High end equipment, and their amp circuits

Well i guess you have decide what you mean by high end.

But this one certainly counts.

http://www.audiovalve.de/rkv/engrkv.html

LF351 used as the front end amplifier.

Lots and lots of negative feedback too.

There is a well known amplifier that is just an opamp and 2 transistors.

Available in both balanced and unbalanced form.

The Crown IC150 used to be considered high end 25 years ago, it was

nothing but opamps.

Many amps and preamps use opamps as the servo's to maintain no dc at

the output without any capacitors in the signal loop. This is a competely

different story as some say they can hear the servos.

No one has ever been able to make a completely complementary monolithic

integrated circuit. The kind of sound you get from completely matched complementary

discrete circuits cannot be done on a single piece of silicon.

I believe the Ray Samuels Apache uses opamps for amplification

circuit capability is far weighted in favor of monolithic today - it is after all 2010 - not 1990

I like the feeling of control you get from designing discrete circuits but the available devices are falling behind - except for the very lowest noise performance, suitable for say Moving Coil phono pre or when you need Vswing beyond +/-15 V (bridge outputs 2x swing)

today's best op amps for line level applications like headphone amps can't be touched by discrete

many manufacturer's have dielectric or other advanced isolation analog processes today that allow as good a match as the different conductivities of N vs P doped Si make possible in discrete or monolithic - and inside monolithic op amps the transistors are much faster, better matched than you can manage in discrete packaging

the demand for A/DSL chips with low distortion driving 25-100 Ohm loads with complex 100KHz-1MHz multitone modem signal have driven a big investment in process and circuit topologies for ultra low distortion op amps - perfect for most dynamic headphone loads

composite multiloop circuits using best in class monolithic op amps low noise/distortion front end and DSL driver for output can give "unmeasurable" distortion performance

all op amp multiloop composites are not common although they clearly allow for better performance

a few multiloops with monolithic unity gain buffers and probably more with op amp front ends and discrete output buffers are made

some are using (or "rolling") recent OPA627 "replacements" or the LME "audio" op amps but most commercial circuits just use a single op amp per channel/circuit function

at the commercial audiophile high end touting your op amp choices would be an uphill battle given the entrenchment of the "discrete superiority" meme in audiophile mythology

simply better (by orders of magnitude, on meaningful complex multitone IMD tests) numbers are not respected - nor do "reviews" include controlled, blinded comparisons

Claiming discrete devices are falling behind is such utter BS that one can just stare a laugh histerically.  The truth is that we have never had better transistors and they are allowing us to do things that were never possible before.  Now for most of high end audio this doesn't matter as the majority is badly designed crap but the true greats are pulling out all the stops. 

It's quite clear that opamps will always be a compromise.  Plenty of feedback to keep them stable and a discrete cap/resistor will always be better.  That said, any design is a compromise (unless you belong to team overkill) and most are just happy with a shiny box that looks impressive. 

Quote:

Originally Posted by JB197 

 

I’ve read contradictory advice about ‘the sound of op-amps’, and it got me thinking about which top of the range equipment might contain op-amps (as opposed to tubes or circuits with individual transistors).

Something hinted at by other posters is the open loop gain.

You can make discrete Op Amps. You can also make discrete Amplifiers. There is a difference. Opamps by definition operate on the theory that they have (nearly) infinite open loop gain. This is nice because as you apply more and more global feedback the THD drops. What is not so nice is that in order to get more open loop gain you add more active stages. This causes more open loop THD... Once they get into the path distortions never go away, they just get shifted around.
 

On the topic of amplifiers, there are actually amplifier chips! what fun. Building a discrete amplifier is an excellent solution when you want something purpose built for an application and are willing to set specific goals for it. The generic answer of an opamp, while fitting may not be the best one - although it is generally the easiest.

On that note, if your goal as a designer is to build a discrete opamp, your a retard. you can buy something better off the shelf today. If you are trying to build a discrete amplifier you have something going for you.

The unfortunate point is that no (very few? Maybe... Im not giving them credit)  MFRs specify that their discrete circuit is an opamp even though it may be. They largely ride on the well earned distrust in the audio comunity of the junk in a dip-8 package (normal IC opamp) because historically opamps have SUCKEDDDDD!!!

