Op Amp 101

[Tom Cloud]

>This M.E. is doing a miserable job of learning about Op Amps.  I
>understand the fundamentals now of what one is and what it does, I'm not
>completely clear what is the goal.

Joe, an op-amp is simply a device that amplifies some sort of signal.
Obviously, there are gotchas and the possibilities of how you can
use it are endless.

 - most op-amps have two inputs -- so they are "differential".  That
means they "look" at each input, compare them (or subtract them) and
operate on the difference between them.

 - the inputs are generally "high impedance".  That just means that
they don't draw much current from the signal they're looking at --
that is to say, they don't change the signal they're looking at.

 - the output is "low impedance".  That means it can supply enough
current to drive other circuits.

 - most op-amp circuits are simple multipliers -- some invert the
input signal (positive transition becomes negative -- sort of like
a "see-saw" where one end is "in" and the other "out"), some don't.
Some sum multiple inputs (summing amps) and then multiply the summed
value.

 - other op-amp circuits do fancy stuff like perform logarithmic
operations to compress a large dynamic range input into a smaller
region (you know what log graphs are for).

For simple stuff, the basic gotchas (from off the top of my head) are:

 - getting correct supply voltages to it

 - being sure the input impedance is set high enough to not load the
signal

 - being careful of the "common-mode input voltage".  This just means
that the op-amp can only measure inputs that are "inside" the extremities
of it's power supplies (called the "rails").  IOW, if you're using
+/- 15 volts, a 741, for example, could measure voltage of ~ +/- 13.5
volts.  It's common-mode input joltage range is typically within about
1.5 jolts of the rails.

 - being careful about it's frequency limitations.  A 741's gain drops
to 1 by the time the input gets to 1 MHz -- and it's output won't be
very large at anything over 10 kHZ or so. .... and it'll be even less
if you're "driving" a load (i.e. supplying very much current).  I prefer
the LM/LF-356.  It has an fT of 5 MHz.  The LF-357 is 20 MHz, but you
can't use it at closed loop gains of less than 5 (see, it gets sorta
deep real quick  8^)

 - being careful to not expect too much out of the output (see above).
For slow signals (say below 1 kHz or so), the op-amp will pretty well
work like the spec sheet.  But, if you try to draw current from the
output and/or expect a wide joltage swing out of it at much higher
frequencies (depends on the op-amp), the output voltage will drop
(be incorrect).

The advantage of the LM-324 is that it can be used with a "single
supply" (a "bipolar" supply means one that has a positive and a negative
output in addition to common or ground).  Most op-amps "want" a bipolar
supply.  The LM-324 has a common mode input range that includes zero (the
negative rail (well, danged near) and it's output will pull to zero (the
negative rail) -- if assisted with a "pull-down" resistor and you don't
expect it to "sink" too much current.

BTW, remember that ALL op-amps are "single supply" op-amps.  Note
that none of them have a ground pin so, for all they know, your
+/- 15 volt supply is actually a 30 volt supply.  The center "common"
is for your reference only (pun intended).  They just have that "large"
common mode range that excludes both rails, so you can't use them to
measure things that operate close to zero, like an EGO, unless you
give them a negative supply to move the negative rail away from zero.

  ........  snip  .....

>The question of the post is: 
>Is the purpose of the op amp to filter the wild fluctuations at idle in
>the reference signal because of the cam overlap, or is the purpose to
>bend the vacuum signal of the wild camshaft to mimic the family sedan
>vacuum curve, or both?

the "purpose" is to do whatever the circuit designer told it to do.
It can be made to do each, both or none of what you described above.
Smoothing the fluctuations is filtering or integrating.  To "blend"
the signal would probably require an intricate circuit to produce
that function and would be more easily done with a micro-presser.

really cheap and quick way to learn the basics is to get the Forrest
Mims books sold at RatShack for about $10.  They have (had ??) one
specifically on op-amps with a kit of parts for that price.  You can
buy a pocket, autoranging DMM (digital multimeter) from them for
about $25 (note that I hate their instrumentation and recommend
spending a little more for something worthwhile, but $25 isn't
a big deal and, when you've outgrown it, you can curse some cub
scout down the street with it  ;-)

To play with the circuits in the Mims book, you'd probably either
want an oscilloscope (not really necessary) or an analog meter so
you can see the needle move and get a feel for what's going on.  The
digital meters are too "sterile", IMO.

I'm outta breath (carpal tunnel setting in  8^)

Tom Cloud

     Madness takes its toll   .....   please have exact change