Op amp based preamp with discrete output

[hags]

Being new in this hobby I'm wondering if anyone has done this: Built a line driver with something like an OPA627P and then added a discrete buffer to it, using something like 2SK170 jfets or the like.
Are there any commercial products like this?
Seems like it would be a stupid question but I am curious.

[jcx]

LT appnotes 18,21:
http://www.linear.com/designtools/app_notes.jsp

Walt Jung's Classic articles, lower half of page for multiloop/buffererd op amps:
http://waltjung.org/Classic_Articles.html

[hags]

Originally posted by jcx
LT appnotes 18,21:
http://www.linear.com/designtools/app_notes.jsp

Walt Jung's Classic articles, lower half of page for multiloop/buffererd op amps:
http://waltjung.org/Classic_Articles.html

Is this kinda like putting make up on a pig? A discrete jfet buffer output and an op amp gain stage?
I was curious as to the practicality of this, I guess it would be better to go all discrete.
I have an OPA627P line preamp that I'd like to add a buffer to. I thought about just throwing a simple discrete jfet buffer circuit on it.
Would I put the buffer circuit in the feedback loop?

[jcx]

Jung's articles outline the advantages of buffers/high current gain stages inside the feedback loop of good quality audio op amps

normally jfets wouldn't be useful on the output - today's best CFA op amps are probably better than anything a individual diyer can desigh and build out of discretes

[hags]

Originally posted by jcx
Jung's articles outline the advantages of buffers/high current gain stages inside the feedback loop of good quality audio op amps

normally jfets wouldn't be useful on the output - today's best CFA op amps are probably better than anything a individual diyer can desigh and build out of discretes

You might get some argument on that. There are alot of diehards out there.
CFA? current feedback amplifier?

[myhrrhleine]

Hello,

It's been done.
OP-Amps can perform quite well.
Discreets can be better, *if* one is sufficiently talented in design.
But it does take skill.

A buffer doesn't necessarily gain much. Op-Amps can drive modest loads.

Personally, I prefer an OPA134 OP-Amp.

[QSerraTico_Tico]

Originally posted by jcx
LT appnotes 18,21:
http://www.linear.com/designtools/app_notes.jsp

Walt Jung's Classic articles, lower half of page for multiloop/buffererd op amps:
http://waltjung.org/Classic_Articles.html

Ironically the buffers presented in the LT1088 datasheet are more interesting.

[Gordy]

It would seem reasonable to use a good op-amp for gain and then an open-loop buffer to separate the op-amp from the line-driving work. In this way any issues with the line (such as RFI pick-up) do not get applied to the op-amp feedback stage. The buffer acts quite literally as a buffer.

If done this way (i.e. with the buffer outside, and hence not corrected by, the feedback loop) then the buffer must be very linear by basic design.

I would recommend that you search the net (and this forum) for 'diamond buffer'.

An interesting example is provided by forum member PMA on his site: http://web.telecom.cz/macura/dispre.png (the right hand part of the circuit is the output buffer)

Good luck.

[hags]

Originally posted by Gordy
It would seem reasonable to use a good op-amp for gain and then an open-loop buffer to separate the op-amp from the line-driving work. In this way any issues with the line (such as RFI pick-up) do not get applied to the op-amp feedback stage. The buffer acts quite literally as a buffer.

If done this way (i.e. with the buffer outside, and hence not corrected by, the feedback loop) then the buffer must be very linear by basic design.

I would recommend that you search the net (and this forum) for 'diamond buffer'.

An interesting example is provided by forum member PMA on his site: http://web.telecom.cz/macura/dispre.png (the right hand part of the circuit is the output buffer)

Good luck.

I'm familiar with that type of buffer. One of the criteria for this is that there be no coupling caps in the signal path.
I currently have an LME49710 with a BUF634T as a buffer. I have no coupling caps anywhere in the signal path as I don't need them.
I think using something like a jfet source follower outside of the feedback loop would need coupling caps.

[Ryssen]

I currently have an LME49710 with a BUF634T as a buffer.

I had that too,to drive my headphones,the BUF634 sounded to "dark" for my taste,so now i use a QRV-05 Discrete buffer.
Inside the loop..