high-power flyback or SEPIC (or equivalent) converters?

[Alex]

Hi,

I need to design a converter that can raise or lower the input voltage
(D/1-D or equivalent transfer function). The converter needs to
tolerate 5kW at voltages between 20V and 60V and must be over 90%
efficient (at nominal power). I found that flyback and SEPIC
converters can do it, but i didn't find any such high-power
applications, so I fear that they can't be used. Anyone has heard
about the topologies in high-power applications or any equivalent
topology that can do?

Thanks,

Alex

 

[Rene Tschaggelar]

Hi,

I need to design a converter that can raise or lower the input voltage
(D/1-D or equivalent transfer function). The converter needs to
tolerate 5kW at voltages between 20V and 60V and must be over 90%
efficient (at nominal power). I found that flyback and SEPIC
converters can do it, but i didn't find any such high-power
applications, so I fear that they can't be used. Anyone has heard
about the topologies in high-power applications or any equivalent
topology that can do?

Highpower convertes tend to use a transformer.
With a specific turn-ratio, almost any voltage can be reached.

Rene

 

[Bill Sloman]

Hi,

I need to design a converter that can raise or lower the input voltage
(D/1-D or equivalent transfer function). The converter needs to
tolerate 5kW at voltages between 20V and 60V and must be over 90%
efficient (at nominal power). I found that flyback and SEPIC
converters can do it, but i didn't find any such high-power
applications, so I fear that they can't be used. Anyone has heard
about the topologies in high-power applications or any equivalent
topology that can do?

I'm not an expert on switch-mode power supplies, but I've fooled
around in the area enough to know that at the 5kW level you want to
use your ferrite as the core of a transformer. Set the turns ratio
such that the lowest input voltage can give you the highest output
voltage you need, and modulate the mark-to-space ratio of the drive to
hold the output voltage constant as the input voltage rises. You will
need to low-pass filter the output with a LC filter (or worse).

Your 90% efficiency should be attainable. There are probably
integrated circuits around to handle the control problem - the real
experts will certainly post their preferences. The switching
transistors and the gate drive chips will have to be substantial - 5kW
at 20V is 250A - but Farnell stock individual devices (in big
packages) that can carry this sort of current. You will probably end
up winding the transformer with layers of copper strip, separated by
transformer tape, which could make for a fun project.

 

[Alex]

Thanks to all, i'll take your suggestions into account. But let me precise my
application, just in case it gives you any new ideas.

Good guess, I need two of these. The objective i'm pursuing is the replacement
of a boost converter and a buck converter by two identical converters that can
either raise (boost replacement) or lower (buck replacement) voltages. Since the
converters already in place are quite efficient (certainly over 90%), the new
one must be at least as efficient. I know that this is quite ambitious, but we
need to rework the old ones anyway, so I just thought about this "universal"
converter.

Some complementary information:

Actual boost : Vin: 20V-30V
Vout: around 50V
Power controlled

Actual buck : Vin: 44V-60V
Vout: 25V-50V
Power controlled

Thanks,

Alex

 

[hosseinkd]

Hey Alex,

Did you build you Sepic converter? what is the rating and efficiency?

Thanks,
Hossein.