Watt Sucking Fireball Series

[Roadbagger]

I was just having a chat with John Hall and he was commenting that the 1-1AMP circuit posted might not have enough voltage across the input fets to optimize noise and that the higher GM fet cascoded would consequently have a higher noise than the input fet placed in series with the same. YES. Noise increases with GM and the input FETs like any low noise FET will have a "sweet spot" when the right Vds is employed. This circuit illustration is for discussion purposes. The front-end is obviously a borrowed thing that I see frequently in a number of designs in various forms. Nothing original here, just a familiar place to start with other points to follow. That front end will be further assessed in time, but I suspect that somewhere in this sphere is a fairly thorough analysis of it. Illustrated first is the use of the matched transistor pair full Wilson mirrors, a habit of mine, current folding mania, then the output buffer. The advantage here is twofold. Normal NPN-PNP emitter follower buffers will, as I mentioned before, reflect the load impedance backward divided by the beta. The other phenom is an emitter follower's tendency to become unstable with capacitive loads due to input negative resistance C loading. The follower employed here is a variation of a circuit conceived by R. Fulks at General Radio. It's advantage is that here, a FET is used as the follower in a N-channel/PNP source/emitter output structure. The negative resistance loading effect is eliminated by the source follower, low noise impedance transform is provided by the FET and the FET source load is a low impedance emitter slaved to the source current demand by the upper PNP Iq sense feedback. The folded current summing node is now able to have it's highest gain potential being loaded by only a FET input capacitance and virtually no load resistance. Often a 20-50 pf cap is added to the current summing node to lower upper bandwidth and prevent oscillation. Slew rate is then dominated by I/C of the summing node. Not much different than the internals of an OP-AMP.

K-wood

PS: The Fulks buffer is also the basis for that cockamamie amplifier further back called the 2N4401-3, as an experiment in thinking I built it into a power amp with junk on the bench and modified it to have some gain.

[Roadbagger]

Ya know,
There are a number of books that I think are paramount to developing a focused perspective about audio reproduction. Many of these are available on the Web as downloads and some are just good to acquire as a shelf item to refer to. For those of you who have not stumbled onto this site, here is a large amount of time involvement you will find in conflict with the other things you know you should otherwise be doing. Technical books online
I think the Radiotron handbook 4th edition is also here but then, that’s one that I think needs a shelf. As well is the Audio Cyclopedia by Howard Tremaine copyright 1959 & 1969 and one that pops up from time to time, Amplifier Handbook by Richard F. Shea copyright 1966. Though the technology thoroughly illustrated here, is, as some feel, “Retro” the fundamental principles are basic physics and invariant with time. I can still hear the effect that the JBL Paragon system had on me driven at that time by a bias stabilized triode zero feedback amp way back. That effect of being there is always my goal and I think that linear transfer function is not as difficult a task as complex circuitry would otherwise represent. Though, it seems we usually get remarkably complex in order to arrive at the simplest solution. A speaker for instance, the magnetic path formed to push a field flux across a gap should be a highly permeable something and have a relatively constant cross-section so as to make the gap field uniform, past practice being steel. That situation only exists in practice with an electromagnetically generated field (EM). Permanent magnets (PM) all have permeabilities of unity even though when polarized they can have high B and H numbers. So the reaction of a voice coil to the EM gap flux is a tighter coupling than a PM speaker because it has no secondary field leakage gap made by the magnet. Though there are great advantages in using NdFeB magnets using a shorter gap across the magnet thickness to energize the VC gap and making a tighter coupling than thick Barium Ferrite ring plates, the coupling coefficient will not be as easily accomplished as that in a EM speaker field assembly. Though not immediately obvious, the EM structure for high clarity is simpler than a PM pole assembly to accomplish the same driver. Better coupling and damping in the “motor” action of a voice coil. So maybe it is “retro” but I like it, and I think there may be a lot of lost engineering that on second inspection might be an improvement toward the goal of ultimate fidelity………………….Massless Speakers.....BAF 2010
Prattling on…………………K-wood

[kenpeter]

Does this mean investing in Maglocks for future AMT or Linaeum experiment?

[Roadbagger]

Originally Posted by kenpeter

does this mean investing in maglocks for future amt or linaeum experiment?

nope !..............

[albin]

I don't understand, but I do somehow.
Regards Max Albin

[kenpeter]

Originally Posted by Roadbagger

nope !..............

Field coil Maggie?

[Roadbagger]

The real question being..........When is a Balance control set to anything other than center anyway? Why do we have this knob? If it's out of balance, something is wrong. Fix it ! !

[kenpeter]

Obvious, the knob its there so you can turn it back and forth to work the grit out.

[john curl]

Everyone, I want to introduce you to K-wood Rough, one of the best engineering consultants in Silicon Valley. He can teach you, if you let him.

[Roadbagger]

Originally Posted by john curl

Everyone, I want to introduce you to K-wood Rough, one of the best engineering consultants in Silicon Valley. He can teach you, if you let him.

Huh? maybe you have that wrong John, isn't "Insultant" more the case here.........Tarring reality and feathering it with symbols,................... K-wood