[Nemo de Monet]

I was reading about biasing opamps into Class-A operation a while ago, and saw some comment somewhere about how some of the more desirable audiophile-approved(TM) opamps couldn't deliver enough current to be of serious interest as Class-A audio amps. That got me thinking... has anyone ever tried running a high-power chipamp like the TDA2030, LM1875, LM3886, and so on as a Class-A headphone amp? Full-sized "Gainclone" amps using these chips are immensely popular, and while I rather like tubes in my amps these days, I can see a certain fairly silly appeal in powering your headphones with a, erm, 14WPC amplifier that sucks a half-ampere of power and dissipates 8-10W of heat, constantly...

The TDA2030, as one very inexpensive example, runs happily in the +/-9 to +/-18VDC range, is stable down to a gain of 24db (near enough the standard CMoy gain of 11x), and can supply, ahem, 500ma of current into an 8ohm load, which should keep most Grado owners happy. The LM1875 is similar; stable down to a gain of 10, and can safely output over 1A under most circumstances.

Obviously, these wouldn't be particularly suitable for portable amplifier use, and even at reasonably conservative values (250ma biasing?) heatsinking would be necessary, but it seems like it could be a fairly fun project for people who like moar power at the expense of THD and slew rates and all the other highly technical things that the Gods created the Internet for people to fight about...

 

[FallenAngel]

Noise floor and gain would likely be an issue.

 

[the_equalizer]

Quote:

Originally Posted by Nemo de Monet /img/forum/go_quote.gif I was reading about biasing opamps into Class-A operation a while ago, and saw some comment somewhere about how some of the more desirable audiophile-approved(TM) opamps couldn't deliver enough current to be of serious interest as Class-A audio amps. That got me thinking... has anyone ever tried running a high-power chipamp like the TDA2030, LM1875, LM3886, and so on as a Class-A headphone amp? Full-sized "Gainclone" amps using these chips are immensely popular, and while I rather like tubes in my amps these days, I can see a certain fairly silly appeal in powering your headphones with a, erm, 14WPC amplifier that sucks a half-ampere of power and dissipates 8-10W of heat, constantly... <snip>

If it's power what you crave for (and heck, I love a headphone amplifier that heats up!) I'd recommend a Millett "Starving Student" Hybrid.

This baby has it's output MOSFETs biased in class A at 19 volts@150mA (2.85 watts of power dissipation per channel). The same power is dissipated by each of the tubes, for a total of 11.4 watts of dissipation, constantly...

According to Pete Millett's website it can output 3 volts RMS into 32 ohms at 5% THD which if my math is correct means 90mA RMS and more than 1/4 watt into the Grados (e.g. insanely loud )

Oh!, and it sounds great: dynamic, warm and punchy.

cheers!

 

[Nemo de Monet]

I'm not after power, thanks, nor really looking for an amp that heats up. I've got several variations on the Super Simple 6DJ8 amp that a bunch of us built last year; they dissipate around 15W each, sound excellent, and don't require expensive and eventually unobtanium tubes, either.

Quote:

Noise floor and gain would likely be an issue.

Yeah, if you happen to care about such things. I just figured, given that there's a fairly popular headamp out there that uses the freaking LM386, of all things, someone must have tried even more ill-suited chipamps to drive headphones...

Aperture Science - we do what we must, because we can...

 

[FallenAngel]

I don't know about "popular", but sure, there are definitely crazy things out there. If you want deep class-A, just look at the SOL Headphone Amp as originally designed. It's a single MOSFET burning at 600mA. That's per channel.

 

[jcx]

I've built a TPA6120 Class A amp

general disscussion of Class A output bias of chips, photo's of my amp and circuit ideas futher down in this thread:

http://www.head-fi.org/forums/f6/ad8...8/#post2261919

 

[robzy]

Quote:

Originally Posted by jcx /img/forum/go_quote.gif I've built a TPA6120 Class A amp general disscussion of Class A output bias of chips, photo's of my amp and circuit ideas futher down in this thread: http://www.head-fi.org/forums/f6/ad8...8/#post2261919

I haven't been here for years, and it's kind of awesome that the post you were replying to in that thread was mine

Ah, memories.

Rob.

 

[m1abrams]

I only looked at the spec for the TDA2030 so not sure about the other chips you listed but I suspect they will be the same.

The TDA2030 has a min. gain requirement of 24db. That is pretty freakin high even for the harder to drive headphones. The noise levels for that chip are not great either and considering you have to have a gain of 24db and mix in the high efficiency of most headphones and I think it would make for a rather unpleasant experience.

 

[jcx]

there's not much you can do for the slow, decompensated audio power chip amps


I recommend output buffer/gain stages with much higher BW than the the "audio" input op amp that servos the loop to simplify stability


put the fast chips in a Jung multiloop and you can bias their outputs Class A - separating the thermal modulation from the input op amp

look at DSL driver op amps or one of the (few) power packaged fast op amps like LT1206/10

 

[Nemo de Monet]

So, pessimism and negative attitudes about the unsuitability of big, slow audio amp chips aside... can someone point out the undoubtedly really fundamental error in this schematic:

I've prototyped this with the LM1875, and... nada. It's the basic CMoy schematic, only with a larger chipamp (I haven't added the bias transistor yet), and it seems like it should work, but... no. No heat, no smoke... no oompa-loompas.

What am I missing here? I've even changed R4 to increase the gain to 21, but still nothing.

 

[jcx]

how about a negative supply?

 

[FallenAngel]

Uhm... that's a gainclone chip, look at datasheet for implementation instructions.

 

[AudioCats]

LM3886's can do ok, pretty black background and details are not any worse than a buffered Cmoy.
Not a bad way to drive current hungry phone I guess (such as the HE-5), if your power supply is good enough.

 

[nullstring]

I have no idea what that transistor is suppose to be doing where it's at..

3 is just suppose to go to ground..

(I'm a novice, so I just figured I'd poke around at it for experience)

 

[Nemo de Monet]

Quote:

Originally Posted by FallenAngel /img/forum/go_quote.gif Uhm... that's a gainclone chip, look at datasheet for implementation instructions.

Yeah, ya think? (Page two: Typical Single Supply Operation.)

Thanks, Captain Obvious.

Nullstring: If the amplifier actually produced music, which it doesn't, for reasons I haven't quite worked out yet, the 2N5486 (or actually a J310, same pinout) would bias the chipamp towards Class-A operation when driving (comparatively) itty-bitty loads like... oooh, I know, headphones. The sort of audiophiles who like to buy very expensive solid-silver or solid-gold wire, then attach it to nickel plugs with tin solder, will be along any moment to inform you in no uncertain terms that ZOMG!!!11 that's a terrible, horrible, no-good, very bad way to produce a constant-current source, but... meh. It is, after all, "just" a "slow", "decompensated" chip from the primitive dark ages of 2002, when Texas Instruments, of course, hadn't the foggiest idea how to fab an audio semiconductor.