Can PSU Be Connected to DC Source?

[iulian28ti]

Can a PSU be connected to a DC source ?
Asking because...
http://images.electronicsinfoline.com/View/Photo/10001/Atx_power~0.gif"
...i see the AC input, followed by capacitors and inductors (or are they transformers?), then the line is rectified. But we don't necessarily need AC to pass through a rectifier. DC should be used with no problem. Am i right ?

 

[MATLABdude]

I don't know if I completely understand your question, but yes, an ATX power supply supplies DC voltage to your computer system (a whole multitude of voltages, in fact, varying form 1.2V up to 24V--I think, might actually only be 12). It makes a nice, cheap, fairly linear, and often quite clean (depending on who actually makes it) DC supply. Unfortunately, you don't get really fine voltage control nor do you (usually) have overvolt or short circuit protection, so be careful!

 

[iulian28ti]

Yes, yes, yes, it provides 3.3, 5 and 12 volts.
But my question was... due to its internal construction, is it possible to hook it up to a DC input ?

 

[MATLABdude]

Yes, yes, yes, it provides 3.3, 5 and 12 volts.
But my question was... due to its internal construction, is it possible to hook it up to a DC input ?

Also provides some others, but if you're asking if you can power the AC power supply from a DC power source? I would highly doubt that, unless your power source is outputting between 120 and 170 Vdc (that's likely what you see after the rectifier and big capacitors).

EDIT: What is it that you're trying to achieve? DC-DC conversion?

 

[iulian28ti]

Yes. Basically i want to jump over the rectification stage and provide it DC which it will convert to 12/5/3.3v through switching.

 

[MATLABdude]

Yes. Basically i want to jump over the rectification stage and provide it DC which it will convert to 12/5/3.3v through switching.

Depending on the internal construction, it may, or may not be able to do what you propose (my money's on not, however). The rectifier will produce a rectified sine wave with a 170 V peak, and with filtering you'll still have something like 100 Vdc. The DC-DC converter block is usually designed to expect this voltage, and will be unlikely to output the voltages that you see in a computer system (if it produces anything at all).

You'd be better off buying a DC-DC converter, or just a linear voltage regulator (only converts voltages down). Depending on what you wish to use to supply the power to the module, you may need some additional filtering (e.g. a car alternator--very noisy-- vs. a battery--very stable). Switching supplies are sold either naked (with none of the supporting inductors, diodes, etc.) or integrated in a module (voltage in, black box with no external components, voltage out).

 

[vk6kro]

In your circuit, it looks like the 110 volt input is voltage doubled to get enough voltage to run the power transistors. This would not be possible with DC.
However, if you had a source of about 325 volts DC, it would probably work OK on 250 volt mode.

You would have to arrange a "power good" signal for the LM393 but this is probably just 5 volts.

 

[waht]

That's what the first stage of the power supply is doing, it is converting AC to DC; probably 325 volts as vk6kro said,

 

[berkeman]

Yes. Basically i want to jump over the rectification stage and provide it DC which it will convert to 12/5/3.3v through switching.

Three things. First, you are doubling the stress on two of the four input diodes. They may be able to handle it, but they may not, depending on thye margin that the design has built-in.

Second where are you getting this high DC voltage from?

Third, how much training and experience do you have with high voltage circuits? There is a significant shock and fire hazard with high voltage circuits.