What To Do With Those 2SJ28's

Jump to Latest
#41
We could approximate one by flipping a couple of the polarities in the behavioral model like so. Close enough for government work. Don't try to land a probe on Europa with it.

As presented, it's simply a mirror of the 2SK82, feel free to adjust to taste.
Click to expand...

I have been playing around with the spice-model you posted, and the 2SJ28 is one strange device....

I don´t know enough about spice to "adjust to taste". I am happy I managed to get the spice model working.

I tested the 2SJ28 and an IRFP054 in a DEF amp and there was a lot of distortion. I guess the transfer-curves does not match well between a SIT and a normal pentodish mosfet.

With the more triode-like IRFP7430 instead of the IRFP054 the distortion drops substantially. Probably due to the similarities in transfer-curves enabling distortion-cancellation.

The 2SJ28 and IRFP7430 seems like a very nice couple in a DEF-amp.
(anyone wants to sell me a pair of reasonable well matched 2SJ28?)

DEF-2sj28IRFP7430.jpg

View attachment 2SJ28-IRFP7430-DEF.asc

Infineon Technologies

(link to the spice-model of the IRFP7430)
 
#42
FFT-DEF28-7430.jpg

THD-DEF28-7430.jpg

I have no idea if these simulated FFT and THD graphs is of any interest to anyone, but I post them anyway. I guess they do indicate a reasonable level of performance from two such dissimilar devices as the 2SJ28 and the IRFP7430.

The 2SJ28 and IRFP054 had 1.0% THD at the same output-level. 0.14% THD is quite an improvement.

I have tried the IRFP240, IRFP250, IRFP7430, IRF1404, IXYS IXTH500N04T2, and they all give much more distortion in the simulated 2SJ28 DEF amp.

Cheers,
Johannes
 
Last edited:
#47
Let's talk seriously about potential 2SJ28 amp structure choices: front-end with output stage.

DEF with vfet is ready made by Mr.NP and sound fantastic !

Could we design different amplifier with touch of new "innovate" elements ?

Sony's 28 is only two pcs, not matched in lucky Diyer's drawers i guess...

My proposition:
Front-end | muscle 2SK79 with autoformer striped Edcor gain stage.
+ use of available small signal Sony's so we get new amplifier with two different triode semi's.

Output stage | Single-ended with CCS
+ low distortion , THD with frequency response are relatively flat, efficiency.

Bear with me ~ thoughts ?
 

Attachments

  • D193E0FF-0889-4D9D-8262-1B02831B7CB1.JPG
    D193E0FF-0889-4D9D-8262-1B02831B7CB1.JPG
    92.1 KB · Views: 683
#48
2X-2SJ28-DEF.jpg

2X-2SJ28-DEF-THD.jpg

2X-2SJ28-DEF-FFT.jpg

And for those lucky enough to have enough 2SJ28 for paralleling two for each DEF amp channel can expect some real THD improvements (according to LTSpice at least).

I guess the curves match much better with some more current through the IRFP7430.

This is simulated performance only. Real life might throw some surprises for those brave enough to test the 2SJ28+IRFP7430 DEF amp.
 
#58
This is derived from l'AMP with CCS.
Voltage is a given from me as I will be using a 18VAC transformer with the two secondaries in series. It is still far from being optimized but I'm now using it mainly to understand what are the parameters that govern the distortion and its harmonic content.
 

Attachments

#60
gior said:
This is derived from l'AMP with CCS.
Voltage is a given from me as I will be using a 18VAC transformer with the two secondaries in series. It is still far from being optimized but I'm now using it mainly to understand what are the parameters that govern the distortion and its harmonic content.
Click to expand...

The other night I put on the first season of The Outer Limits in which
the scientist manages to put his hand into a magnetic portal to
another dimension, and ends up with two right hands.

Of course the solution as in many other old science fiction sagas is to:

REVERSE THE POLARITY!!!!
 
You must log in or register to reply here.
Top Bottom