🚀 go-pugleaf

An integrated operational amplifier that would probaly perform
quite well for audio (hifi?) applications might be the Analog
devices AD318K.  This chip can be characterised as a wideband, low
drift, FET input opamp.

Here is a synopsis of its characteristics:

open loop gain:		100,000

output:			+/-12v (+/-10 min) at Rl=1K
			20 mA short circuit output current limit

dynamic:
	unity gian, ss	5 MHz
	full power bw	500 KHz
	slew rate	30v/uS
	.1% settle	700 nS
	.01%		1.2 uS

input bias current:	10pA avg, 50 pA max

imput impedances:
	differential	1E12 ohms in parallel with 7 pF
	common mode	1E12 ohms in parallel with 7 pF

input volrate range:
	differential	+/- 20 v  (maximum safe limit)
	common mode     +/- 12 v (+/-10 min)
	CMR		80 dB minimum

noise figures:
	0.1 - 10KHz	2 uV peak-peak
	10 Hz		70 nV/Hz^.5
	100 Hz		45 nV/Hz^.5
	1 KHz		30 nV/Hz^.5
	10 KHz		25 nV/Hz^.5

case style:		TO-99

This device is at least somewhat exemplary of the sort of device
that was alluded to in the preceding article.  Most notably, the
device exhibits a wideband frequency response and high open loop
gain concurrently.  At the same time, excess noise is held to
acceptably low levels.  Analog devices does not publish a
scheamatic representation of the AD381K's circuit topology, so I
can not comment on that.  Personally, I don't care about the
topology inside the device as long as the terminal characteristics
are what I want.  I am not sure when the AD381K was conceived, that
really doesn't matter either as long as it performs correctly.  I
would feel that the above amplifier would be acceptable for use in
such things as CD palyer active filters, dbx de/encoders, etc,;
certainly in mundane applications such as tonal shaping control
circuits.

I am not particularly impressed with the Signetics 5534 relative to
anything equivalent from other fabricators; who cares who makes the
part, given reasonable quality control is imposed.

For a dual amplifier chip, The LF442 from National has many
characteristics similar to the AD381K.  National says that
interamplifier coupling is no greater than -120 dB.  That is
certianly low enough for anything I do.  National does not publish
the full power bandwidth, but the unity gain point is 1 MHz.  I
really don't see any reason why multiamplifier chips should not be
used.  My test equipment can not measure to -120 dB limit; I assume
that National's claims are correct.  I doubt that I could decouple
other parts of my designs much better than that anyway and still
have them be easily assembled.

Probably the desire to keep costs low and maintain easy board
loading keep parts such as the AD381K out of mass produced consumer
grade audio devices.  (I'm gessing that it is more difficult for
automated equipment to insert a TO-99 case device than a DIP or SMD).
For snob-appeal audiophile equipment, there should be no reason not
to use high quality parts, since the cost of the parts is a
relatively minor portion of the cost of manufacturing the
equipment.  However ... it frequently would seem that logic and
common sense are seemingly minor factors in design, manuafacturing and
marketing (even purchasing!) of snob-appeal equipment.

As far as fully integrated subsystems such as dolby or dbx
en/decoders go... I don't design chips for a living, so I feel ill
prepared ot comment on the praticality and merit of such things.
However, I do suspect that some such devices are not optimally
designed, which is not to say that is impossible or difficult.  It
seems likely that standards are often no better than is necessary
to satisfy the public's desire.  Most likely the manufacturer of
equipment is to blame by surrounding a decent integrated circuit
with suboptimal discrete components.

No one fabricator of integrated circuits is superior above others.
The speific devices that have been mentioned above are only chosen
as exemplary of similar devices that are available from any number
of suppliers.