Discrete vs opamps... - Page 9

missed this :

Called an headroom and where even a boatload of good Class-A amps hold a bit back and go into Class-AB for music peaks.To use the entire output all class a all the time means added cost,possibly a shorter lifespan for the parts,added cost ( )..

The options here are stay in class at all times but back off on ultimate power output thus posssibly strangling a peak signal event (runs out of steam),retain a high output in pure class a then (solid state) increase the heat sinking and venting accordingly (plus the added power upply emphasis,again $$$) or stay in class A up to a certain point then for peaks bounce into AB.
Class AB is not as destructive to the sonics at high levels but at low so the higher up into the signal you can stay in call a the better but real world cost and real world restraints means sometimes you need to comprimise where it does very little damge.

You "headphones can not handle 1 Watt of power" example means little in an actual system where you may only need 300 mW for most music but may hit much higher than that for music peaks.Add to that you don't want the peak right at your power limit or what you get is either compression or clipping and neither is good though I will take compression over clipping personally.
A 2W amp is not overkill for even a Grado or I would have stayed with the original 1W design.The power increase opened up the sound and that is what ity is all about not what might be but what actually happens.

BTW the 2W class A mosfet amp uses HUGE heat sinks for a headphone amp at 5x6x1.5 and still they get hot ! My LH0066 class A buffer the same size but with many more fins and it is only warm to touch.If 2W AB I could use a heat sink 1/4 that size and still the heat sinks would be totally cool to the touch so a real worl example on just how much heat can be geberated by even a low power amp and why an I.C. would fry at these extremes

This is all good from a purist's perspective-- i.e. you don't want anything added to or taken away from the music. I suppose I'm not really a purist since anything that sounds "good" is fine with me. Is treble smearing bad? I would say so if what you're after is transparency, but I love my sr71 exactly for that reason -- without the treble smearing I probably would have returned it long ago. But who knows, maybe I'll change my mind after hearing a few good tubes.

Btw, my central station has this marketing flyer:

While the headphone amp definitely sounds very transparent/uncolored, it also sounds very clinical and cold to my ears. I may have to get a Gilmore Lite to compare if this is what discrete SS amps typically sound like.

Do you mean like PPA and M^3 when you say discrete Op-amp?

From all of this class A etc. discussion, have I correctly understood that it is is technically possible to bias a monolithic opamp into class A through all of it's output range, however the issue is heat dissipation in such a small package? Has anyone actually tried taking for example an OPA627, some appropriate cooling solution (thermal past -> heat sink -> maybe a fan) and fully biased it into class A?

I suggested this earlier and he said something about taking the plastic off and finding transistors; frankly, I was so baffled by his suggestion that a monolithic op amp is a plastic covering with a few transistors inside that I didn't bother following it up.

I think he was trying to say the housing of the opamp is plastic and the real circuits lay beneth, and you'd have to remove the plastic to get a metal to metal connection between the heatsink and opamp. Imagine if the Heatspreader on the Athlon 64 was plastic, and only the core beneth it was metal, you couldent put a heatsink on the plastic part. Think back to the old AthlonXP chips that had an exposed core with no heatspreader. Obivously a plastic heatspreader would never be used on a microprocessor, but they can get away with it with opamps apparantly. And like it was previously mentioned, even if you could, with enough effort put a big ass heatsink on an opamp and run it class a, you'd be better off doing it right with a discrete design built from the ground up.

Ahh, okay. Makes sense now. It's a plastic housing? What the hell do they do that for? Seems like it'd be less expensive to fab it all in silicon, seeing as it's a chip op-amp anyway, right?

Well my idea was more that you could probably use small heatsinks to allow the opamp designs to be biased further into class A. My understanding is that increasing the bias is limited to the law of diminishing returns. I am guessing that a large improvement could be gained from current monolithic opamps by adding some small and simple heat dissipating solution, allowing the opamp to go much further up on the output scale before resorting to class AB. Am I crazy?

Thanks...

Just wanted to thank the participants in this thread, I've learned a lot and gathered plenty of food for thought. I think my next amp will be a Gilmore Lite or other solid state discreet design... transparency is very important to me, and I don't feel that I'm gaining all that much with opamp-based amps, even going higher-end than a Cmoy! Sure, they drive the headphones better but there's significant loss of transparency and noticeable colorations that detract from the music -- I'm in the "straight wire with gain" school of thought, and the Gilmore Lite (or similar design) looks like just the ticket.

just skip it and get the dynahi instead

I'll take my Reference over a Dynahi any day Found the Dynahi less resolving with an artificial glow to everything played. That and it burned the cat everytime he tried to sleep on it

On topic err... discrete for the win

I don't know if you knew this, but fewtch [used to] be called "the budget hi-fi guy" so I was referring to the Dynahi PP which is not that much more than a Gilmore Lite.

Anyone know this?

Yes, those are monolithic opamps...

EDIT More correctly, those use monolithic opamps...

Did you feel this before or after reading this thread?