Inside the ubiquitous 741 op amp, the circuits in silicon

Cheap op amps made everything in electronics work better. In the tube era, an op-amp was a huge device. It's also hard to make a DC amplifier with tubes and have the output voltage be near the input voltage, hard to get enough gain, and hard to keep the zero point from drifting. Tube op amps were used only for analog computers, including ones for gun aiming.[1]

Negative feedback was rare in the tube era, because gain was so expensive. An op amp with negative feedback throws away gain to get linearity. In the tube era, that was wasteful. In the IC era, having a few stages of amplification to get some insane but nonlinear gain, then throwing it away to get linearity, became a win.

Once op amps became cheap, they could be used routinely whenever you needed some gain. So audio gear filled up with them. A negative feedback amplifier behaves like an ideal linear amplifier up to a frequency limit determined by the slew rate of the device. But for most op amps, that limit is up in the megahertz range, so there's no problem using them for audio. (There is an audiophile cult which denies this, of course.)

Op amps with differential inputs can also be used as comparators; the output goes to one extreme or the other depending on which signal is stronger. Comparator ICs are op amps with a logic gate stage at the end. That's how differential signals such as twisted-pair Ethernet get turned into logic levels for processing. Any interference which hits both wires is subtracted out. This is why we can pipe gigabit Ethernet around over rather ordinary wire, rather than coax.

[1] http://www.analog.com/library/analogDialogue/archives/39-05/...

Amusing having seen, decades ago, detailed articles describing how op-amps were engineered ... and now practically the same article showing op-amps being reverse-engineered. The excitement of uncovering the past is just as palpable as the excitement was of creating the future.

Ah the 741, for a long time it was ubiquitous everywhere. I particularly like the Evil Mad Science "discrete" version (http://shop.evilmadscientist.com/productsmenu/tinykitlist/76...) which still works pretty much exactly like the real one except its easy to stick probes on it if you want to see the internal signal paths.

I still have half a tube left of the quad package one (four 741 equivalent op-amps in a 14 pin package. Really useful for building analog computer circuits.

What a beautiful circuit. Recalling my first encounter with op amps as a struggling undergrad, faced with seemingly incomprehensible analog electronics, it was not pretty. I graduated but must note the significance and beauty of these circuits was lost on me.

Not surprising university was my last serious encounter with analog electronics. I do slightly regret not using the opportunity to explore and appreciate the elegance of it all. Working with software now, elegance and beauty are seldom the first thoughts that come to mind while debugging code.

Maybe Shaw was right, youth is wasted on the young :)

>Doing away with the external capacitor made the 741 extremely popular, either because engineers are lazy[14] or because the reduced part count was beneficial.

Small comment on this, doing away with the external capacitor meant 99% of the time the things just worked. Where the previous designs were a royal pain to tweak so they were stable, not only on the bench, but in production.

I love the letter from marketing at the bottom; the excitement is palpable.

I recall being told a rule-of-thumb to use when figuring out what a circuit using an op-amp does: the voltage difference between the two inputs is always zero, at least to a first approximation.