[sdiy] more power questions

[harry]

Hi MM  (inline)

Mountain Man wrote:

> Like Rob, I've been working on circuit power this weekend.  I've put
> together a 7812/7912 circuit, intended to daisy-chain off my 9700 supply
> (and killed the 9700 power supply in the process :(    Anyway, I have the
> supply working, but studying it's behavior in light of the discussion a
> couple of days ago has raised a bunch of questions for me.
>
> After the discussion about needing more than 12V, I thought, "hmm, I wonder
> if the 12V AC wall wart that I'm powering off of is supplying enough?,
> Hooked it up to my voltmeter - 14.7V.  That's not 12V.  That's not even
> within 20% of 12V.

The smaller the transformer, the worse the load regulation. They make the core
too
small, and expect to run the unit "hot" to get the current (near saturation of
the core)
at light loads... the windings are TOO hot... so voltage rise of 20-30% is not
uncommon at all. Your AC line could be a little high as well...

> Not that I'm complaining, I'm getting extra power for
> "free" <g> ... but I am curious.  What's the story on "power ratings"?
>
> I have the "12V" AC feeding a pair of power diodes (1N4003).  When I
> measure the (DC) voltage at the other end of the diodes (using ground
> coming from between the caps in the 7812/7912 circuit) I get +18.2 and
> -19.2.  What causes the voltage increase?

AC voltages are sine waves... so a 12V "RMS" AC voltage is actuallt 1.4 times
that
peak... or around that 18V range.  Loading on the positive and negative
supplies is probably different, and that accounts for the voltage difference.
Especially if there is ripple on the caps.

> Is the difference between the
> two simply due to mismatched diodes (not a problem, but I'm curious  :)
>
> I hooked my scope up to the output of each diode in turn (using my grabber
> hook attachment - thanks, folks!).  I expected to see a series of "M"s (or
> "W"s).  What I'm seeing is a flat line, with upward "bumps" every 18 msec
> (must be 60Hz!  I think I'm getting the hang of this scope :).

16.6mS on 60Hz line... set your clock by it (hehheh).

> The bumps
> are about 2msec wide, and 1.5V high.

Ripple... The diode only conducts when the voltage at the anode is greater than
the cathode... happens at the peak of that AC cycle. You are seeing the cap
voltage
"decay" between charge (current) pulses. More on the positive side... because
the load is draining it faster.  Make the cap bigger, and the decay will be
less... but
too big can blow your diodes or transformer by drawing too much current at turn
on!

> When I check negative, I see downward
> bumps a little less wide, and .5V.  Why is so much of the cycle at 0V?  Why
> are the + and - peak voltages a factor of 3 different from each other?  And
> why is the peak voltage I'm seeing so low, particularly given the DC
> voltage readings?

You ARE using the scope with DC coupling arent you... if not you are high pass
filtering the DC and all you see is the AC ripple.  That would account for the
being at
zero volts most of the time.  Use DC coupling to see overall voltages, and AC
coupling to "zoom in" on the ripple.

>
>
> BTW, I *am* still learning to use my scope (and I suspect will be for a
> while :)   I'm using the "holdoff norm" control to get the waveforms to
> "hold steady."  I'm unable to prevent a small amount of horizontal drift,
> however, which makes measuring time a pain.  Is there a trick to this?

Does your scope have "triggered sweep" ???  If so, you choose a channel to
trigger from, and set a level. Normally such scopes have "auto" and "trigger"
modes.  What model scope is it... or send me a picture of the front panel and
ill try to assist you.

Great work by the way. Get to know that scope... its well worth it!

>
>
> Thanks much,
> Elby