SE Preamp With Moderate Current NFB

[VladimirK]

This simple SE preamp schematics with moderate local current NFB provides rather high voltage amplification (22dB) and swing (15V AC undistorted sine wave at the output), nearly 1 MHz passband, shows stable and nice behavior with square wave signal (correction cap in NFB is not needed). At the square wave figures, small bump belongs to input signals. Moderate level output impedance 440 Ohms allows direct coupling to the output stage. This could easily be 2SK1530 or 2SJ201 transistors, so this preamp is easily extendable to an integrated amp.
Voltage amplification can be decreased with corresponding lowering of the output impedance.
With the values indicated, the only adjustment needed – selection of VT1 2SK246 with Idss close to 5mA.
The input cap could be removed, but this would bring up definite difficulties and limitations.
At the moment, I have only measured a prototype, and can not say about its sound, but with a shunt power supply, and eventually with the follower-like output stage, good sound is expected.
Figues show sine and square signals at 100kHz, levels of inp and outp signals are adjusted. Also to be mentioned, that clipping behavior is quite soft, no too sharp edges.

[juma]

Very nice Vladimir
Topologically almost identical to Aleph L 1.2 preamp - I built it with k170 and IRF9610 few years ago to drive F4 and it sounded really good, you won't be disappointed.

[VladimirK]

This version of the schematics differs by power supply polarity and types of transistors used, two times higher VT2 current, also the source degeneration resistor was removed from the source of VT2.
The result is definite improvement, Zout has become 75,6 Ohms, it is also faster now. With negative ps rail choice of transistors is more favorable for performance. Clipping is soft.
The pictures show clipping at 1kHz, 100kHz square, 500kHz square, 1MHz sine (input signal with small bump, input and output levels adjusted)
Negative PS rail is also more convenient for coupling with follower-like output stage.

[VladimirK]

Originally Posted by juma

Very nice Vladimir
Topologically almost identical to Aleph L 1.2 preamp - I built it with k170 and IRF9610 few years ago to drive F4 and it sounded really good, you won't be disappointed.

Thanks, Juma
As for sound, my experience says that it is unpredictable from THD or nice square wave reproduction. More or less reliable recipes for this, are shunt-like stages, shunt power supply arrangement, compact PCB, minimum elecromagnetic coupling, that can cause some post-effects like in audio cables. I had several niciest schematics, without output caps, but did not manage got them sounded.

Next step in developing this schematics, I plan to add shunt PS. Then follower-like output stage.

[VladimirK]

The Shunt PS to the 2SJ103-BSP110 preamp version is ready now, I'll post the schematics and some measurements in few days. In particular, I am curious, what passband will be with relatively tuff output load 1k...3k3.
Also I would like to pay attention to the differences between the 2SJ103 and 2SJ74 (2SJ109) j-fets, since we frequently put the last into input stage of preamps. Being used in common source gain stages, input capacitance of 2SJ74 can figure up to 600pF (according to Borbeley measurements), although theoretically it is only Ciss + Crss x voltage gain = 110pF + 32pF x gain and presumably must not go that high. Fixing this problem is cascoding and NFB.
For the 2SJ103 j-fets, Ciss=18pF and Crss=3,6pF , and this allows to have quite a wide passband with much lower NFB level, with no need in NFB compensation cap. The last situation is empirically more favourable for good sound, IMHO.

[VladimirK]

just have forgotten to add along this line, that the mentioned values of parasitic capacitances of j-fets (non-linear with applied voltages) can explain quite easily, why many people prefer tubes. In this relation, I think about using HF transistors more and more, since they are closer to tubes, than general purpose ones.

[WuYit]

Hi,
my choice would clearly be the 2SJ103. Crss is most detrimental as it suffers from voltage amplification lowering the bandwidth. I would implement a cascode.

[MiiB]

This concept is further developed into a power amp in the thread...

JFET input, MOSFET VAS, LATERAL output = Perfect!!

[lineup]

Originally Posted by WuYit

Hi,
my choice would clearly be the 2SJ103.

WuYit
and where do we buy 2SJ103
farnell - no
digikey - no
mouser - no
We call these exotic. Those almost impossible to find. Except in fantasy.

[bracla]

Originally Posted by lineup

WuYit
and where do we buy 2SJ103
farnell - no
digikey - no
mouser - no
We call these exotic. Those almost impossible to find. Except in fantasy.

It seems, that ampslab have them in stock.
See: Audio Signal JFETS - Buy OnLine @ AmpsLab

Claus