[Yaaarc] DC-DC power filters

[Janet Plato]

Hi Guys,

Does anyone have experience with
power supply filter design?


I need some help with a quick and dirty 
power supply filter problem.  I have a 
small black box with RF in (BNC connector), 
audio out (via line-out to PC soundcard 
line-in) and a serial connection to control 
it.  

It works adequately and I am ready to permanently 
mount it.  It draws about 1A and can tolerate 
0-16V, and run on 11-16V.  I strongly suspect
it would take 18V, but I'd rather not find out
I am wrong.  With a little help from my friend
Dremel I've mounted it into an old CD ROM
drive, and intend to power it from the 12V 
rail of an ATX power supply.

Herein lies the rub.  I don't really want 
to buy a filtered power supply, and I don't
have an oscilloscope to verify my design.  So
I'd strongly prefer I get it right the first
time, before I haul everything in to ask a
friend to examine my waveforms.

My handy 2003 ARRL handbook offers the stock
designs but is really focused on removing
ripple from a rectified AC wave to tolerable
levels for a transceiver.  They don't have
any engineering advice for DC-DC filters.

I have a variety of chokes and caps, and could 
easily build something and measure it's 
performance, but like I said, no oscilliscope
means being wrong is a bit tedious.

1) do you recommend capacitive or inductive
inputs?  Below is a stock capacitive input
design, note the filter inductors are after
the first capacitive bypass.  I am slaughtering
the vernacular but hopefully you get my meaning.

 + ---+---UUUUU---+---UUUUU--- +12V
      |           |            
      = ? uf      = ? uf
      |           |
 - ---+----+------+---------- -
           |
         ----
         \ \ \

Here is an inductive input, note the cacaptive
bypass after the inductors.  Other the design
is largely identical.

 + ---UUUUU---+---UUUUU---+---UUUUU--- +12V
              |           |            
              = ? uf      = ? uf
              |           |
 - ---UUUUU---+----+------+---------- -
                   |
                 ----
                 \ \ \

Any idea of what values to use?  On a theoretical
level it's obvious the bypass caps have to provide
the voltage/current to cover the parts of the
cycle when the voltage sags.  But the 12V rail
is probably running at 10's of KHZ, so it 
seems like a capacitive input would present a
dead short.

2) any preference for number of stages?  Space
is tight ideally I could meet my goals with something
small, in a single stage.

I know, I really need a scope, but I use it
all of once a year, and I've had some rather 
brutal personal expenses the last two years.

Cheers,

Janet

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com