4
\$\begingroup\$

I am new to electronics. I am presently studying DC to DC converters.

One thing I noticed is that most of the people are using isolated stepdown converters like half bridge, full bridge, or resonant LLC converters for conversion such as 400VDC in and 28VDC out at 1kW. But there are no non-isolated converters like buck converters.

I am just curious to know why this is so? I am trying to build one AC to DC converter at 1KW, where my boost PFC gives 400VDC output.

I think 7% duty cycle is normal to drive given with the technology we developed (as D = 28/400 => D = 0.07), still i am unable to find any manufacturer who makes such PMIC with that small duty cycle at high power such as 1KW.

\$\endgroup\$
5
  • 5
    \$\begingroup\$ This is speculation, but 1. When you have a large Vin/Vout ratio (like 400 V in to 28 V out), it may make the design more efficient to use a topology that includes a transformer. 2. Once you are designing in a transformer it doesn't add much to the cost to make the design isolated. 3. You can always use an isolated design in an un-isolated application. \$\endgroup\$
    – The Photon
    Jun 19, 2019 at 16:01
  • 2
    \$\begingroup\$ But you can look for off-line regulators for a class of converters that is often designed with unsafe input voltages but without isolation (however, these have AC inputs, not DC) \$\endgroup\$
    – The Photon
    Jun 19, 2019 at 16:03
  • 1
    \$\begingroup\$ @ThePhoton good suggestion; most "off-line" designs will work with a DC input as well. I use almost an exact copy of the reference design for the ST VIPER26L to provide auxiliary 14V power for flight instruments from the 350-450 V DC battery bus in my electric airplane. \$\endgroup\$
    – pericynthion
    Jun 19, 2019 at 19:02
  • 1
    \$\begingroup\$ Well, it ain't exactly high-power so J won't write an answer but look at desktop PCs motherboards and realize that those 200W going into the CPU are mostly at 1.2-1.5V. It is common to see a 12-phase design in more expensive/overclocking motherboards. \$\endgroup\$
    – jaskij
    Jun 19, 2019 at 22:14
  • \$\begingroup\$ @percicynthion I have searched offline converters for the DC DC conversion, there are many manufacturers out there making those controllers (for example, UCC2880 by Texas Instruments, VIPER26L as you said) but they are for low power (at least as per datasheet) like 12W, 25W etc. My requirement goes to 1KW \$\endgroup\$
    – Voldemort
    Jun 20, 2019 at 6:24

3 Answers 3

Reset to default
7
\$\begingroup\$

Like Aaron said, a topology with a transformer allows for better control of the step down/step up voltage. A duty cycle of 7% seems fine on paper, but in reality you should aim for closer to 30%. In an application with changing load that 7% nominal duty cycle can drop down to 1 or 2% with low or no load.

When using a PFC, especially a boost PFC, the 400V(ish) output is chosen for the AC line. A universal input converter (90VAC-264VAC) will have a rectified voltage of 128VDC to 374VDC. If you are only doing low line (90-130VAC or so) then it doesn't need to boost to 400.

Then there is safety concerns. Any company is going to have to get UL/CE/TUV/KC or whatever other safety agency you can think of. Think about a buck converter, if that top switch fails, how are you going to protect the output from that HVDC? A transformer/coupled inductor provides galvanic isolation.

Then you start getting into efficiency concerns. A buck converter will always have turn-on, turn-off and conduction losses in the switches. There is nothing you can do about it. A resonant topology like a PSFB or LLC can lessen or eliminate the turn-on losses. Look up Zero Voltage Switching. Phase Shifted Full Bridges are really cool and a topology I work with often.

You can use a buck converter for a 1kW source, there are just better options. Of course, better is the enemy of good! For a medium power source like that, you can look into interleaved buck converters. These are essentially 2 or more buck converters connected in parallel.

\$\endgroup\$
2
  • \$\begingroup\$ I did found some ICs by various manufacturers like texas instruments UCC2889 IC which can drive PWM at 7% duty cycle but, the output cannot be delivered with high power, don't know why is it so ? \$\endgroup\$
    – Voldemort
    Jun 20, 2019 at 7:46
  • 1
    \$\begingroup\$ Think about it as the transfer of energy. If you have a short duty cycle that means you have a short pulse of energy being transferred per period. Power is energy over time. If you're delivering a smaller amount of energy over a constant time period you have a smaller max power available. \$\endgroup\$
    – user187594
    Jun 21, 2019 at 13:21
5
\$\begingroup\$

With a simple buck converter, you only have one degree of freedom, the duty cycle. And as you are noticing, the duty cycles for those input:output ratios are very small.

With a transformer design like a flyback, you get two degrees of freedom, the duty cycle and the transformer winding ratio. Usually you design the winding ratio so that you can be in the sweet spot of the duty cycle.

\$\endgroup\$
2
\$\begingroup\$

For consumer use, it is not popular, for light industry it is not common, but for heavy industry (MVA generation) as long as somewhere in the system design there is isolation and safety earth ground, there is no restriction for isolation if no service access is needed or safety hazard imposed in use.

You are likely never to see the latter unless you are involved in high power system designs such as Megawatt PV Grid Tied Inverters (GTI) powered from Solar Farms.

All I can say is , they exist in your power range, but you must be aware of the consequences of lightning strikes and lack of safety certification on home insurance.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.