another bizarre audio circuit

At high frequency, you get noise from the output feeding back via Miller capacitance. Cascoding can help that.

Cheers

Phil Hobbs

Indeed, when I discovered their existence back in 2001, I could hardly believe the datasheet. I'd never seen a JFET with a ratio of yfs/Cin that high. That, and with only 0.8nV/rtHz input-referred noise. I love it.

Jeroen Belleman

You need to spend some hundreds of dollars in one buy in order to get close to that price/unit, don't you?

Jon

--- Being the grunt work that it is, then, one should be able to take a schematic devoid of component values and then hand it over to a grunt with the expectation of getting back a working circuit some time later?

One thing about topologies which is misleading is that if the component values aren't defined, the topology might look fine but the realization of the circuit will be impossible under economic or technical constraints.

An example which springs to mind is a circuit which was posted some time back which looked good, but which on closer inspection you said needed a choke with an inductance of near 1 henry, as I recall, and a Q of about 200 somewhere in the audio range.

I went looking for one, just for grins, but found only unobtanium so, unless I missed something, (got a source?) your guess was wrong and the topology bogus.

--- JF

"One" should? I couldn't say. I usually design all of a circuit myself, but if we brainstorm/whiteboard circuits, we may include values, or one of the people might go off and finish it. Could be me, could be somebody else, whoever is responsible or volunteers. We usually check one anothers' work before we release a board. We have no techs, so everybody does grunt work.

But all circuit designs start with ideas and topologies. If someone has never seen a folded cascode, or an opamp's V+ used as a signal output, or a bootstrapped photodiode, all I may need to do is present the concept, and then they can run with it.

The laser controller I'm working on now, all the resistors on one sheet are 100 ohms and all the caps are 0.33u. That reduces the grunt work considerably. This scheamtic will run about 25 B-size sheets, and I'll do maybe 18 or so of them, with other people contributing others, like the FPGA, the ARM, and the PCIe interface. This is a rush job, 7 weeks from start to 1st article, so several people are tossing in sheets.

Sure. A complete electrical/thermal/mechanical design ought to be done before something goes into production. Lotta grunt work. Newsgroups are for playing with ideas without consequences.

Bogus? Because you can't order some part? Nothing wrong with playing with a topology if it might work. Whether it's practical or affordable is part of the downstream analysis. If every idea has to be immediately and exhaustively analyzed for cost and parts availability and subtleties, you won't come up with many ideas.

As usual, you're just being bitchy. Why are you so hostile to playing with ideas? Why do you refuse to do it yourself?

John

Millions of the "GE" circuit have been used for decades. The mosfet hybrid is a very reasonable headphone amp.

Post a circuit, doofus. You've forgotten how to do anything but whine.

John

All you ever criticize is my personality. You said nothing about this circuit. On the rare occasion when you do post a circuit, or critize one, you're usually wrong the first few times.

Old hen.

John

You're whining about everything but the circuit. Diversionary tactic.

Old hen.

John

Stop whining and clucking about personalities and design some electronics. That's not a straw man, that's what this ng is about.

Think about what happens to my bizarre headphone amp if the inductor is replaced with a constant-current source.

John

Good. I'm glad we finally agree on something.

I suppose it would be cruel to trick you into discussing electronics. So, keep on cluckin' !!!

John

I didn't "infect" this thread, I started it. So why have you posted so much cluckey blather here?

You refuse to discuss this circuit, then you attack me personally for not doing give-and-take discussion of this circuit!

John

I bought myself a reel for $600 or so from Newark. Given that JFET makers have been dropping like flies lately, I'll probably buy a couple more reels over the next few months. That way I can keep designing them in even if they go away. (Crossed fingers.)

Cheers

Phil Hobbs