WOTS Pre-amplifiers

The IXCP10M45S is sold as a current regulator so the usual mosfet parameters are not shown in its datasheet. It could be a normal depletion mosfet but I wouldn't take the risk of using it as a follower. It cannot be used as a CCS in the PSU, the current to feed the zener is 1.7mA, and the IXCP10M45S has a minimum of 2mA. It could be used as a CCS (10mA) in the follower stage, but I don't know about the result. It would be a $3 saving each. Again, I wouldn't take the risk.
I bought a bunch of 10 IXTP01N100D and selected them for the minimum |Vgs| to use as followers and maximum |Vgs| to use as CCSs.
An STF3LN80K5 (cheap but good) can be used as a follower (also as a voltage regulator in the PSU), but the midpoint will go down around 5/6V because of its positive Vgs. This is only cosmetic, with no impact on the sound.
Before some ask: I tried the DN2535/40, as a CCS, but it seems below a certain current it doesn't work. Some in this forum say its sound is not the best.
 
I am thinking to use two J113 parallel as the follower.

When diying BabyHue tube power amp, two mosfets with samll Ciss(300p) and Crss(0.85p) were used as buffer (follower with BJT CCS) between two tube stages(12AX7 and KT88). Subjectively speaking, I think it sounded not like a tube amp(I might be wrong).

If just from dadasheet, JFet has much smaller Ciss and Crss is comparable. I think it is worth trying.
 
I wouldn't use J113 because of its Vds. The Tre used the ZVNL120A. In the Talia I used the IXTP01N100D which has a low Ciss. However, The Ciss is negligible because the follower works in common drain, and the previous stage only sees a couple of pF.
 
My WOTS DIY

The schematics I used is marigno's WOTS original, JFET is BF244B and M1 is ZVNL120A.
The system consists of 4 PCBs: transformer, rectifier/smooth, capacitor/regulator and preamp.
 

Attachments

  • Fig05.jpg
    Fig05.jpg
    39.6 KB · Views: 111
The regulator is similar as marigno's virtual battery. I like to use LED array as a portion of reference voltage as in Salas DCSTB power supply. By shorting out some LED segments, the output voltage can be adjusted. As show in the photo, 3 LED have to be jumped out to get 60VDC voltage.
 

Attachments

  • Fig13.jpg
    Fig13.jpg
    101.4 KB · Views: 129
  • Fig14.jpg
    Fig14.jpg
    161.7 KB · Views: 129
  • Fig15.jpg
    Fig15.jpg
    197.5 KB · Views: 123
Although my DIY is the original one, but the designed PCB can be used for multiple configuration, like Talia by adding or omitting some components. GND is start style like described in marigno's design. The output delay circuit is different. I used a version I did and tested.
 

Attachments

  • Fig16.jpg
    Fig16.jpg
    68.4 KB · Views: 128
  • Fig17.jpg
    Fig17.jpg
    167.1 KB · Views: 122
  • Fig18.jpg
    Fig18.jpg
    305.3 KB · Views: 87
  • Fig19.jpg
    Fig19.jpg
    257 KB · Views: 123
THD at 1, 3, 5, 10V output. When at lower output, it has almost pure H2 distortion.
 

Attachments

  • Fig24 BF244 ZVNL 1V.jpg
    Fig24 BF244 ZVNL 1V.jpg
    81.7 KB · Views: 66
  • Fig25 BF244 ZVNL 3V.jpg
    Fig25 BF244 ZVNL 3V.jpg
    81.9 KB · Views: 68
  • Fig26 BF244 ZVNL 5V.jpg
    Fig26 BF244 ZVNL 5V.jpg
    81.6 KB · Views: 64
  • Fig27 BF244 ZVNL 10V.jpg
    Fig27 BF244 ZVNL 10V.jpg
    82.1 KB · Views: 63
After one week listen, adjustment, listen, change, it the time to report my first impression.

This is indeed a very excellent preamp. I have built few kinds of preamps, so far it is the one of two I like at most(another one is M7 tube preamp).

Also my personal impression, BF256B is slightly better than BF244B, so I use BF256B in my final configuration.

One thing is worth to mentioning is that ZVNL120A is very hot. I think I will increase its Rss to 3k or so reduce Pa.