John Curl's Blowtorch preamplifier part II

[syn08]

Originally Posted by john curl

Probably good enough for you? (referring to your phono stage)

I don't know, that's why I'm asking.

Anyway, one of the beauties in my phono stage is that it doesn't need any expensive dual or handmatched JFETs, so a J74 equivalent together with the LSK170 would do just fine.

[john curl]

That is interesting.

[Curly Woods]

Originally Posted by syn08

I don't know, that's why I'm asking.

Anyway, one of the beauties in my phono stage is that it doesn't need any expensive dual or handmatched JFETs, so a J74 equivalent together with the LSK170 would do just fine.

Could you hear a difference if one was obvious? Just curious as to how you can evaluate how a component sounds when you do not believe that any differences could exist in the first place. On your web site you allow others to provide you with testimonials as to the sound quality of one of your designs, but you do not seem to agree with the possibility of sonic differences in components. I am confused.

[john curl]

I always thought that the self-noise of the input devices was the most important source of overall noise, that can't be fixed.

[syn08]

Originally Posted by john curl

I always thought that the self-noise of the input devices was the most important source of overall noise, that can't be fixed.

Ok, I'm in a good mood today, so here's a free lesson:

In a two stage amplifier, the input stage dominates the total noise only if the input stage gain is larger than 1/3 the total gain (closed loop if any feedback is involved, not the case with BT). Why 1/3? Because noise adds squarely and (1/3)^2~1/10, so that the second stage noise can be neglected (it's contribution to the input referred noise is small).

Now, what is the BT input JFET stage gain? It's the ratio of the drain and half of the source resistors, that is, about 1.6. And what is the BT total gain? If memory serves, is about 7.6 (ratio of the output load and half of the source resistor). The condition above is not fulfilled, so the input stage is not the dominant noise source. It's the output MOSFET noise dominating the input referred noise.

I'm not even talking here about what's the purpose of using 4 x 1nV/rtHz JFETs (equivalent to 0.5nV/rtHz), with a source equivalent noise resistor of 130ohm (that is, 1.5nV/rtHz).

Of course, if the BT schematic that was posted is incorrect, so am I.

[john curl]

The Blowtorch had a phono stage as an optional extra. The pictures of the Blowtorch show this phono stage. The visible part of the phono stage was first designed over 25 years ago, and has an equivalent noise of about 9 ohms or 0.4nV/rt Hz. It is NOT the line amp, which has a 25K pot in front of it, in any case. The input stage of the phono preamp consists of exactly the same devices that I gave to the consultant to Linear Systems, the 2SK146 and the 2S73. The actual noise limitation was the 2.5 equivalent ohms that was the effective noise of the 10 ohm wirewound pot that I initially used for self biasing the jfets and to give AC balance for lowest 2'nd harmonic distortion. The input stage current was approximately 25 ma. If I had used a 5 ohm pot, or even a 1 ohm multiturn pot, I could have even lower noise, but I might not have the range for nulling the 2'd harmonic that is present.
You are quite correct that the BT line stage is more noisy than I would like, but it is still essentially inaudible.

[syn08]

Originally Posted by john curl

The Blowtorch had a phono stage as an optional extra. The pictures of the Blowtorch show this phono stage. The visible part of the phono stage was first designed over 25 years ago, and has an equivalent noise of about 9 ohms or 0.4nV/rt Hz. It is NOT the line amp, which has a 25K pot in front of it, in any case. The input stage of the phono preamp consists of exactly the same devices that I gave to the consultant to Linear Systems, the 2SK146 and the 2S73. The actual noise limitation was the 2.5 equivalent ohms that was the effective noise of the 10 ohm wirewound pot that I initially used for self biasing the jfets and to give AC balance for lowest 2'nd harmonic distortion. The input stage current was approximately 25 ma. If I had used a 5 ohm pot, or even a 1 ohm multiturn pot, I could have even lower noise, but I might not have the range for nulling the 2'd harmonic that is present.

That's news to me - thanks. So, there is a BT phono stage option, is there any difference to the Vendetta phono stage? Following your description, it seems like these differences (if any) are very small.

I think it is also important to add for those looking at the BT line stage that using very low noise JFETs is not really required in this application.

[john curl]

The BT phono stage was a modified Vendetta Research phono stage. I use it, every time I play a vinyl record.

[john curl]

Perhaps, NOW, it might be more obvious why I am concerned about subtle compromises in noise with the Linear Systems devices. I do NOT design BT line stages any more. I am working on phono stages, and super low noise line stages, instead.

[ikoflexer]

There's something very confusing to me here. You're talking about sub 10nV/rtHz self noise for the input stage; is the supply noise of that stage anywhere near that value? Does it matter if it's not?