Sony vFET Amplifier Part 2

Jump to Latest
#463
WalterW said:
........
What I understand from the CSX1 article is that the input transformer would not be very capable of loading a few paralleled VFETs. So was hoping this new frontend is.
Click to expand...

there is always possibility to buffer* transformer secondary (NB that 95% of them (xformers) need proper loading for best results)

though ...... when done properly , it's hard to beat** xformer in multiple duty role

* if needed
**read - vividness
 
#464
WalterW said:
ZM, I 'm not a designer and have a simple brain.....I was thinking of doing this: replace the N and P VFETS, top and bottom.
And bias the VFETs different.
Would this work?
Click to expand...

I like your thinking, and not sure if its do-able, but what I would love to see is the Sony in the Aleph topo, alas I'm just a greedyboy.. and a fan of the Aleph. Line level ready single-ended SIT goodness, with the typical PS.

Even better if it involved k170s and the plethora of sj28s still available. The Aleph JPVFET. It was coined here lol.
 
Last edited:
#465
res07njc said:
I like your thinking, and not sure if its do-able, but what I would love to see is the Sony in the Aleph topo, alas I'm just a greedyboy.. and a fan of the Aleph. Line level ready single-ended SIT goodness, with the typical PS.

Even better if it involved k170s and the plethora of sj28s still available. The Aleph JPVFET. It was coined here lol.
Click to expand...

Since all I have are 6 sj28s, that's my dream! 🙂 Or a topology that incorporates the Tokin 2sk180 that is still available.
 
#467
Zen Mod said:
in other words - ref to pic in #461 - if you bias output Fets properly , and if you cascode them properly (other way around ) and if you take modulation(s) (two phases) from Q15 and Q16 sources ....... it'll work

:clown:
Click to expand...

Hello Zen Mod. The cascoding technique of the power output SITs as seen in the schematic of WalterW [as shown by using R17, C7 and R18, C8] are counterproductive to operation as you noted above. The other way around which you suggested works. It has consequences; more rework as noted earlier. By example, the pair of 2SJ28s will be sandwiched between two power Mosfets, one at their sources [upper] to increase PSSR, and the second Mosfet [lower] at the drains of the SIT to lock their Vds constant [cascode]. The power output of the resultant amp may be decreased if Vds of the lower Mosfet is kept to a minimum active region.

Another consequence is the dynamic impedance of the 2SJ28s [for example] is now in series with the dynamic impedance of its lower Mosfet which is acting as a common gate amplifier. Does one wish to compound voltage variable resistors [SITs] with a voltage variable current source [Mosfet]? Maybe a different sound will emanate from this new amplifier. More bench and listening tests.

Best regards.
 
#469
dantwomey said:
Any PCBs on the horizon?
Click to expand...

Tea-Bag has a functional design, ready to go as far as I know, and I'm not
looking to reward him by stepping on his toes.

That said, I do have one, but I think I'll revise it so that it will fit the DIYAudio
heat sinks, kindly sampled to me by this site, and also to accommodate
TO-247 parts for the regulators and possibly single pairs so that more can
be built and matching is not required.

The parts have adequate dissipation to support this, and the difference in
output stage transconductance is only 30% (you halve the transconductance
by two with only one part, but then you increase it by the square root of 2
because the single part has double the bias)

Anyway, I'm working on it. Those of you who want to do this now might
want to contact Tea-Bag about availability of the boards. I do not expect
the revisions to perform better - I'll be happy if they are about the same.

😎
 
#470
Nelson Pass said:
Tea-Bag has a functional design, ready to go as far as I know, and I'm not
looking to reward him by stepping on his toes.

That said, I do have one, but I think I'll revise it so that it will fit the DIYAudio
heat sinks, kindly sampled to me by this site, and also to accommodate
TO-247 parts for the regulators and possibly single pairs so that more can
be built and matching is not required.

The parts have adequate dissipation to support this, and the difference in
output stage transconductance is only 30% (you halve the transconductance
by two with only one part, but then you increase it by the square root of 2
because the single part has double the bias)

Anyway, I'm working on it. Those of you who want to do this now might
want to contact Tea-Bag about availability of the boards. I do not expect
the revisions to perform better - I'll be happy if they are about the same.

😎
Click to expand...

Would you share your minor value modifications for the existing design 🙂
 
#475
Antoinel said:
Hello Zen Mod. The cascoding technique of the power output SITs as seen in the schematic of WalterW [as shown by using R17, C7 and R18, C8] are counterproductive to operation as you noted above. The other way around which you suggested works. It has consequences; more rework as noted earlier. By example, the pair of 2SJ28s will be sandwiched between two power Mosfets, one at their sources [upper] to increase PSSR, and the second Mosfet [lower] at the drains of the SIT to lock their Vds constant [cascode]. The power output of the resultant amp may be decreased if Vds of the lower Mosfet is kept to a minimum active region.

Another consequence is the dynamic impedance of the 2SJ28s [for example] is now in series with the dynamic impedance of its lower Mosfet which is acting as a common gate amplifier. Does one wish to compound voltage variable resistors [SITs] with a voltage variable current source [Mosfet]? Maybe a different sound will emanate from this new amplifier. More bench and listening tests.

Best regards.
Click to expand...

This is way over my head 😀
First I didn't know what ZM meant at all with cascoding the SIT's properly. After reading the article again and Antoinel his comments, I understand that the IRF's cascode the SIT's.
I think I understand a little bit what R17, C7 and R18, C8 are doing, but it stops here for me.
Why ZM suggest to take the modulation from Q15 and Q16 sources?
Is it because Q15 and Q16 than become common drain and the OS is common source, so the whole amp is not inverting anymore? Or do you mean something totally different ?
Sometimes explanations on DIYAudio are given in a short sentence, special code, which are difficult to decipher 😱 for me...
 
#476
Here is the concept art for the VFET v2 board dubbed "AL"
I have a couple people besides myself looking for errors on the boards. It looks like based on Botte's work that we will need to widen C1 and C2 a bit.

Comments are welcome.

17.png


Click this link below for a larger pic.

http://www.diyaudio.com/forums/gallery/showfull.php?photo=12937
 
#477
WalterW said:
This is way over my head 😀
First I didn't know what ZM meant at all with cascoding the SIT's properly. After reading the article again and Antoinel his comments, I understand that the IRF's cascode the SIT's.
I think I understand a little bit what R17, C7 and R18, C8 are doing, but it stops here for me.
Why ZM suggest to take the modulation from Q15 and Q16 sources?
Is it because Q15 and Q16 than become common drain and the OS is common source, so the whole amp is not inverting anymore? Or do you mean something totally different ?
Sometimes explanations on DIYAudio are given in a short sentence, special code, which are difficult to decipher 😱 for me...
Click to expand...

Hello WalterW. I am always glad to clarify my previous comments. I'll post a hand-drawn attachment showing a simplified schematic on cascoding the SITs in your schematic per the direction of Zen Mod and my previous explanation.
 
#478
Tea-Bag said:
Comments are welcome.
Click to expand...

Since there is some real state left next to the output board bias trimpots, do you think you could flip them 90 degrees and mix in the 3296P's footprint? I prefer the horizontal version of the trimpot, with the adjustment screw pointing to the amplifier case's top.

http://www.bourns.com/docs/Product-Datasheets/3296.pdf

Also, what are the measurements of each board?
 
You must log in or register to reply here.
Top Bottom