[Re: Power source]

[Harry Bissell]

You can use two matched resistors and an op-amp voltage follower (such as a
741 because it's well behaved) and decoupling caps at the output to both
"rails". But it will still suck !!! Essentially, you have a single supply with
a "pseudo-ground". One of my favorite sayings to engineers, techs, etc (anyone
who will listen...) is "If you pour a glass of water into the ocean, what
happens to the water level." Most people say "nothing" but the right anwser is
it goes up just a tiny bit. We just hope not to notice. Moral of the story
is... Use a supply with a real GROUND, big wires to keep resistance low, and
big output filter caps. Run all ground leads to the junction of those big
caps... Try to make your ground as unmovable as the ocean.  Also, The
half-wave wall wart solution is not as good as the Full Wave Center Tap. The
FWCT has a ripple frequency of 120Hz (not 60Hz) which means the filter caps
are replenished with current more often, and the same size cap will give more
filtering. You can get info from the Signal Transformer catalog, or the
National Semi Audio Handbook. A center tap transformer is required, but
surplus they can be quite affordable. Many a good design has been f@@@ed up by
an inadequate supply. WHILE I'm on a rant... dont forget to bypass the big
electrolytic caps with maybe a Tantalum, and a Ceramic disk  (like 1000uF
electrolytic, 2.2uF Tant, and a .1 or .01 ceramic. The Electrolytics filter
really low frequencies, when thier inductance and internal resistance gets too
high, the Tantalums will take over, and then the Ceramics at even higher
frequencies. Bypassing at the active devices is also a very good idea. I don't
see on-board decoupling caps located at the IC supply pins in many board
designs on the web. They really need to be there.... 
If anyone wants to argue this one, flame me at harrybissell at netscape.net
:-) Harry Bissell

Martin Czech <martin.czech at intermetall.de> wrote:

> > I agree, I really need a good power source and making one would be good
> experience.  I realized that this power source (from other members on the
> list and the display on my "new" scope, an RCA last calibrated on 1974) is
> simply fulwave rectified and no smoothing -probably because its just used
> to push little cars across a track. The cap I used worked well but now I
> have a problem.  I used two identical resistors to divide voltage and get
> a ground out, Problem is once I connect a circuit the voltages go crazy
> and ground moves to about two volts from the positive rail. 
> 
>  How can I buffer the power source so that other resistances in the
> circuit dont change the scale of the power source?
> 

I think the basic scheme to get bipolar power from a single coil
wall wart is to use half way rectifying.
Use one trafo output as gnd. The other will feed two caps (to gnd) 
via two diodes, one for the positive and one for the negative
cycle. This way you get a stable gnd, but now the power rails
are only half way rectified. Should be ok for experiments.

Of course, you have to get the voltages right at the trafo,
not behind the full wave bridge.

If you downlad all lepers effect schematics (in one file)
there are a couple of such circuits, to feed those effects
that need bipolar power.

m.c.


____________________________________________________________________
More than just email--Get your FREE Netscape WebMail account today at http://home.netscape.com/netcenter/mail