1
\$\begingroup\$

I'm trying to make a buck converter which produces an output voltage of 3.3V with a load of around 1 amps when the input is around 220v AC. This basically means converting from 311v DC to 3.3v DC. I'm basing my circuit on

What I wanted to ask, is such a circuit feasible? Practically not theoretically. It seems quite simple. I calculated the inductance at around 26uH, and capacitance to be a minimum of around 50uf, However I'm using 100uF on both the input and output capacitors.. I have a duty cycle of around 1% since the output voltage is tiny in comparison to the input.

\$\endgroup\$
  • \$\begingroup\$ For safety purposes, you would want some kind of isolation from mains voltage. \$\endgroup\$ – mkeith Jul 31 '18 at 7:22
  • 1
    \$\begingroup\$ take care of the switching frequency, if you take the Fsw to high the 1% duty becomes very short. And as a result, the turn-on and turn-off time of the FET can become a problem. \$\endgroup\$ – Navaro Jul 31 '18 at 7:29
  • \$\begingroup\$ What’s your output current and max allowable voltage ripple on the output and max current ripple on your output cap? \$\endgroup\$ – winny Jul 31 '18 at 7:45
  • 3
    \$\begingroup\$ That is a very simple circuit used to describe the basics of how a buck converter works, it is not a practical circuit that will work reliably, efficiently and safely. \$\endgroup\$ – Steve G Jul 31 '18 at 7:47
10
\$\begingroup\$

What I wanted to ask, is such a circuit feasible? Practically not theoretically.

I'll go with a "no". The time you'll spend in such an open-loop configuration to actually make things stabilize at the voltage you want will be quite substantial, and the slightest change in temperature will change how the transistor works, so: don't try this. Also, you'll need a >> 300V breakdown bijunction transistor, and that's going to be unnecessarily expensive, and probably waste a lot of power, thus will be very rapidly changing in characteristics due to heating up, thus making everything worse.

Also: at 1% duty cycle it is very hard to maintain stability of anything.

It seems quite simple.

That's because it's a very textbook example, seemingly taken out of context. No one in their right mind would recommend building something like this!

I calculated the inductance at around 26uH, and capacitance to be a minimum of around 50uf, However I'm using 100uF on both the input and output capacitors.. I have a duty cycle of around 1% since the output voltage is tiny in comparison to the input.

Exactly, and because this circuit was meant to illustrate the principle, not to work at edge conditions, it will work miserably, if at all.

Don't do that: copy circuits from random websites which try to teach you basics. If you want a power supply, there's ICs for that, which take all the burden of actually building something that works from your shoulder: In practice, a switch-mode converter would measure its output, and adjust its duty cycle automatically to achieve the desired output. That's much cleverer than just calculating things and hoping for the best in terms of component tolerances and parasitic effects (which will absolutely ruin your day here).


regarding how to build that:

Actually, I'd recommend not building something like this yourself. 300V can and will kill you and the ones you love. Considering a minimum textbook example might be a sign of "needs a bit of experience" before you want to build things that will potentially electrocute you.

However, if you insist: You'll need something with galvanic isolation between the primary and secondary side here to even fulfill the most basic safety requirements (i.e. to satisfy your will to live).

That means you want a flyback transformer bases switch-mode power supply, where you probably have a switch which switches the primary side of a transformer on and off – on the secondary side, you'd have the voltage and/or current sensing, and you'd have some way to transport the signal "voltage high enough" to the switch controller on the primary side.

The fact that your transformer can have any winding ratio (not just 1:1) means that you don't have to operate at a 1% duty cycle – which would be very good, because such marginal duty cycles make it very hard to regulate stably.

You'll probably have to wind that transformer yourself, or order it special-made. Maybe you can buy one that fits your needs.

Such switch controllers can be bought relatively cheaply, and they're not totally out of this world in terms of complexity of what they do internally, but you'll need quite a firm grasp on electronics and control theory to fully appreciate what happens. However, you don't need to understand them fully – they come with good application notes explaining how to build such a step down converter.

Look at ON Semi's website. They have a lot of such "off-line controllers"; things like the FSCQ1265RT integrate the switching transistor (which is a MOSFET at these voltages), and you'll "only" have to calculate the values for the rest of the components. For many of these, evaluation boards with online schematics are available. Again, it's not overly complicated, but it's not a good beginner project, because you can kill yourself, and because shorted out things connected to your 300V power source probably go "poof".

\$\endgroup\$
  • \$\begingroup\$ +1 in total agreement - the op needs to use a flyback design and not buck. \$\endgroup\$ – Andy aka Jul 31 '18 at 9:22
  • \$\begingroup\$ "Load of 1 amps" - it's right there in the first sentence. \$\endgroup\$ – Arsenal Jul 31 '18 at 10:03
  • \$\begingroup\$ @Arsenal now I feel rightfully stupid! \$\endgroup\$ – Marcus Müller Jul 31 '18 at 13:58
  • \$\begingroup\$ @MarcusMüller well it's not gonna change much of your answer - basically the circuit + question combination speaks for little experience. \$\endgroup\$ – Arsenal Jul 31 '18 at 14:42
  • \$\begingroup\$ @Arsenal true, changed the answer accordingly. \$\endgroup\$ – Marcus Müller Jul 31 '18 at 14:54
1
\$\begingroup\$

There are several reasons to not implement your circuit without transformer:

  • It has no isolation between input and output - it's dangerous for most use cases.
  • Switching transistor needs to have both high voltage and high current capabilities.
  • If switching transistor breaks, all 300 V will be applied to the load and will burn it.

Transformer (flyback topology) solves all that problems, and I suggest to use it.

Also, making AC/DC coverters is not right starting point without enough knowledge and experience. It's complex and literally dangerous. May be, it's better to start with low voltage converter circuits (less than 30 V).

\$\endgroup\$

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.