UCC256404: Useful for wide ranging output voltage?

Joe,

Can you please share more info about the application, system requirement or end equipment? Is it for TV?

For wide Vout application, we'd recommend UCC25660x, our latest LLC controller. You can find additional info here: https://www.ti.com/product/UCC25660

Best,

Ning

Ning,

Thanks.  I have also looked at the UCC25660, but my fundamental question still remains.  Is an LLC halfbridge capable of performing decently over an output voltage range of 80-285V?  I'm a little outside my typical design space, and was trying to go through the excel design sheets, but I get vastly different results when I use a Vout of 80V vs 285V.  I wasn't sure how easy it would be to split the difference.

The application is an off-line LED driver, mostly for outdoor lighting.  My customer just wants to be able to use a single LED driver for a wide rage of light fixtures, hence the wide range of output voltage.

The reference designs for both of the LLC chips show a constant output voltage, whereas I want to design a constant current output across that 80-285V range.  I wasn't sure if I was looking into the correct topology and chipset.

Thanks,
Joe

Joe,

No matter what LLC controller to use, the fundamental of LLC topology doesn't change. >3x output voltage range is doable but requires careful design and will need to trade off certain performances, like efficiency. You can consider add a boost stage to make the design easier.

Depending on your thermal budget, you can also consider active clamp flyback (UCC28782) topology. In order to change the Vout, the FB loop needs to be changed either dynamically by another controller (like PD controller) or via resistor change, regardless of the topology. 

Here is a 220W flyback reference design. Power stage needs redesign to accommodate the higher Vout.

https://www.ti.com/tool/PMP23224

Best,

Ning

Ning,

Thanks.  My plan, after some more thinking about trade offs, is going to be a 3 stage design: Boost-PFC to create a 450V rail, then a LLC half bridge to create an isolated 75V rail, and then a boost to regulate the current through the LEDs (with a Vf between 85V and 280V).  I'm still go the base the first two stages around the PMP40580 demo board.  I do have a few questions related to the UCC28056 PFC chip.

1) In the datasheet, equation 17 uses T.onmax0 * G.ff1, but the excel design file T.onmax1.  The give different answers, and I'm not sure which one I should be using.  This affects the calculation of the max boost inductor value, which affects the Rcs calculation, and therefore the required rating for the inductor's peak current.

2) In the excel design file, on the schematic sheet, there is a node between Ros1 and Ros2 that is labeled BLK.  I am assuming this is meant to connect to the LLC stage's IC's BLK pin.  Is that the intention?

3) To calculate the max rms current through the boost inductor, equation 26 in the datasheet of the UCC28056 is different than cell C63 in the excel design file, and they give significantly different results.  I could use help understanding why.

Thanks,

Joe

Joe,

Is your input AC or 450V DC? If DC 450V, there is no need for a boost PFC stage before LLC. 

Please clarify.

Ning

AC - 120 to 277.  450V was just my input to the LLC stage.

Hello Joe, 

You raise some interesting questions about the Excel tool and the datasheet. 

Unfortunately, I think I can answer only one of them at this time with any certainty.
That is #2:  Yes, the Vblk node between Ros1 and Ros2 is intended for the UCC256301 LLC controller, and specifically at a 330V bulk voltage threshold on a 390-V PFC output.  See rows 159-161 in the Excel tool which drives the Ros2 recommendation.  
If you are not using this node, you can choose Ros2(actual) = 0 and the Ros3 recommendation adjusts to achieve the target PFC output voltage value.

The answer to #3 may be linked to resolution of question #1, but I'm not sure.  I'll need some time to study these equations and trace their derivations.  

I feel vaguely confident that equation (26) in the datasheet is generically correct, so that suggests that the Excel tool "short-cut" equation relies on suspicious cell C62 results whose equation may be linked to the discrepancy of question #1.   

Please allow me a day or two to investigate further. 

Regards,
Ulrich

Ulrich,

Thanks.  There isn't a huge urgency, here, so no worries about taking a couple days.  It was just a discrepancy I'd like to understand before finishing up my design.  I'm still a ways away from starting a board layout - a bunch of other stuff to figure out that doesn't have to do with these questions.

I figured as much, regarding question 2.  I have a independent R-divider for the LLC stage that connects to the BLK pin, so I can just eliminate that connection from the PFC's R-divider.

I am starting to get into the LLC design with the UCC256404.  I'm not too far into that, but I did notice a difference in topology between the datasheet (and excel file), and the PMP40580 demo board.  The demo board doesn't appear to have a resonant inductor nor a single resonant cap.  Am I correct in assuming that the resonant inductor is just the leakage of the transformer's primary?  Do you have a specification for that transformer (BOM lists the part number as BCK-50-722D).  And am I correct in assuming that the resonant cap is split between C203 and C204 on the demo board?  If so, what is the benefit/detriment to having two resonant caps vs a single Cr?

The output of my LLC stage will be 75V (along with a primary and secondary aux voltage of ~15V), and will require about 200W of capability.  On the demo board, the 12V output, which is what's regulated, uses a center tap secondary, and the 150V rail uses 4 diodes for full-bridge rectification.  My understanding is that the tradeoff is number of components vs. electrical stress on those components - that with a center tap, the diodes have to handle twice the reverse voltage, and only half of the output winding sees current at any given time (so more copper is needed).  Is that accurate?  And if so, my other assumption is that neither topology really affects the design of the LLC converter (outside the transformer construction).  Is that also accurate?

Thanks,

Joe

Hello Joe,

Uli is looking at your questions not related to the LLC.

I looked at the PMP40580 and it looks like they are using the leakage inductance as the resonant inductor.  So you are correct.

The resonant capacitor is formed by C203 and C204.  The total resonant capacitance would be the sum of the 2.

The total resonant capacitance is 30 nF.

I don't have a data sheet for the transformer.  However, the transformer is listed as having 400 uH of magnetizing current on the BOM.  The leakage inductance of most transformer is 3%.  In this case I would estimate the leakage inductance to be 12 uH.  This would put the resonant frequency at  265 kHz.

I look at the test data and it look like the design is operating at 250 kHz when the design is at resonance.  This is pretty close on what I estimated. 

A lot of design do use the leakage inductance for the resonant capacitor.  It is also not uncommon to use two capacitors as the resonant capacitors connected across the bulk capacitor.   By doing this you will have half the input voltage at the resonant inductance and resonant capacitors nodes.  If you use just one resonant capacitor before switching starts this node will start out at 0 V until switching beings.  It would take a few cycles for the node between the resonant inductor and resonant capacitor to settle at half the input voltage.  When you uses the two capacitor technique for the resonant capacitor.  The resonant voltage does not swing below ground during startup.

Regards,