Why these parts?
I've received several emails with questions on the power
supply circuit.
I don't know if I can answer them completly, but this should
serve to
start the discussion.
Please, read all your email before you reply to any, just to be
sure
that there are no duplicate answers.
All the questions centered on a theme. WHY did the designer
choose this
particular component?
Lets also keep in mind that many components are chosen for
reasons other
than function. There are many parts that could be replaced with a
number
of parts and function properly. Often the choice comes to
economy,
availability and stock. Dave may have 10,000 1N4001 diodes in his
stock.
That would move him to use these components in his designs
anywhere
possible. Physical size is also a consideration, will it fit on
the
board? Component choice can also be "designers
preference".
D13, why a 1N4001?
I think I can answer this one. What characteristics do you need
in this
particular application? It must be able to carry all of the power
supply
current through it (up to .5 amps). It must have a reverse
voltage
rating greater than the supply voltage (15 volts). It does not
have to
be fast, as it is not a detector or a switching diode.
Capacitance
really does not matter as it is normally forward biased. It
should be
inexpensive and readily available.
So, the 1N4001 fits these characteristics perfectly. It handles
up to 1
amp forward current, 100 volts reverse voltage, characterized as
a
rectifier. And they are really cheap.
You could replace this component with any rectifier diode that
has a
PIV of 30 volts or more(input voltage plus the charged Power
supply
capacitor) and a forward current rating of 1 amp (0.5 amps plus
surge
current) or more.
C112 and C102
Why these types and values? I think I can answer part of this
question.
C112 is a 220 microfarad electrolytic. C102 is a .01 microfarad
ceramic.
C112 is for filtering low frequencies and C102 is for filtering
high
frequencies (RF). The electrolytic has a low impedance to lower
frequencies (120Hz) but has a substantial impedance to high
frequencies,
due to it's construction. This is why C102 is there.
There are lot's of different types of capacitors with widely
varying
characteristics. Does anyone have a summary of different
capacitor types
and their characteristics? That would be a good post.
U2 78L08
Why? I believe this one was already answered in the original
post. It
keeps a steady 8 volts on sensitive circuits even if the supply
drops to
10.4 volts.
Why this particular part? It is a standard three terminal
regulator.
They have been around a while. Take a look at the date on the top
of the
data sheet (January, 1976). From what I understand there are two
types,
the L series in the TO92 (plastic transistor) case, and the
standard
series is in a TO-220 (like the final output transistor, Q6). The
standard series can supply lots more current. It is also more
expensive,
larger and has a higher bias current. Since the circuit
requirements are
for <100mA , the 78L08 is the correct choice.
You could "roll your own" regulator with discrete
components. However,
at $0.25 each, and such a small size, I don't think you can beat
the
78Lxx series.
I hope this is a start in the right direction to answer the
questions
posed. Let's hear from those "in the know" on the
details I missed.
More Whys.
Here are the answers to the capacitor selection questions. I
got this
info from Dave Benson, the designer.
>> Why do we use a 220 mF instead of a 100mF to filter the
low freq junk?
>The 220 microfarad cap was the largest value available in
that size package.
>> Why do we use a .01 mF instead of a .1mF to filter the
RF?
>Yes, they'd all work, the .01 is cheapest.- I buy caps in
lots of a thousand-
>the difference between $.02 and .05 fills up my car's gas
tank a couple times. >;-)
>
>I do remember someone asking why the bypass caps on U1 and U2
were different
>values, and I saw some good explanations. The real story is
that they would
>have both been the cheaper .01s but one of the locations was
physically too
>tight and I upgraded to the smaller-package .1 cap!
I hope this clears this all up. Thanks Dave for the feedback.
So, it should be clear that often a component choice is not made
only on
it's function. Economy and size are also a large factor.
Power supply rejection ratio error
>Also, variations on the input do appear at the output,
although greatly
>attenuated. The spec is 48dB at 120Hz (full wave rectified
line ripple).
>If my math is correct, a one volt change in input voltage
will cause a
>15.8 microvolt change in the output.
Sorry .... the power supply rejection ratio is in units of power,
so the
power is indeed reduced by 48 db. But, this corresponds to a 4 mV
change
in voltage. This is one of those famous 'db voltage' vs. 'db
power'
problems.
10^(-48/20) = 0.004
Lesson is very good! At least, for an engineer :-)
Schematic error - C102
I recently posted that C102 was at U2's INPUT side. WRONG!
I had looked at the schematic to determine where it was, not
on the board itself....
On the schematic, C102(0.01uf) is shown at U2's(78L08)
INPUT. But how is C102 connected on the board? Its at U2's
OUTPUT side.
So if you measure the DC voltage across C102, you'll
see +8v. Not +12v.
Everyone should scratch out C102 on their schematics, and
draw in a new C102 at the top of the tuning pot (J2, pin 1).
-Glen VE3DNL