I've learned just enough SMPS theory to be a danger to myself From the amazing claims made about the wonderous abilities on SMPS units, it would seem possible to take one and feed it ~12 volts (like from a car) and expect it to function properly, assuming the 12 volts goes to the right place. But being as far along in education as I am, I've learned theory rarely is reality. So, say I tried to feed the PSU power from a car right after the full wave bridge rectifier on a typical PSU, what sort of fire am I looking at making?
Let me ask you this: do you even know that the input AC supply is rectified to a potential of around 12-15 volts after leaving the bridge rectifier? Have you measured it? Because if it isn't, you are asking for trouble. Actually, no matter what, you're asking for trouble until you get some advice from a SMPS professional.
If a PC power supply behaves as a theoretical SMPS, then the input voltage is irrelevent. The theory of operation of a SMPS is radically different from that of a standard linear supply. Think of them as storing power in a big container and squirtng it to the load when it's needed instead of trying to limit the flow of a constant stream. Because of this storage, they can deal with a lot of different input voltages, at least in theory. I was hoping that a PC supply is close enough to theoretical to deal with an input about 12 times less voltage than it typically would (based off of US mains voltage). The theory of SMPS says yes, but my area of study is digital processors and DSPs, not power electronics, so I on't know what I can get away with. An old AT supply might just become a victim of an experiment complete with blast shield and fire extiguisher. But I had hoped somone might know enough about the real things to tell me if I can get away with it.
I would probably guess the power supply itself would just be out of regulation and just either quit or output a voltage that is lower than the input. The reason being is that the regulator circuit will notice that the ouput voltage is not in tolerance and so increase the frequency of its PWM to compensate, but because of voltage drops, I am assuming the battery is exactly 12v, the 12v rail will never be exactly 12v unless the voltage input is higher or if there is a built in step-up circuit. If the input voltage is say... 14v, then that may be a possibility. The reason being is that aren't power supplies for a computer designed to be step-down and not step-up? There are different types of switch mode power supplies: step-up, step-down, inverting, ect. Therefore the design of it will probably limit you unless you specifically find/build a power supply designed to do such a thing. You can measure the output voltage of the bridge rectifier and see how much voltage is being sent out and see if it is possible. Other then that... go for it if you want. (I accept to responsibility) Here is a website that should further help you in the theory and construction.
Well, the additional info is very welcome, though iI'm still scratcing my head a bit over it. I can definatly see where the PWM controller might go crazy, but I'd assume the soft start circuit would have a fit first. One potental victum I have for experimentation does appear to have a step up circuit in it, or at least the LC circuit, transistor and other parts sure looks a lot like an oscillator that feeds a transformer. I'm hoping that once the main resevior caps charge, it'll be enough for the lower voltage of the car's system to keep it charged. As I said previously, the theory seems to agree with me, but theory and paractice aren't always the same thing So I'm off to build a protective box now, and some dummy loads.
Its easy to say, if you try to draw say 200W thats 16+A at 12v, the HV side is only meant to handle 6A on a large PSU(400W with some extra for inefficiency). So it would quickly overload the HV side of the PSU. I would think that you could feed 300vdc straight to the caps and it should run then. But it might be kind of hard to get 300vdc in a car.
