Thanks Esperado on your input of the Siliconix app note of the modification of the ne5534. They were trying to sell jfets, not IC's with that app. The real problem with that modification was that it was NOISY because the second stage in the IC was so noisy. It did eliminate thermal feedback to the input stage, however.
It´s the JC-2 (or ML-1, not JC-1 as I wrote before), and it´s not a wish list,
it´s what you can get with discrete. For IC opamps it seems you can never get
low noise, Fet input, high slew rate and a class A output stage in _one_ package.
John, thank you for the detailed explanation. I have mixed up JC-1 and JC-2. What I wanted
to say was "a JC-2 line stage built with modern (well modern, the P-channel part is already
out of production) Thoshiba Fets". I´m not aware of a integrated solution that will match that.
Last edited:
I think I lost it years ago. IIRC you pull the inputs to Vee to disable them and just put a JFET long tailed pair on the comp/offset pins.
This demonstrate one time again that the first stage of an amp is the most important, regarding the poles. I believe the slew-rate of this composite OPA was better ?
I am thinking to conventional Power amp, and the fact that many amp use the same kind of transistors and schematics for their input stage. And i wonder why CI manufacturers does not propose complementary fast and high voltage matched pairs, both BJTs and FETs in a integrated circuit optimized for temperature matching.
Scott ?
Last edited:
The question was not about doing in ICs the same things (topology, biases, differential complementary stage, etc...) you can do with discretes, but to provide the same level of objective performance. When you evaluate a discrete solution vs. an IC solution you don't necessary know about topology (Class A) output stage bias current, etc... OTOH, knowing such details in advance and evaluating based on the "sound quality" only is equivalent to peeking in a blind test.
On the same note, you could argue there isn't any IC providing 200W in 4ohm load, so ICs must be inferior. As much as I can argue you can't build a 1000V/uS gain stage (or anything close to the THS4631, like 325MHz unity gain frequency and 80dB gain) with discretes on a regular PCB.
So, as I said, I see this going nowhere, therefore I'll stop here.
That's a good reason.
What is a 3.6? What is a Big Bembeh? Yes please to your list.
I am assuming that the "3.6" refers to the Magneplanar 3.6 (a planar-magnetic/ribbon hybrid loudspeaker from Magnepan Incorporated (MG 3.7)), a loudspeaker that presents a fairly resistive, largely 4-ohm load.
Sorry, I have no idea what consitutes a "Big Bembeh"...
Good luck!
They've been around a while. More than 2 decades if I recall correctly.
Are the fet numbers of any interest? I know the rep, and could ask.
jn
- Status
- Not open for further replies.