OK, I see now: what worries you is the higher rms current of the post-converter.
That is completely normal: resonant converters, be they ZVS, LLC, or anything else require some storage and exchange of energy with the passive components, and this translates into a higher rms current for a given load.
Depending on the topology, this higher current can be limited (to some extent) to the passives, but it can also apply to actives.
That is the price to be paid for lower switching losses and quieter operation.
This is the reason why to this day, high power switching supplies are still... switching (hard): switching losses are almost a non-problem with modern MOSFets, and the industry knows very well how to manage the EM pollution aspect.
On the other hand, higher rms currents translate directly into costs: more copper, bigger FETs and heatsinks.
It is only when the supply becomes so large that one is forced to adopt resonant topologies, because parasitic parameters have to be taken into account, and instead of wasting power into large snubbers, it is more advantageous to use some kind of resonance mechanism to usefully recycle the reactive power
That is completely normal: resonant converters, be they ZVS, LLC, or anything else require some storage and exchange of energy with the passive components, and this translates into a higher rms current for a given load.
Depending on the topology, this higher current can be limited (to some extent) to the passives, but it can also apply to actives.
That is the price to be paid for lower switching losses and quieter operation.
This is the reason why to this day, high power switching supplies are still... switching (hard): switching losses are almost a non-problem with modern MOSFets, and the industry knows very well how to manage the EM pollution aspect.
On the other hand, higher rms currents translate directly into costs: more copper, bigger FETs and heatsinks.
It is only when the supply becomes so large that one is forced to adopt resonant topologies, because parasitic parameters have to be taken into account, and instead of wasting power into large snubbers, it is more advantageous to use some kind of resonance mechanism to usefully recycle the reactive power
But that means I have a 350W output with an 800W input? That's a really low efficiency!
No, that's not possible: contrary to a widespread belief, active power is not simply the product of rms current by voltage.
To compute active power, you have to make the product of the average current by the voltage (DC) at the output of your PFC.
Just chucking an idea in for LLC converters in general.... several years ago I had to fix an HV SMPS and it used a resonant output circuit for efficiency reasons but the really clever bit was that it had self-tuning based on current draw so it was always operating at the optimum frequency and maximum efficiency.
Ok, quoting the application:
So the peak current on the primary is:
sqrt[ ((Iout*PI)/(2*n))^2 + I1^2 ]
In my case this is 2.79A. The average current should be Ipk*(2/PI) that is 1.78A. Assuming the typical voltage of 390V that is 694W. So the input power at the LLC SMPS is 694W with an output power of 350W, that is a 50% efficiency.
Are the equations in the app note right?
This is the PFC stage, the input voltage range is from 180VAC to 260VAC and the output voltage is from 380VDC to 400VDC:
And this should be the LLC Half-Bridge converter:
My main concern is how much output power my PFC should be. If my SMPS has a 350W output power my PFC should provide about 400W of power to the resonant Half-Bridge. If you also have any suggestion for the circuits please feel free to tell me!
Woops. I think I got it. I'm the problem once more 
The primary average current is 1.78A but the voltage across the primary is half the PFC voltage! The maximum current is calculated with the minimum PFC voltage, so, 1.78A*(380V/2) = 338W! The exact output power of the SMPS is 337.5W so now everything is correct!
The PFC output power has to be 400W and the RMS current should be used to dimension the transformer wires.
The primary average current is 1.78A but the voltage across the primary is half the PFC voltage! The maximum current is calculated with the minimum PFC voltage, so, 1.78A*(380V/2) = 338W! The exact output power of the SMPS is 337.5W so now everything is correct!The PFC output power has to be 400W and the RMS current should be used to dimension the transformer wires.
When you properly write down the details of a problem, they generally (not always!) tend to vanish. This is why I ask for them: I felt there was room for simplification.
Try to think outside the box/ tablet/PC/notepad, etc.
I typically do this during my insomniac episodes, at around 4am when I cannot sleep: no distractions, nothing else to do and the benefit of a half-night sleep.
Some problems can be simple when you take them at their face value
the blocking cap is part of the tuned circuit so it has to optimized accordingly.
http://www.ti.com/lit/ug/sluu361/sluu361.pdf
http://www.ti.com/lit/ug/sluu361/sluu361.pdf
Yes, you can't be more right! Thank you very much for all your help Elvee!
Yes, in my case Cr is 22nF. I've read that using Cr splitted as in the schematic is better so each cap is Cr/2. Do you have any recommendation?
The splitted way is imho preferable, because both caps in series provide a bypass cap for the supply rails as well.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.