Simple Power Supply Help

[PhasetheMusic]

Okay, here's the story: I'm trying to build a super-simple power supply for a LM386 based amplifier.  I need somewhere between 9 and 18 volts - not very close tolerances.  I've got a 25.2 volt center-tapped transformer that I'd like to use for full-wave rectification at roughly 12.6 volts AC.  With the transformer primary plugged in (my outlet was measuring 127 volts AC) I was getting an even 15 volts AC from ground to each secondary (30 volts AC between secondaries).  Okay, I can allow for some variation in transformer specs, esepecially without a load.  The center tap connects to ground, each secondary passes through a 4004 diode and the silver bands connect together at the output with the positive end of a 35V 1000uF filter capacitor whose negative is side connected to ground.   Pretty standard FWR set-up, from what I can tell.  The weird part:  I'm getting 20 volts DC at the output.  Maybe somebody can help me with the calculations because I am having a rough time of it.  Any help would be very much appreciated.

[Gearbuilder]

Hi,
I thing you could take a variable regulator like the LM 317.There's a lot of source about that on the web(check for DIY regulated power supplies )
Bruno

[spongebob]

Isn't DC voltage = AC voltage * 1.41 after rectifying? So 15 Volt AC would result in 21V DC, minus diode drop maybe.

If you want a smooth and stable power supply go with a voltage regulator (15V for example).

[smoguzbenjamin]

You're right, spongebob. What I'd do is use a regulator for 15V and use that. You might wanna heatsink it though because in my experience they sometimes burn out because of heat.

[R.G.]

You *could* read the power supplies basics article at GEO for how to do this.

[PhasetheMusic]

...if I were to swap out the 12.6 volt CT transformer for a 6.3 volt CT transformer could I expect about 7.497 volts AC (6.3 volts AC* 1.19 (the ratio of the 12.6 volts AC expected from the secondaries in the previous experiment to the 15 volts AC I actually measured from them)?  This 7.497 volts AC * 1.414 would yield 10.6 volts DC, which after the .7 volts DC dropped by the diode, would leave me with 9.9 volts DC, correct?  I'd like to avoid using a regulator, so this would be a workable situation for me, if my math is correct.  Thanks for the help.

[humbuck]

You could go to http://www.maplin.co.uk They have some fixed regulators at 5, 9, 12, and 15 VDC - They are about 50p. Look in the semiconductors section - you'll find it under "components". but I have some, if you want I could trade one with you,

H.

[smoguzbenjamin]

You know PTM that sounds like a clever idea But how much current would that be able to supply?

[spongebob]

The LM386 has only a mediocre 50dB PSRR (power supply rejection ratio), so that without a voltage regulator you might hear supply ripple on the amplifier output, depending on how good your filtering is.

[smoguzbenjamin]

Absolutely. My LM386 oscilated nastily with a normal 12V wallwart :?

[Gearbuilder]

Hi,
Yes Smoguz,he could use a 15v like you did but it's interesting to have a variable voltage regulator too.

Its exactly what i said at the top of this topic, use a LM317 regulator with a trim pot and you could change the voltage. If i remenber ,R.G. has a good shematic on his site :wink:
Bruno