why not use high switch frequency in dc/dc converter??

[batistuta]

switching frequency dc-dc converter

hi,all

just as the topic, i want to know in dc/dc converter, why not use high frequency switch frequency, as we know, use high frequency can get small passive component, but high frequency will causes more switch loss,
but as the process goes down, i think the parasitic does not change so much, so why not use high switch frequency?
this is my ideas, if wrong, please check it...
so i need your idea, at high frequency just consider switch loss ? and other things that influence efficiency at high frequency?
thanks

 

[TekUT]

dc-dc converter switching frequency

This is a complex issue, well I can try to give you just some idea.
Switching frequency is strictly related to power dissipation issue into the switch elements (transistor, diode), parasitics (capacitive, inductive) and to EMI emission.
Small sized converter can be obtained by increasing the switching frequency in order to small down the size of the storage element, mainly inductor and then capacitors.
A way to increase the switching frequency is change the topology from the traditional hard-switching power cell (the switch turn-off and/or turn-on is dissipative then with current and voltage on the switch different to zero) by using more complex topology like the soft switching quasi square wave, quasi resonant converter, resonant converter, multi resonant converter or more generally by using soft switching techniques.
Basically soft switching means change the state of the switch element when current or voltage or either is equal to zero, this in theory is able to null the power dissipation into the commutation stage into the switch element, less voltage or current stress of the converter cell devices, and also lower EMI emission than traditional hard-switched topology.
With these topology the voltage and current waveform inside the active switch and auxiliary switch take soft shape like piecewise linear with sinusoidal portion attached, pure sinusoidal, sinusoidal with different frequency value. Of course there is a drawback, modeling the converter behaviour is not simple and the control strategy is more complex than simple hard switched PWM converter (often we have on time control, off time control, frequency control and so on).
To give you an idea, about quasi resonant converters (QRC's) there is a unified theory that is able to study the dynamical and steady state behaviour of the converter under quasi resonant condition and PWM hard switching condition. Also parasitics like leakage inductance into the transformer for Flyback converter can be used to achieve the soft switching action, then most often parasitics is a good matter instead a problem like into the traditional converter topology.
To give you a idea I've read paper about quasi resosnat converter working up to 10 MHz then very high frequency in order to have small converter size and high power density. Commercial device wich use soft switching technology are actually available, for example VICOR (http://www.vicr.com/) use the quasi resonant convertert topology with ZCS technique (Zero Current Switching control action).
Here below some link about this type of converter:

http://ece-www.colorado.edu/~pwrelect/book/slides/Ch20slide.pdf
http://www.fairchildsemi.com/an/AN/AN-4146.pdf

and some reference to book and paper about high frequency converter:

**broken link removed**
"High-frequency quasi-resonant and multi-resonant converter technologies"

"Industrial application, 1 MHz, quasi resonant converter"
Carrasco, J.M.; Ridao, F.P.; Quero, J.M.; Janer, C.; Franquelo, L.G.
Power Electronics and Applications, 1993., Fifth European Conference on
Volume , Issue , 13-16 Sep 1993 Page(s):178 - 183 vol.3
Digital Object Identifier

To give you also a idea I've attached a image of a converter that I've builded for my master thesis on power electronic, this is a quasi resonant ZCS Flyback converter (half wave) with switching frequency between 300 kHz and 400 kHz (variable frequency control).

Hope this can help.

Bye
Pow

 

[nebo123]

POW can you please post a pdf copy of your Master's Thesis on Quasi Resonant Flyback ; it would help us a lot!

IV

 

[chuckey]

Once you have found rectifiers that will rectify say 12V at 5 A at 10MHz, then your dreams will come true. Big current needs big junctions which have large capacitance, so the efficiency of rectification goes down. Ordinary power rectifiers's performance falls off at 400 Hz.
Frank

 

[dick_freebird]

Your usual process tradeoff is speed for voltage. A
high frequency silicon process is unlikely to stand off
high voltage. If you're stuck with the same high VIN
then your options are not as broad as you'd like to
think.

But I think GaN FETs are going to bust this apart,
with their fairly small gate drive amplitudes and
high frequency, high voltage capability.