Transformer utilisation with *BIG* choke input filter

[Circlotron]

Say we have a 100VA transformer feeding a bridge rectifier and then straight into a large filter capacitor. The power factor presented to our transformer will be about 0.6 and so we would be able to pull 60 watts DC from our 100VA transformer for it to be fully loaded.

Then we put a big iron cored choke between the bridge rectifier and filter capacitor. And not just any old choke, but an ultra-choke (the dual of an ultracapacitor ) Because of its huge inductance the current through the choke is essentially constant and so the current in the transformer secondary is a square wave.

Seeing the transformer instantaneous current is now constant and not having gaps that need to be filled by current peaks with their consequent I2R losses, could we assume that the transformer losses are even lower than what they would be if the transformer was supplying a purely resistive load? Just how hard could we push the transformer if we used a ridiculous size inductor? What might be the maximum utilisation factor? This is obviously not a practical approach, just a thought experiment.

[Jay_Diddy_B]

Hi,

If the inductor is infinitely large so the current is a square wave and the diodes are ideal, no voltage drop.

using the concept that PF = VA/W

Apparent Power/Real Power

The Power Factor is  2 x SQRT(2) / Pi = 0.9

This the ratio between the Average of a sine wave and the RMS of a sinewave.

So you can get 90W from a 100VA transformer.

You get a 100W if you use a PFC that is 100% efficient.


Regards,
Jay_Diddy_B

[trobbins]

Circlotron, it is also worth noting the additional transient stress this applies to secondary winding waveforms and diodes.  The secondary winding current has to essentially flip in direction.  The diode currents have to effectively start/stop instantaneously.  Nothing happens instantaneously, and so you get dragged in to the underworld of stray/leakage/parasitic components that can't or shouldn't always be simply ignored.

So the thought experiment should really start to look at secondary leakage inductance and how large that is, as it effectively lowers the utilisation time by introducing a transitional glitch twice each mains period.  There are other secondary losses introduced also, like the cyclical charging/discharging of stray capacitances to the rectifier/choke node.