Advantage of jfets in LTP ?

[AndrewT]

Why do you believe that a dual package has better distortion performance cf. a pair of singles?
Have you read some technical literature that you could link to?

 

[Bigun]

In most gnf amps with reasonably high feedback factors the LTP is not experiencing large differential input signals - so they are going to be operating where they are less sensitive to small differences between devices providing they are nicely biassed into Class A. Perhaps this helps explain why matching etc. is less critical.

 

[Ian Finch]

I think the only "parameters" affected by a monolithic or close-coupled LTP pair is thermal uniformity and that relates to drift of current balance, DC offset and anxiety in audiophiles. Considering the minor role this actually plays in audio amps, I'd only use them in Jfet LTPs where thermal effects can be more dramatic. That's not likely unless we select SMD parts now, so its moot, IMHO.

I recall the performance of a supermatch pair; the venerable LM394(H) which was credited with all kinds of superior performance in low level sensing but, from the datasheet, it seemed aimed at very low level instrumentation. I still use an MC headamp with these. It is certainly quiet and good sounding but that is not related to thermal balance.

Yes, I've read several forum posts where precise balance of the LTP is claimed to do wonders for stereo imaging and how the soundstage grew x-fold times etc. - just like when a bypass cap. was swapped for an audio grade one. Any reliable evidence?

 

[davidsrsb]

A jfet LTP with low Vgs devices can be operated at the zero temperature coefficient point (Vp-0.6V), avoiding any thermal distortion effects. Whether this is important is another matter.

 

[Telstar]

Why do you believe that a dual package has better distortion performance cf. a pair of singles?
Have you read some technical literature that you could link to?

Andrew, I just said that the k389 have DIFFERENT specs from the k170, likewise for the complementary part. And that because I have no such parts I didnt experiment in the real world (only simulations which are as good as the models themselves).

 

[Telstar]

In most gnf amps with reasonably high feedback factors the LTP is not experiencing large differential input signals - so they are going to be operating where they are less sensitive to small differences between devices providing they are nicely biassed into Class A. Perhaps this helps explain why matching etc. is less critical.

My case is the exact opposite (0gnfb and no preamp), so the matching seemed very critical.
Also let's not forget that jfets provide about 1/10 of gain of bipolars in a similarly optimized LTP. I believe only Charles Hansen utilize an IPS with gain, and I bet he's using bipolars (except maybe a jfet cascode).

 

[AndrewT]

It could, if they were the same transistors, but they are not.

which parameter does the dual perform better than selected singles when there is no specification for the dual's performance?

Distortion, according to spice models.

Why do you believe that a dual package has better distortion performance cf. a pair of singles?
Have you read some technical literature that you could link to?

I just said that the k389 have DIFFERENT specs from the k170,

I said two selected singles can perform better than a dual.
You disagreed.
It appears the only evidence you have is Spice Model comparisons. Could the Spice Models be inaccurate?

 

[Telstar]

I said two selected singles can perform better than a dual.
You disagreed.
It appears the only evidence you have is Spice Model comparisons. Could the Spice Models be inaccurate?

Oh yes they can be. But the datasheets are also different.
Let's hope some real-world tests will be showed in this thread

But yes, you're right, in some circuits they may perform the same or better.