I want to build a 12 V DC power supply. Which one is a better approach?

  1. If I have a step down transformer 220/12 volt and after doing its full wave rectification and smoothing its wave using capacitor, place the load in parallel to the capacitor.


  1. I should rather get a 220/15 volt transformer and after doing its full wave rectification and using smoothing capacitor i should use a 7812 voltage regulator at the rectified output along with capacitors recommended to use with voltage regulator? and after doing all that I should place my load to the output of voltage regulator?

Because I heard that voltage regulators reduces the rippling. Is it true?

  • \$\begingroup\$ Linear voltage regulators drastically reduce ripple, yes, but they make it more complex and consume some power. \$\endgroup\$
    – Hearth
    Oct 8 '18 at 15:44
  • 1
    \$\begingroup\$ Also, rectified+filtered 12VAC does not result in 12VDC! Research rectified voltage, diode voltage drop, filter capacitor ripple and sizing. Your plan 1. may result in a significantly higher voltage than 12V, and in your plan 2. you might not need a 15VAC transformer (12VAC might be enough). \$\endgroup\$
    – marcelm
    Oct 8 '18 at 15:49
  • \$\begingroup\$ Before you try 100 year old unregulated power supply designs, why not learn how modern efficient ones work? You can scope an unregulated design on many simulators if you learn the ESR or DCR of each part and see the crest factors in current that cause ripple. Lspice, Vspice & Falstad etc.... < free and then get a free Audio card scope software then buy a cheap PS and get schematic and learn. \$\endgroup\$ Oct 8 '18 at 15:53
  • \$\begingroup\$ @TonyEErocketscientist Are you suggesting the OP should make a SMPS? I think that's way too ambitious (for most people!). Consider the flyback transformer selection, the layout issues for high-frequency stuff, and the requirements for electrical safety, for example. A mains-frequency transformer on the other hand is simple and safe. Don't get me wrong, modern SMPSes are great, but if it's going on mains, designing them is something best left to experts. \$\endgroup\$
    – marcelm
    Oct 8 '18 at 22:33
  • \$\begingroup\$ @marcelm where did I say “make” ? I said buy a cheap PS \$\endgroup\$ Oct 8 '18 at 23:21

Adding a linear regulator will significantly reduce the ripple, but it has its downsides.

1) The regulator has a minimum "dropout" voltage, in this case about 1-2V depending on current. To account for this, your first stage needs to have an output voltage greater than 12V by at least the minimum dropout voltage - so, maybe 14V.

This means you're dissipating some power (V_dropout * I_load) in the linear regulator. In addition to power loss, you may have to pay attention to heat sinking it appropriately.

2) Depending on the ripple before the regulator, you may have to increase that "drop out" voltage even further, because you have to do the dropout calculation based on the minimum value of the ripple waveform (not the average "DC" value).

The advantage of the linear regulator is that the final waveform will be perfectly clean - the final ripple will be plenty low enough for almost any load.

So, what's the right approach? Depends on your load. If you're powering, say, a heater, the ripple probably doesn't matter. If you're powering a circuit with a secondary power conversion (IE, it takes your 12V and converts it to 5V), you're probably ok with high ripple, depending on the quality of that 12V to 5V converter. If you're directly powering something sensitive (say, an amplifier or other analog circuitry), you probably want to use the regulator.


I will not say anything extra about ripple reduction. Others have already made the necessary explanations.

I just want to warn you about the other things more important than ripple reduction.

Heat dissipation: Selvek has made a good explanation in his/her answer.

Size: If you use a 220V-to-12V step-down transformer then the output voltage will be around 16VDC after FW rectification. So you can directly place a 7812 regulator right across the smoothing capacitor (Remember the power dissipation here). BUT the RMS current drawn from the secondary winding will be twice (or even higher) the DC output current due to the need of a big smoothing capacitor. For example, if you're planning to make a 12V/1A power supply then the secondary winding should handle at least 2ARMS. For this reason the required transformer can be "bigger" than you imagined.

I just want to recommend you to make an offline SMPS but it looks that you don't have enough experience about the topic.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.