Hi Omid,

There are various ways of viewing the increase in losses, depending on how you analyse the circuit.

One easy way to do this is in terms of increase in current harmonics as the converters approach discontinuous conduction mode.

Since the active power transfer in a DC-DC converter can only be by the DC voltage and DC current components (since voltage and current have to be of the same frequency), all harmonic currents will contribute solely to losses.

So as the quantum of harmonics builds up, the losses increase too.

With best wishes.

-Sanjay

Thanks for your respone Mr. Roy,

The considered approach to determine power losses in the IGBT module is based on lookup table (data are extracted from data-sheet).

Is there any relation between output current and power losses?

I saw a figure in a paper that showed the increase in output current leads to a decrease in efficiency.

Hi Omid,

First of all, let me clarify that my previous answer referred to losses in circuit - not device losses. (I am not deleting the previous reply because as far as the circuit  losses are concerned, that is still a correct interpretation.)

Device losses (or as an aggregate, module losses) have little to do with the circuit current harmonics, because device switchings are always very fast in comparison the circuit switching period. Device losses, and therefore device efficiency (module losses or efficiency) depend more on the way the device itself flips between ON and OFF states.

In an IGBT, the device losses depend mostly on the losses of the BJT power side device traversing its active region. Note that when the BJT is in cutoff, the current is negligible, and when it is in saturation, the collector-emitter voltage is very low. In both these extreme cases, the device losses become very small. But when the device is changing from one to the other, it goes through the active region, in which the instantaneous product of the device voltage and current is never negligible except at the ends!

Going back to your original question, when the buck convertor operates at very low duty cycle or the boost convertor operates at very high duty cycle, waveforms will show you that the IGBT turns ON and OFF in very quick succession. So if the switching is slow, the BJT has trouble getting out of active region in such cases ! Visualise a switching where the device begins to turn ON, is not fully ON, but in the meantime the gating change wants it to turn OFF - you will immediately see why the device has more losses, and why the device temperature will consequently rise !

In general there are three major factors that can contribute to slow switching in IGBT's:

Other than #1-3, the device tends to heat up when you operate it at a very fast operating cycle frequency, simply because the active region traversals occur more frequently per second (this effect is similar in all switching devices, nothing exclusive about the IGBT).

That should explain the phenomenon.

-Sanjay

Thanks you so much for your grateful comments,