Which JFets for audio use? (other than the 2SK170,389,109 and 74)

[Bricolo]

Hi,


I'm looking for good Jfets for audio use. All the projects I saw used 2SK170, 2SK389, 2SJ109 or 2SJ74, are those the absolute best ones?

I'm wondering why other part numbers are rerely used.

Thanks
Alex

[SY]

As Tevye once observed, "Tradition!"

There are lots of great FETs out there, but you have to start by outlining a specific application.

[Bricolo]

that's for line level applications, nothing particular in mind

small gainstages, buffers, I/Vs...

the 1st project may be a buffer, maybe borbely's white follower

[SY]

Line level? There's piles out there. Anything with an en of 40 nV/rtHz or lower will be quiet enough. FWIW, on the p side, I use 2N5462 in my line stage.

[Dave]

How about for input differential pairs?

The 2SK389 are the only thing I've seen used but not only are they hard to get but I've also heard they are going out of production.

This looks like a good alternative, the U401.

http://www.ociw.edu/instrumentation/...rts/SST404.pdf

Never seen these used in audio before though.

[Bricolo]

and any of those?

http://www.onsemi.com/site/products/...outputSize=50&

[StevenOH]

Here is one reference source for japanese transistors (although not complete) for audio on the web:
http://www.geocities.co.jp/Technopol...in/tr_data.txt
Use babelfish translator to read it.

[SY]

Input diff amp for line stage or power amp: 2N5566 has been successfully used. Very nice pair matching.

[Jocko Homo]

I'm wondering why other part numbers are rerely used.

Why wonder? There is only one reason:

Nothing else is worth anything for audio.

I don't want to have to 'splain this again.

Jocko

[wrenchone]

Jocko, what you say ain't necessarily so. Yes, I have gotten good results using the 2SK170 and the old matched pairs in aluminum cans (2SJ73 and 2SK146). I also have 2SJ74s, 2SK389s and 2SJ109s, but I haven't come up with an application for them just yet that edges them to the head of the line over other projects.

However, I am also getting excellent audible results using the PN4393, an inexpensive device which is readily available in the US. The device was originally intended as a fast chopper or analog switch, but it also works very well in the linear region. True, it does not have the extremely low noise and high gain characteristics of the 2SK170, but it is quite adequate for a lot of applications. Its higher IDSS brother the PN4391 is good for cascoding, as it will give one about 4-5V Vds when positioned above a PN4393 - probably more if used with a low current device like the 2SK170. The PN4393 has a high IDSS and high cutoff voltage compared to the 2SK170, so you need to be a bit careful how you use it. If you can cater to its limitations (which is not too hard) and design to its strengths, the results can be rewarding.

I am currently running a discrete open loop preamp with a modified Pacific RIAA stage using both 2SK170 and PN4393. The PN4393s are used as current source loaded followers at the output of each stage, an application where they excel due to their high current capability compared to the 2SJ170.

The line amp portion of the preamp is a unity gain stage using cascaded complementary followers with current source loading. The devices used are the J270 and the PN4393, both of which are available at low cost from Mouser.

I am currently in the process of designing a new RIAA stage (again, a modified Pacifiic type preamp) to dump the 2SK170 altogether in favor of the 4393, with a 2X decrease in distortion for both input and output stages (at least, that's what simulation has told me, and simulation and measurement have agreed reasonably well so far). We'll see whether they still agree for this new design.

So far, the audible performance of the open loop preamp I have described above has been very pleasing, with an increase in resolution in both LP and CD material compared to the stock Nikko preamp I was using before. Anyone who is interested in the circuits I'm using can find them in the thread "Open Loop Follies" in the Analogue section. They are nothing very exotic, as I was deliberately trying to keep the circuits simple for this design to see how good a simple open loop circuit preamp could sound. The answer - very good, indeed...