UCC28251: Asymmetric half-bridge operation for off-line converters 400Vin to isolated 24Vout

Hello,

UCC28251 OUTA and OUTB are symmetric, this is good device for offline application with many advanced features. 

If you want to use asymmetric control, maybe you can take one OUTPUT, then INVERT it to get complementary output for the other switch. It is complicated, but doable.

Nitish,

UCC28251 OUTA and OUTB are asynchronous D outputs not suited for the asymmetrical half-bridge (AHB). AHB requires a controller capable of producing D and 1-D outputs and TI has many controllers of this type. Any D, 1-D Application Specific PWM can be used from TI. these include the PWMs assigned to the active clamp topology and I parametrically searched them here for you.

Why have you chosen the AHB topology?

The AHB has some niche equipment applications but is not such a popular topology because it really works best when D is close to 50% (narrow VIN range) but the LLC converter also works best with narrow VIN and fixed VOUT and with the large number of LLC controllers in the industry, the AHB has sort of fallen by the wayside. Still, the AHB is a fixed frequency topology suitable for higher power compared to other 2-switch topologies and the LLC is a variable frequency 2-switch topology. There are some nuisances with the AHB converter topology you should be aware of. Please be sure you understand the trade-offs and limitations which are summarized below:

Regards,

Steve M

Nitish,

The asymmetrical voltage stress exists because of the wide range of D. In the graph I showed, you are only seeing a ~2:1 D variation and you can see how widely displaced the voltage distribution is between the two rectifier diodes. If you can find rectifiers that are safely rated to the worst case voltage/power stress than no worries. From a transient response point of view, the blocking cap in the AHB is similar to the resonant cap in a active clamp. Make it large and you reduce the ripple but take on the poor transient response because the cap voltage can't change instantaneously. Make it small and the transient response improves but the voltage ripple voltage increases. You need to find the sweet spot when you tune/test the hardware on the bench. 

Steve

Nitish,

Since AHB is not very much promoted at TI, I can't find and y direct comparison to LLC. I've done these types of comparisons and wrote some training materials you can find in the public domain here. Pay attention to slide 63 which I've clipped below from the slide deck.

Regards,

Steve

Nitish,

1) Sean mentioned that UCC28251 is for symmetric driving and Steve mentioned that its outputs were asynchronous D outputs.

1a) From the above I take it that both the upper and lower switches of the half-bridge at driven with the same duty cycle D....which would meant that D<0.5. Would you please point me to where in the datasheet the symmetric drive is mentioned?

[TI]: "symmetric" just means what you stated and is shown in Fig A, "Symmetrical Half-Bridge Diagrams" from the first diagram. I don't see the word "symmetric" mentioned in the UCC28251 data sheet but is implied by the waveforms below. You can see that OUTA and OUTB each respond to every alternate clock pulse and the combined result is symmetric:

1b)  Kindly elaborate further your statement "UCC28251 OUTA and OUTB are asynchronous D outputs".  Since there is only one feedback pin whose voltage determines the duty cycle to drive the upper and lower MOSFETs with, I am really confused by "asynchronous outputs" mention in your statement.

[TI]: Asynchronous (symmetric but 180 out of phase) means the same as "symmetric" from above. Sorry if that term caused confusion.

1c)  Would you please point me to where in the datasheet the "asynchronous outputs" feature is mentioned?

[TI]: Refer to 1b above

2)  LM5036:  What does input voltage feed-forward mean?

[TI]: In VMC, the control loop gain varies as a function of input voltage variation. VFF modulates and adjusts the RAMP signal proportional to input voltage variation to help improve regulation and transient response by overcoming this limitation of VMC.

3)  Since Vin in LM5036 is limited to <100V, can I use it to drive a half-bridge connected to 400Vdc bus if I use isolated gate drivers and supply it with Vin generated from an auxiliary source?

[TI]: LM5036 has internal integrated HV start-up and the pass transistor is limited to 100-V. You can start-up from higher line voltage but you need to bypass the internal start-up as described in the LM5036 data sheet, section 8.2.2.4.

Regards,

Steve M

I am closing this thread. Thank you.