Resonant converters (LLC/PSFB) for low power<50W

[chaitanyab]

Hi,

Has anyone used a LLC or PSFB for power less than or equal to 50W?
I am trying to achieve more than 90% efficiency including the PFC stage. I think the PFC operates at around 95% efficiency so I need a DC to DC with 95% efficiency. I am designing for Universal input 110-230V, 5V output.
Active clamp Flyback is another one comes to mind but I wonder whether it can achieve 95% efficiency.

Please comment.

Thanks

 

[cupoftea]

The most efficient <50W you will get would be with a flyback with synchronous rectifier.
Power integratiosn make one...i think its called innoswitch.

The LLC and PSFB both relay on having an extra circulating current in the primary, and so they would not end up being very efficient at <50W. Also, the primary current travels through 2 fets instead of just one like on the flyback. Also, all the power for the fet drives of the LLC /PSFB would add to the power budget.

 

[mtwieg]

I'd recommend looking at laptop AC adapters as a reference. They typically have very high efficiency in order to shrink the form factor with only conduction cooling. I've seen reference designs based on active clamp flyback with peak efficiency around 94%. For 95% at 110V input, you may need to use a bridgeless design though. With a 5V output, your synchronous rectifier will be absolutely critical.

PSFB isn't suitable for such low power levels. Usually PSFB is used as a means to pass EMC, but at <100W that's normally not an issue.

 

[chaitanyab]

Hi Both,

I see there is 100W LLC converter so I wondering can there be 100W PSFB? so that I can give paralleling feature easily to modules too.

 

[cupoftea]

Its possible to do PSFB at 100W, but if you dont use synchronous rectifiers, then it'll be no more efficient than a flyback with synch rects.

 

[chaitanyab]

Its possible to do PSFB at 100W, but if you dont use synchronous rectifiers, then it'll be no more efficient than a flyback with synch rects.

yes I intend to use synchronous rectifier, however I was wondering to my knowledge the secondary diodes will go through ZCS switching (at least it happens in LLC, don't in PSFB ) and unlike mosfet there wont me a turn ON loss so how much really synchronous rectification would help.

Also I am seeing for synchronous rectifier driver ICs are available for center tap version, for diode based full bridge secondary rectifiers, I am not seeing any synchronous rectifier driver IC, has anyone done that before, synchronous bridge rectifier for PSFB or LLC secondary side?

 

[cupoftea]

I woudlnt use a bridge , because then you will need 4 sych fets.
"Turn on loss"...synch rects are all about reducing conduction loss rather than switching loss.

If you do full bridge output, then you have to make DIY synch rect driver.

 

[chaitanyab]

I woudlnt use a bridge , because then you will need 4 sych fets.
"Turn on loss"...synch rects are all about reducing conduction loss rather than switching loss.

If you do full bridge output, then you have to make DIY synch rect driver.

I was under impression the center tap would cause the Transformer utilization factor to go down as against full bridge implementation. I remember in another high power application when we analyzed transformer losses, found that the additional windings needed because of center tap were causing more loss and thermal rise. Especially proximity losses because of more windings.

I was thinking can a mosfet based full bridge replacing the diode bridge will help? I have seen in Dual Active Bridge but it needs an inductor on both sides for CLLC, now I don't know for PSFB can we simply replace the four diodes by a full mosfet bridge for unidirectional power flow?

 

[FvM]

I was under impression the center tap would cause the Transformer utilization factor to go down as against full bridge implementation.

Yes it does, but do you expect synchronous rectifiers to be lossless?

If you just want to design a high efficiency 5V/10A switcher with acceptable BOM, I believe mtwieg is pointing in the right direction. For a wider research about high efficiency topologies, start SPICE simulations with realistic transistor models.

 

[mtwieg]

yes I intend to use synchronous rectifier, however I was wondering to my knowledge the secondary diodes will go through ZCS switching (at least it happens in LLC, don't in PSFB ) and unlike mosfet there wont me a turn ON loss so how much really synchronous rectification would help.

When your output voltage is just 5V, diodes on the secondary side will cripple your efficiency. For example, if Vf=0.5V and D=50%, diode conduction losses will cause a 2.5% drop in efficiency. Double that for a full bridge rectifier.

I was under impression the center tap would cause the Transformer utilization factor to go down as against full bridge implementation. I remember in another high power application when we analyzed transformer losses, found that the additional windings needed because of center tap were causing more loss and thermal rise. Especially proximity losses because of more windings.

There are several ways of designing the secondary side rectifiers. A current doubler design is more suitable for low voltage outputs. These are commonly seen in PSFB converters, but AFAIK they can be implemented for LLC or any push-pull type topology (not flyback or forward).