Insufficient specification, I'm afraid.
How much ripple?
How accurately 9v and over what load range?
How much noise?
What peak current? Average current?
Short circuit proof?
Is the 12v centre-tapped?
all have to be answered or at least considered.
The normal solution is to:
1)take the ac and rectify it with power diodes.
2)Smooth the output with electrolytic capacitors.
3) Reduce the resultant dc voltage to that required with a regulator.
The starting point is typically the regulator, if one is needed. So, say
you choose an LM138.
"The LM138/LM238/LM338 are adjustable 3-terminal
positive voltage regulators capable of supplying in excess
of 5A over a 1.2V to 32V output range. They are
exceptionally easy to use and require only 2 resistors
to set the output voltage. Careful circuit design has resulted
in outstanding load and line regulation comparable
to many commercial power supplies."
The data sheet and application notes, available online, will give you
the circuit diagram needed.
Of course, that could be way more than you need. If ripple, regulation,
fault tolerance, etc don't matter - all you may need is a rectifier and
The easy way is to use a bridge rectifier (get a big one 25a or so,
they are cheap) a capacitor a few hundred MFD at 35v or so. and that
gets you a fairly clean 15-16vDC. Then you can use a 3 terminal
regulator to whack that down to 9v and scrub the ripple off the top.
It is not the most efficient design, wasting 10-15 watts but easy.
If you can't find a 5a 3 terminal regulator you can get an adjustible
and use a pass transistor to ramp up the amps.
lots of ways to do this. in addition to the other posts you could use a
Triac dimmer in the primary side and rectify and filter the secondary.
is the secondary center tapped? a simple FW rectifier and filter cap will
get you 8.5 volts... well minus the diode drop
is it really 12 vac or the more common 12.6 vac? in any event if it 12 at 4
amps it will be somewhat higher at 3 amps.
If you are asking about a power supply of 12v ac to 9 v dc 3 amp
consisting of just a bridge rectifier and a series resistor:
A Google search on bridge rectifiers will tell you more than enough
about them! Probably almost any one rated at 5 amp or more would do.
The resistor is trickier. A single, fixed value, resistor needs to take
into account the regulation of the transformer, voltage drop of the
diodes in the full wave rectifier, etc. So, if you know all those
things, the value needed can be calculated.
If you don't know all these things: you start with a high value resistor
and a true RMS meter across the load and decrease the resistor until you
get 9v indicated at the load. I would use a 10 ohm 50W potentiometer and
slide it down until I got the right voltage. Then measure the resistor
and replace it with a fixed one of the same value. You may need a
combination of resistors to get the right value - each of which needs to
be physically rated for the power it would have to dissipate.
Bear in mind what the load will get - a supply with an rms value of 9v -
as requested. But with an instantaneous voltage that drops to zero and
goes way past 9v, many times a second. Any variation in load will see
the rms value of the supply change too. Any variation of the 12v ac
supply will also affect the voltage at the load.
Not many loads will like or even survive this! Hence, for electronic
loads, smoothing/filtering and regulation is almost invariably used in
place of the resistor.
Incidently, if you are replying to someone, then including what you are
replying to, or a part of it, helps a lot.
For the DC input, connect the 12 Vac output from your transformer
to the AC marked pins on a 6 amp bridge rectifier like this one:
Put a large electrolytic capacitor across the + and - pins.
Something like 33000 uF, rated no lower than 25 volts .
Use an LM7809 and the circuit shown here for the voltage regulator:
Mount the transistor and regulator on a good heatsink.
The solution that uses only a resistor to lower the voltage
to 9 volts works *only* with a fixed load that always
draws exactly 3 amps. The solution with the LM7809 works
regardless of how much current the load draws, up to 4 amps.
Without a smoothing capacitor, the voltage at this point would be
Now what a dc voltmeter would make of that would depend on what
technology it used. If sample and hold - then it would be producing
"random" values, depending at what instant on the half-sine wave the
sample was taken. Others would show the average value. Others would show
the peak value. Using a true RMS meter seemed the safest bet.
I have no idea what the load is - I am trying very hard to think what
load would like 3A of 9v full-wave rectified unsmoothed supply. Sticking
some filtering on it and adding a regulator IC has become so much the norm.
I have a strong feeling, that I am sure is shared by others here, that
the OP doesn't really want a supply consisting of just a full wave
rectifier and a resistor. That this won't meet the requirement. However,
the OP has repeatedly been told the disadvantages of such a supply and
the advantages of other, recommended, solutions - but still asked for
details of the rectifier/resistor solution...
you need a well filtered / well regulated supply or the audio section will
if you are unfamilier with electronic construction techniques building a
supply is going to be more of a hassle then its worth (unless you really
want the learning experiance).
the way i might do it: buy a 12 volt (13.8 VDC) regulated supply at the
local R shack. buy a half a dozen or so 5 amp diodes. wire these in series
each will drop about .6 volts. at 9 volts the buck will sing (in harmony
with big mouth billy bass) and life will be good.
It's probably cheaper to just get an old PC supply if that is the way
you want to go. They go on Ebay for $5.
I have one I will give away for shipping (Dell proprietary ATX)
That is cheaper than a radio shack supply and it will be a lot cleaner
power. I have had to add additional filtering to RS (ferro) supplies
to make them usable.