Hi everyone,
I have an question to do.
Can i use an transformer of a pc power supply (atx type) in reverse mode, in order to have an in dc high voltage output with 12vdc in input?
Of course using an appropriate pwm circuit as oscillator to High frequency.
thanks kyu9971

 

You need AC to drive a transformer; not pulsed DC.

 

Ok, thanks for your reply,
but into the pc power supply the ac voltage is first converted in dc and after send to transformer with pulse.
Is it wrong?
can you explain better please.
thank you
Kyu9971

 

Perhaps I was a bit hasty with my reply. Yes, a switched-mode power-supply works that way, but the primary winding is designed to withstand the inductive high voltage generated when the switching device switches off rapidly. The secondary (low voltage) winding insulation may break down if subjected to such high voltage caused by square-wave drive currents, but should be ok if driven with a sinusoidal current.

 

atx supplies go (after rectification) high voltage dc to low voltage dc

you are wanting to go the other way? -> low voltage dc to high voltage dc

simple answer is no.

 

Ok,
so i will must make before one sine wave and after drive the transformer.
some suggestion on to start?
thank you

 

I disagree. The switching currents in the secondary undergo almost exactly the same high-frequency content as the current in the primary. It is, after all, a switching power supply transformer. I see nothing wrong with driving the secondary with switched low voltage/high current, and getting switched high voltage / low current out of what was the primary.

BUT - without knowing a lot about the transformer, such as primary and secondary inductances, mutual inductance, etc., it will take a lot of fiddling around to get an efficient circuit. It never will work as well as a supply (and transformer) designed from the ground up to do what you want. Almost any topology other than flyback will reduce or eliminate driven-side inductive kicks.

ak

 

The secondary (low voltage) winding insulation may break down if subjected to such high voltage caused by square-wave drive currents, but should be ok if driven with a sinusoidal current.

Just a question on this statement. Are you saying they use two types of magnet wire to wind a transformer? I always assumed that the only difference was the gauge of the wire was different not the insulation?

 

A sine wave might work, but that's not how the transformer in an ATX supply is driven in normal operation. Normally, the full 370 volts DC is applied across the primary for a few microseconds, then removed for a few microseconds. This process is repeated maybe 50 or 100,000 times a second. So the voltage input is more like a rectangular wave, and the current input is more like a triangle.

In theory, you could do the same in reverse, by applying (say) 5 volts to the secondary for a few microseconds, then removing it in the same way, and then rectifying a high voltage out of the primary. You would need the electronics to pulse the secondary at the right frequency. The electronics are non-trivial, as you would be designing a switch mode power supply. If you want to try, maybe post a schematic of your concept here, and people can give feedback.

 

Are you saying they use two types of magnet wire to wind a transformer?

I doubt they do. I used the word 'may' intentionally. I've never stripped down a switcher tranny to see if, for example, there is inter-layer insulation apart from the wire enamel. I was just pointing out a possible problem, since the tranny isn't designed to be used back-to-front.

 

My idea was to use the tl494 (most diffuse in pc power supply) and connect 2 transistor on pinout 9 and 10.with these transistors i will have driven 2 mosfet (maybe irf3205) and after the transformer.
the central tap on +12vdc, the "sources" of mosfets to gnd and the drains to the other 2 tranformer pins.
what do you think?

 

What is the purpose of this?

 

are you making an inverter 12 to 240v?

 

I would like, if possible,
But now i am working on a circuit where there are two components that need of 220 vac for some seconds and max 11 watt.
so the idea of to realize an little inverter 12 vdc - 220 vac And so I thought to the transformer of atx power supply.
kyu9971

 

if your making a standard inverter, an atx transformer is no good as they work on 80 to 120 Khz, a standard mains inverter is designed for 50 to 60 hz.

 

I thought to set the first tl494 in order to have an frequency that you say (80/120 kHz) and then with a second TL494 I will drive the h bridge with frequency 50/60 hz As an inverter.

 

You need AC to drive a transformer; not pulsed DC.

A flyback transformer uses pulsed DC - Pre-ATX, quite a few PC PSUs were flyback topology.

 

it would be much simper, since you will need a totally different core material for the lower frequency you want to output, to make an inverter from 12 volts to 220 volts directly. or an inverter running off rectified line voltage to drive the 220 vac output. thematerial of the atx transformer will not work with lower frequencies without a lot of loss.

 

it would be much simper, since you will need a totally different core material for the lower frequency you want to output, to make an inverter from 12 volts to 220 volts directly. or an inverter running off rectified line voltage to drive the 220 vac output. the material of the atx transformer will not work with lower frequencies without a lot of loss.

alfacliff nailed it. A switching transformer will have too low an inductance to be driven at 50/60 Hz. It would be better to simply buy a 220 to 12 volt AC transformer, and wire it backwards.

 

Sorry, maybe I explained badly.
I would drive the transformer ( ATX ) say 100 khz , and get a high voltage that will be rectified ' around 250/300 vdc . Then by means of a "H bridge" drive by a another TL494 , in this case made to oscillate at 50/60 hz , I obtained a sine modified waveform.
I was inspired by an inverter repaired recently.
kyu9971