PS: Why is this in the elitists forum? It would receive many more hits in the low end forum.

Quote:

Originally Posted by spritzer 

Claiming discrete devices are falling behind is such utter BS that one can just stare a laugh histerically.  The truth is that we have never had better transistors and they are allowing us to do things that were never possible before.  Now for most of high end audio this doesn't matter as the majority is badly designed crap but the true greats are pulling out all the stops. 

Discrete devices ARE falling behind for audio applications. The fact of the matter is that high voltage transistors were a biproduct of the CRT/TV industry for driving the display. Nobody is building CRT monitors/TV's anymore, so the need for high voltage transistors is dropping to stuff designed for switching and whatnot. Yea, that sucks. Low voltage stuff is largely moving from analog/linear circuits to digital. Again, the greatly reduced demand for transistors that work well for linear applications (you dont need low noise for a switching transistor..) is stopping production.

I can think of at least a hundred places Kevin Gilmore (sorry Kevin, your just so conspicuous) has complained about various semiconductor MFR's stopping production on semiconductors which are desirable for certain high voltage amplifiers for purely reactive loads... and a few amps he has in his mind that wont get out because of that.

Lets not mix apples and oranges here.

ONSEMI  has some absolutely wonderful audio transistors, that are, and will be in production for

a very long time. Matched pair thermaltrak transistors up to 350 volts. You can build some really

nice monster power amps (800 watts rms/channel) with these things.  All of the driver transistors

to drive them are also available. Lots of cascode front end transistors for these kind of applications

also available. This kind of stuff will never go away. Even if you go all digital like the ICEpower amps

you still need these transistors.

What is not available is matched pairs at 1000 volts or more. The best audio pnp transistor is

rated at 600 volts and in my latest amp, they are stacked. Because of CCFL and other uses

this transistor is not going away either. But the low Cob 1200 volt transistors and the 900 volt

pnp's are no longer in production because no one needs them anymore.

The biggest problem with opamps is the group delay associated with the feedback. You can

make the stuff test absolutely amazing and it will still sound like crap. And you can make

things like the Firstwatt that sound absolutely wonderful, but test awful.

To get the kind of sound that dynahi,B22,dynafet produce requires power supply rails way higher

than +/-15. And idle currents in the hundreds of milliamps.  There are very few monolithic

opamps that can do that kind of thing, and the apex hybrids that do are way out of line

price wise. And in general the apex things don't sound very good.

You do not see opamps in any of the high end gear except for servo's. Notably

Krell, Levinson, Ayre, Roland etc.

my comment was qualified as applying to headphone amps, new, useful op amps suitable for most dynamic headphone loads have been developed in the past 20 yrs while suitable discretes have remained static - with a few losses - so op amps have advanced in absolute terms, discretes have fallen behind relatively

"slow negative feedback" is another unfortunate audiophile meme - the TPA6120/THS6012 CFA op amp can have a corner frequency of 100 MHz at gains up to +10, power gain at 100 MHz is in excess of 60dB, 400 mA output current

in contrast the "fast" discrete small signal transistors shown in Kevin's original project article which became known as DynaLo/Justin's GilmoreLite have ft of ~80 MHz  

(anybody remember my avatar?) I've shown my 48 Vpp 400mA Class A output all op amp multiloop at two national meets and posted the Class A bias scheme and shown the heat sinking that allows >25 W from the 6 TPA op amps

Quote:

Originally Posted by nikongod 

PS: Why is this in the elitists forum? It would receive many more hits in the low end forum.

 

x2 .  double-dutching is de rigeur here .

Quote:

Originally Posted by jcx 

all op amp multiloop composites are not common although they clearly allow for better performance

 

a few multiloops with monolithic unity gain buffers and probably more with op amp front ends and discrete output buffers are made

 

some are using (or "rolling") recent OPA627 "replacements" or the LME "audio" op amps but most commercial circuits just use a single op amp per channel/circuit function

 

at the commercial audiophile high end touting your op amp choices would be an uphill battle given the entrenchment of the "discrete superiority" meme in audiophile mythology

 

simply better (by orders of magnitude, on meaningful complex multitone IMD tests) numbers are not respected - nor do "reviews" include controlled, blinded comparisons

If they told the truth how could they justify $$$$$ for their products any more?