LLC converter: Why do output diodes have ZCS?

[eem2am]

I wish to do isolated LLC Resonant half bridge LED driver.
LED string voltage is 100V maximum, at 1A.

i wish to be able to analog dim it down to 0.1A.

-Is the reason for doing an LLC resonant converter for this, due to the absent reverse recovery in the output diodes.

I hear that LLC resonant converters make the diodes switch at zero current.

..but how is it that having a relatively small transformer primary inductance makes the diodes switch with zero current?

In an LLC converter , i understand that the ratio of L(leak) to L(Mag) is around 1:5 to 1:10.

...this measn that relatively high amount of primary current "sloshes around"

but how does this bring about ZCS in the output diodes?

-Will i be able to get ZCS for all LED current levels?...from 1A to 0.1A?

 

[Orson Cart]

You may have trouble getting a control range down to 0.1 amps with LLC, it is because you are going from inductive source to capacitive load thru the o/p diodes that you get "soft" switching of the o/p diodes.
Regards, Orson Cart.

 

[mtwieg]

For getting down to 0.1A, it may be necessary to do some kind of burst mode operation, where you run the converter in bursts at high load levels for short times, with a rep rate fast enough to cause visible flicker in the LEDs.

 

[eem2am]

Surely i could do the light load case (100mA) by simply allowing the LLC controller to increase its frequency?

 

[mtwieg]

Maybe, but at light load you might lose ZVS on the bridge due to not enough tank current. That, combined with the higher switching frequency, will cause efficiency to drop.

 

[Orson Cart]

You will need to go very high in frequency to get to 0.1 amp, possibly 5-10 times higher than for full power, which may be a stretch for the gate drive, however you will not lose soft switching of the diodes or the mosfets... Regards,Orson Cart.

 

[eem2am]

I have tried simulating LLC bridges with LTSpice, at various fixed frequencies.

I am noticing there that i am indeed getting non-ZCS switch-off in the output diodes at higher frequencys.

I am under the impression that the most important thing for LLC converters is that the Lm/Lr ratio is chosen to be small enough, that ZCS output diode switching is achieved down to the necessary lighter load levels.

At max load, i would expect an LLC to be switching at a frequency just below 1/[2 x pi x SQRT(Lr x Cr)].

where:-
Lr = resonant inductor (=leakage inductance)
Lp = magnetising inductance + leakage inductance
Cr = resonant series capacitor
Lm = magnetsing inductance.





Here is an LLC LTSpice simulation:-
2shared - download LLC.asc

 

[Orson Cart]

To lose soft switching in the diodes, you would need to have significant secondary side leakage inductance, so that one diode was still conducting as the other one is about to turn on, due to transformer drive, even in this case the applied PIV of the diodes is limited by the o/p cap and the fact that the other diode is on, so there will not be large voltage spiking on the diodes unless the layout is very loose.

Regarding the output current, given that you are driving an LED string, you may not need to reduce the o/p volts by much in order for the current to reduce greatly, LLC is a good choice where the o/p volt range is 2:1 max, i.e. 100V down to 50V in your case, this should be enough to reduce the current in an LED string to low levels, easily tested with a varaible power supply, Regards, Orson Cart.