Discrete Op-amp Question

[Agung Kurniawan]

Did anyone know the advantage and disanvantage of using discrete opamp?
Since some of Boss drive pedal use this one inside it, im interesting to make some. And why did boss feed them with different power voltage?
Thanks...

[EBK]

This page seems to cover it well:
http://nwavguy.blogspot.com/2011/08/op-amps-myths-facts.html?m=1

[Agung Kurniawan]

This page seems to cover it well:
http://nwavguy.blogspot.com/2011/08/op-amps-myths-facts.html?m=1

Thanks..

[slacker]

In my experience the simple dicrete opamps you see in pedals distort in a nicer way than real opamps, or have some other tonal quality that the circuit is exploiting, that's probably why Boss uses them. They're basically technically not very good opamps but that makes them do something interesting.

[R.G.]

This page seems to cover it well:
http://nwavguy.blogspot.com/2011/08/op-amps-myths-facts.html?m=1

I love it. Somebody finally gathered it all up in a readable form. Good on him.

And you!

[Agung Kurniawan]

In my experience the simple dicrete opamps you see in pedals distort in a nicer way than real opamps, or have some other tonal quality that the circuit is exploiting, that's probably why Boss uses them.

Agree with you, i like the Boss mega distortion gain boost section. It sound diffrent, warmer and crack nicelly for me

[PRR]

> advantage and disanvantage of using discrete opamp? ...Boss drive pedal use this ... And why did boss feed them with different power voltage?

My opinion--

Boss did a lot of odd things.

I think some VIP's nephew needed a job. He was put to designing circuit details, switching and amplifiers.

At the time the trend was VERY much to using chips wherever possible. They cost a little more, but assembling a dozen discrete parts is a real cost; also the management to make sure all those parts are kept in stock, no shortage or surplus.

But while chip-use was rising everywhere, discrete parts prices were falling. 50-cent transistor for 9 cents, etc. The junior designer saw that clever use of low-low-cost parts already in Boss's system could save a whole penny (over a Yen!) in that sub-circuit.

Saving a penny a pedal, times Boss's sales volume, is BIG money and some respect (less-small desk) for the designer.

There may be some "audible" differences for extreme signals. And extreme signals is Boss's real product. But I suspect the economics played a large part.

Also as the years went by, chips fell out of large product as "better" chips came along. The new chips had to be tested before replacing older chips in a happy product. OTOH a subcircuit built of low-cost generic transistor will probably work with almost "any" similar transistor no-difference. And transistor types typically do not go "out of style" as fast as chips. (Power MOSFETs an exception, but that's not Boss's business.)

[teemuk]

I think the most common reason to use discrete alternative is simply to acquire lower noise figure.

If you look at it, the very first stage of many higher quality studio preamps (especially mic pres) is often a discrete opamp, or at least a discrete differential stage coupled to generic integrated opamp.

Lower noise figure would also make sense in "high gain" applications like those Boss effects that clipping distort the signal because when gain is high all noise produced by the circuit is naturally amplified with that gain ratio. It's not nice when that distortion pedal hisses disturbingly.

Discrete circuitry is often more expensive than integrated but it is common that lower noise opamps are more expensive than more generic opamps with higher noise figures so... Additionally, if low noise is a design goal then I guess a little bit of extra expense to that is noted in the budget to begin with.

In some cases a discrete circuit may feature better overdrive / overdrive recovery characteristics. The dreaded "rail latching" of certain integrated opamps is a great example of what sometimes needs to be avoided. A simple discrete circuit may have poor noise figure, more overall THD, etc. but simply behaves better when overdriven compared to opamp chips that implement all kinds of constant current sources, current mirrors and current limiter circuits.

Of course, that's just a rough generalisation because in practice opamps of all kind have great variety. For example, "rail-to-rail" -type devices tend to behave better under overdrive conditions than more generic chips. I guess what I try to say is that if overdrive characteristics are considered even the generic integrated chips offer so much variety that going discrete is not neccessary.

Also, there are distortion effects that specifically employ clipping characteristics of an overdriven opamp instead of clipping characteristics of shunt diodes or diodes in the feedback loop of gain stage. Generally one achieves "harder" clipping with more aggressive tone in such way.

Someone also mentioned rail voltage limits. They actually apply to both directions. Yes, high rail opamps are expensive and their variety much more limited. A discrete solution might prove to be a better option. A discrete solution might also prove to be a better option when the rail voltage is very low and yet as much voltage swing as possible is required. Which of course brings us to those "rail-to-rail" devices to remind that mostly anything one can build discrete can also be integrated to chip.

Anyway, I've seen a few effects where clipping is achieved by overdriving the gain stage (no diodes etc. involved) and the circuit is deliberately powered from lower supply voltage to decrease headroom.

[bool]

I assume that a simple discrete opamp may exhibit less rail-sticking and quicker recovery when overdriven compared to a (much) more complex chip-opamp ...