DIY Sony VFET pt 1

Member
Joined 2016
Paid Member
If I remember correctly, the 16R - preferred load was for the "small" Sony VFETs (2SJ18 / 2SK60) - that's why they will get a Mu Follower in an upcoming variant of the DIY VFET amp.

I think the VFETs in part 1 (2SJ28) and part 2 (2SK82) like 8R loads - that's why the got a CCS.


Sorry, don't remember the post from the top of my head, will search for it ...


Edit:

Here: https://www.diyaudio.com/forums/pass-labs/370282-diy-sony-vfet-pt-1-a-37.html#post6602753
and here: https://www.diyaudio.com/forums/pass-labs/370282-diy-sony-vfet-pt-1-a-111.html#post6696445


Regards, Claas
 
Last edited:
The one and only
Joined 2001
Paid Member
Agreed, But it is nice to know why.

The spectral content of the thd is dependent on the impedance of the
load, since it affects the load-line of the Vfet/SIT.

Works just like a Triode in this regard.

As a general approach we optimize it against an 8 ohm resistive load.

If you have a different load, it can be biased more toward that, or
you can give it some help, like with a mu-follower, or you can adjust
the load itself...

:snail:
 
Member
Joined 2018
Paid Member
I put these little filter critters in my vfet, one per channel. Seems a bit more 3D and deep in the soundstage. I’m doing a victory lap by cranking a favorite record at proper volume for the style: Tanith - In Another Time LP. Love it. On an SP10 - thanks Jim!

Next up will be some front end boards.
 

Attachments

  • 3A291C18-93C6-4F3E-A05F-4092C3B21234.jpg
    3A291C18-93C6-4F3E-A05F-4092C3B21234.jpg
    979 KB · Views: 467
  • AB3E8E41-9B4D-4407-BD22-2FF8D70A7E9A.jpg
    AB3E8E41-9B4D-4407-BD22-2FF8D70A7E9A.jpg
    824.6 KB · Views: 453
  • 13ACDE22-3441-4FA4-A674-FBB29801E7D1.jpg
    13ACDE22-3441-4FA4-A674-FBB29801E7D1.jpg
    962 KB · Views: 461
  • 52F2C37C-FF52-41BE-8F87-FB2EE2461B3A.jpg
    52F2C37C-FF52-41BE-8F87-FB2EE2461B3A.jpg
    994.2 KB · Views: 454
Glad you enjoyed the additional filters in your VFET!

On top of the better soundstage, with time you may appreciate quite a step re tonalities, resolution (better highs and lower bass, helping in fact the better soundstage) and especialy music flow

And yes, perfect match for horns, I agree!

Enjoy music

Claude
 
Member
Joined 2016
Paid Member
Hi Randy and everybody,

we seem to be on a parallel path here ... I also added the PO89ZB filters to my DIY VFET (N, so 2SK82), one before each output stage.

Music-wise, the main difference I noticed was a bit more transparency in the bass (more info - you can better hear what's going on with all the instruments), and a bit less sibilance in some recordings. Otherwise, I don't have a lot of difference to report. I feel that the little hum that the amp has, has gotten a bit quieter. You have to be very close to my 100 dB/W widebanders to hear it. I took some distortion measurements before and after adding the PO89ZBs, and you can see a slight reduction in the 100 Hz peak. (second picture with later date is with the filter)
I now measure 480 uV AC at the outputs (both channels the same) - unfortunately, haven't measured that before adding the filters.
I guess the about 900 uF of filter capacity the PO89ZB adds helps a bit.

Next step will be some different front-end boards as well. I have a pair of completed and dialed-in Marauder boards sitting on my bench, but want to memorize the sound with the original front-end a bit more before I swap.

I'm really curious if an Edcor-less front-end will change the measured hum (since it's 100 Hz in the first place, and then also the Meanwell PSU is far away, I guess not), but expect much better square-wave response and frequency responce. Let's see ... :)


Best regards, Claas
 

Attachments

  • VFET R warm 2.83V -3dB gen -6dB atten 211023.png
    VFET R warm 2.83V -3dB gen -6dB atten 211023.png
    74.9 KB · Views: 325
  • VFET R warm 2.83V -3dB gen -6dB atten 211116.png
    VFET R warm 2.83V -3dB gen -6dB atten 211116.png
    75.9 KB · Views: 125
Last edited:
It may be interesting to try adding a modest amount of extra capacitance to the power rail on each output stage. The filter board may work to prevent the SMPS from going into a hiccup current limiting mode. Try values from 1000 uF to 10,000 uF.
Judging by the differences you have already mentioned, you can expect better bass instrument articulation, and improved separation.