[sdiy] op amps [was: Re: TI SN74LS624N, worth buying?]

[Magnus Danielson]

On 11/13/2011 09:23 PM, David G Dixon wrote:
>> True, but some circuits are even more sinister, they can
>> depend on the internal design of the 741 such that it will
>> just not work. Jürgen had that experience for instance.
>
> I personally tend to avoid those sorts of circuits like the plague.  If I
> can't understand a circuit based on the so-called "golden rules" of opamps,
> then I don't bother with it.

One has to respect that some of the more elaborate designs around the 
741s is to overcome the limits of the original chip. Modernisation might 
involving making away with the additional circuit. Look for instance at 
the oscillators of the Oberheim SEM. The JFETs and transistors there is 
to overcome the lack if high impedance inputs and lack of good slew-rate 
comparators. Today we do TL-072 and LM-311 and barely nothing extra. See 
the ASM-1 VCO for comparision.

> Concerning oscillations and their avoidance, this seems to me to be a bit of
> a black art, and something I've had to learn by bitter experience.  It is
> tempting to throw little caps into every negative feedback loop to avoid
> oscillations, but this can also do more harm than good (particularly in
> filter cores).  For me, the following general rule has evolved: provide for
> stability caps on layouts whenever possible, but leave them off unless they
> are truly required.

There are many things to consider, such as phase-margin, dominant poles, 
de-coupling, slew-rates and PCB layout. There is no easy answer that 
will always work. The higher speed, the more things will be less ideal 
so it can be a bit challenging to master unless you learn.

Cheers,
Magnus