"ScottW" wrote in message
...
Now I'm confused again....isn't all that power supply stuff
technically decoupling?
The distinction between decoupling and bypassing seems to have escaped you,
Scott. You can decouple without bypassing, but it is harder to bypass
without doing at least a little high frequency decoupling.
The purpose of decoupling is to isolate the load from variations in the
source (example: the power supply). The purpose of bypassing is to provide a
source that provides a low source impedance. The process of providing a low
source impedance inherently provides some decoupling.
Many op amps are not stable unless their power supplies provide a low source
impedance, especially at high frequencies. An easy means for providing a
power source with low source impedance is to add bypass capacitors near the
power supply pins of the op amp. Therefore we see a lot of small value caps
clustered around the power supply pins of op amps.
A good op amp will tend to reject variations in its power supply,
particularly at low frequencies where the op amps open-loop gain is
extremely high. Therefore, there are only a limited number of sitautions
where op amp power supplies need to be heavily decoupled.
There are many situations where op amp power supply pins need to be bypassed
at high frequencies. An op amp with high bandwidth is much more likely to
need careful bypassing. In some cases the power supply pins need not be
bypassed to ground at the op amp, but simply effectively connected together
at high frequencies. Therefore we occasionally see cases where just one
relatively small bypass cap is connected between the two power supply pins
of the op amp.
Here's an example of an independent source supporting much of what I've just
written:
http://electronicdesign.com/Articles...leID=1478&pg=2
"The usual arrangement for a pc board is to have a relatively large bypass
capacitor for each supply line at a point close to where dc power enters the
board. Practical values between 1 and 100 µF may be used. As a rule, larger
capacitors are needed in circuits that must deliver large transient currents
into the load. In these applications, the power-supply bypass capacitor
stores the necessary transient energy, which otherwise wouldn't be available
directly from the supply itself due to lead inductance."
"If the pc board holds several op amps, relatively small (0.1-µF)
local-bypass capacitors can be used at each op amp's power-supply pins. A
much larger capacitor, located where the supply line enters the pc board,
can serve as a common bypass for each supply. The effectiveness of a common
bypass depends on several factors, including the length and width of the
wire runs between this bypass and each op amp, the value of the bypass, and
most importantly, the transient current required to drive the load. In many
cases, the power supply will have a very low output impedance, and the pc
wire runs will be very short. Therefore, the power supply itself will be
able to provide the transient energy needed to drive the load. If so, only a
small local bypass is needed at each op amp—just enough to supply
short-duration energy to the op amp."
Bypassing is shunting to ground at the source...not at the
power supply.
I quote my source:
""If the pc board holds several op amps, relatively small (0.1-µF)
local-bypass capacitors can be used at each op amp's power-supply pins."
That's not shunting to ground at the source, it is shunting to ground at the
load.
Here's a reputable app note on high speed digital
circuits which have far greater need .....
http://hsi.web.cern.ch/HSI/s-link/de...c/decouple.pdf
Wrong again Scott. You need to read and understand the titles of articles.
The article is titled:
"Using Decoupling Capacitors"
It is a fine article as far as it goes.
You need to find an article about "Using bypass capacitors".
**Nope. I checked. ALL use electros for decoupling. Every single one.
Some have film caps as well.
Well...I picked one op amp (BTW..I didn't see any obviously in the
audio circuits..only video)..and it was decoupled with a ceramic.
No, it is bypassed with a ceramic. Decoupling and bypassing are two
different things.
The most common symptom of poor bypassing is high frequency oscillation and
ringing. The most common symptom of poor decoupling is low frequency
osciallation and ringing.
So would you mind specifying a schematic and an op-amp.
http://www.hifiengine.com/download_c...-1_service.pdf
Please see page 35, figure 7-13, around IC 171 and 271.
http://www.hifiengine.com/download_c...85_service.pdf
Please see page 62, "Input/Volume Unit", IC 781-784. There are a total of
8 op amps with individual power supply bypassing, but shared decoupling.
In addtion IC 816 on the same page shows an example of an op amp with power
supply bypassing by means of a single relatively small cap connected between
the power supply pins.
PITA to scroll back and forth between schematic and parts list.
It takes a certain lack of intellectual laziness to understand these things.