How to build your own opamps from discretes.

[ModernRonin]

I've been having a good time at home with this page, which shows you how to build your own simple opamp from 6 transistors and 2 pots:

http://www.allaboutcircuits.com/vol_6/chpt_5/17.html

Also been looking at Widlar's LM301 and Fullagar's 741 - http://en.wikipedia.org/wiki/Operational_amplifier#Internal_circuitry_of_741_type_op-amp

Learning about differential amplifiers (long-tailed pairs and variations), current mirrors, Vbe multipliers, and other staples of analog design with BJTs.  It's fun stuff, and maybe might make good fodder for an EEVBlog episode one day.

[Zero999]

Of course in real life there's no much point but it does have some educational value.

One thing you could consider is using a transistor array IC which will give better results as the transistors will be better matched.

[ciccio]

Discrete operational amplifiers are still used to-day in high-end audio apparatus.
In fact, there has been a recent "revival" of them, and I see many new offerings.
Many are pin-to-pin replacement of the old API 2250, others can replace standard 8-pin single and dual op-amp.
These are some links:
http://www.sg-acoustics.ch/analogue_audio/discrete_opamps/index.html
http://www.avedisaudio.com/1122.html
http://www.tonelux.com/opampinfo.html
Being capable of higher voltage operation, they offer an increased dynamic range, and sometimes  a better noise performance, when compared to standard audio op-amps.
Their circuitry is, most of the times, more complex than the simple schematic presented in the original post.

[Bored@Work]

Notice that they don't specify temperature drift for these discrete opamps.

The big breakthrough in building opamps was when the IC was invented so the temperature drift became controllable, because everything, esp. the differential amplfier, was finally at almost the same temperature.

And notice that they add some typical audiofool trash talk, like "It all sounds very warm". When will the audiofools get it that it isn't about sounding "warm", but about sounding right?

Or "Using no IC's, no JFET's, and no surface-mount components, the 1122 is an evolution of where our favorite opamps from the past left off." Dude, you seriously call it a evolution skipping decades of progress?

[Zero999]

Being capable of higher voltage operation, they offer an increased dynamic range, and sometimes  a better noise performance, when compared to standard audio op-amps.
Their circuitry is, most of the times, more complex than the simple schematic presented in the original post.

Better noise performance and increased dynamic range my arse.

The correct way to use an IC op-amp in a high voltage application is to add transistors aground it so the op-amp never sees the full power supply voltage.

Here are some links to some examples:
http://cds.linear.com/docs/Application%20Note/an18f.pdf
http://www.tenmilecreek.net/images/Op_Amp_Booster_Stages_pt1.pdf

[ciccio]

I was only reporting the fact that somebody still sells discrete op-amps.
I'm not endorsing them, and I do not usually speak or write audiophile jargon.
The fact is that some circuitry sounds "better" than others, especially when your source is a microphone with a known voice (maybe your voice).
We could open a debate about the meaning of "better", but I'm not sure this is the place.
30 years ago I built some discrete op-amps, starting from a design from Robert Burwen (http://www.burwenaudio.com) published on Audio magazine (USA).
I wanted  to test them in the mixing stages of a big desk I was designing.
The results were amazing for noise and distortion performance, but we discovered that it was (at least for my company) very difficult (and expensive)  to get a consistent quality production, so we did not continue in that direction, and used the new entry : Signetics  NE5534.

[shadewind]

There's definitely a place for audio electronics that doesn't sound "right" in recording studios. The simple reason is that reality sounds like shit so you don't want that.

That being said, "warm" is a rubbish term which seems to mean entirely different things to every person. To some people, it even seems to mean "it sounds nice".

[alm]

Isn't the task of the recording engineers to alter the sound so it sounds good, and shouldn't the playback equipment reproduce this as accurately as possible? Or do you mean that the recording engineers and artists are incompetent and audio equipment needs to compensate for this?

[shadewind]

Isn't the task of the recording engineers to alter the sound so it sounds good, and shouldn't the playback equipment reproduce this as accurately as possible? Or do you mean that the recording engineers and artists are incompetent and audio equipment needs to compensate for this?

The goal of playback equipment is indeed to reproduce things as accurately as possible though you sometimes need to listen to a mix in crappy speakers as well to get a feel for of how it will sound on crappy speakers.

Of course it is the recording engineer's job (or mostly the mixing engineer's but often they're the same person) to alter the sound so it sounds good and there are many different tools for doing that. The interactions between a microphone and a microphone preamplifier is admittedly not as complex as many people claim but it's still hard to reproduce that by manually applying distortion, filtering and dynamic compression since it's so subtle. It's not that easy to tell just by listening what is happening to the signal. It can often be simpler to just use something that you know sounds nice. You still have to make a judgment call as to what will work in the final mix etc.

There's also the matter of "learned sounds". The Shure SM57 is a classic dynamic microphone used a lot for recording electric guitar and drums. The sound of a snare drum that people associate with rock and pop records is rather far from reality, you could almost say that the SM57 is part of what people expect a rock snare drum to sound like.