as in this circuit, then 0.1mA is 7%.
I posted a spice file before, so you can simulate what impact 0.1mA mismatch would have, if you want to find out before building.


Patrick
 
I see that the data sheets do indicate that the 2SK2145 is rated for 300mW dissipation and the 2SK209 is rated for 150mW dissipation. Maybe I did not put enough detail into my question. Does the increased surface area of the 2x single transistor packages help with the thermal loading concern or is there another factor? Thanks!
 
I see that the data sheets do indicate that the 2SK2145 is rated for 300mW dissipation and the 2SK209 is rated for 150mW dissipation. Maybe I did not put enough detail into my question. Does the increased surface area of the 2x single transistor packages help with the thermal loading concern or is there another factor? Thanks!
tl;dr Dual pkg is still rated 150mW per transistor. A single SOT-23 running at half power would be generating less heat toward max. rating (junction), of course, which may be the gist of your question. Toshiba only specifies SOA chart, though, so that's the important thing.

Also, it's generally better for both transistors' temp to track each other and thus maintain balanced characteristics, which they would better do by sharing same pkg.

Anyone can quickly get in the weeds by trying to understand how heat dissipates on a PCB, which you don't need to do in using these parts.
 
  • Like
Reactions: 1 users
I have received my DIY Front End 2022 Kit, thank you Nelson.

I am in the process of collecting parts, and I just happen to have 4 Cornell Dubilier 5µF/100V metalized PP capacitors in my parts collection. The size is not a problem, since I will be mounting the DIY Front End card on an existing socked circuit card, and I can mount these capacitors on that card. My question is, since I will be running the DIY Front End at unity gain, and the negative input will be grounded, how critical is C2, or do I even need it?

Thanks
ceulrich
 
Member
Joined 2012
Paid Member
So I ordered some parts from Mouser yesterday. Nothing exotic, just film and electrolytic capacitors, zener diodes, rectifiers, RCA connectors, terminal blocks, and power transformer for my DIY 2022 Front End build, plus 30 J113 to sort for future use. I will be building a linear power supply for the Front End.

Now it is stuck in Export Review, and I can't get an answer from Mouser why it is being reviewed. There's nothing exotic - the highest tech component in my order are the J113. What do they think I am planning to build? o_O
 
I don't believe I would be able to hear a difference between military grade and regular resistors in my projects. :)
I recently did a small test, a common source on which I tried 3 types of drain resistors, no name metal film, rn55 and caddock mp915.
The caddocks were the lowest thd followed by the rn55 with 2x the thd and on the last place there were the no name with 4x thd of the caddocks.
 
I recently did a small test, a common source on which I tried 3 types of drain resistors, no name metal film, rn55 and caddock mp915.
The caddocks were the lowest thd followed by the rn55 with 2x the thd and on the last place there were the no name with 4x thd of the caddocks.

Interesting! Thats pretty Substantial differences, if you ask me. And thats of course if the tests and measurements is done right.

Did you set up any serious listening tests and noticed any substantial sonic differences in real life?
 
  • Like
Reactions: 1 user