UCC25600 design questions

Q3 and Q4 would need an external Syncronous rectification driver, of course you can use schootkys. With Sync. rect. you may boost up efficiency on the price of complexitiy.

You must sure made some mistakes in the spreadsheat. Could you share your design parameters?

Thanks for your reply.  The parameters for the spreadsheet are a bit different than the values used in the Power Stage Designer Tool... but they are as follows:

For the spreadsheet:

DCinMin: 240V

DCinMax: 405V

Switch freq: 200khz

Max. switch freq: 350khz

Min switch freq: 85khz

Max. Power Limit: 1100W

Max. Output Power: 1000W

Full load eff: .92

Output Volt: 170V

Ratio Lm/Lr: 8.8

Qr:0.5

Everything else left at default values.

In Power Stage Designer Tool:

VinMin: 240V

VinMax: 405V

Vout: 170V

Output Current: 5.9A

Switch freq.: 200khz

Diode Voltage drop: 1V

Inductor Ripple Current: 10%

Max. Duty: 90%

Mag Current: 5%

Choose Turns ratio: 0.6:1

Choose Transformer inductance: 550uH

Choose Inductance: 320uH

I checked and the Power Stage Designer Tool calculates values for a Hard-switched Half-bridge converter, not for an LLC resonant one.

The parameters you designed cannot be realized with an LLC resont Half-bridge converter. To have such a wide input range, you would need an Lm/Lr ration of 3-5 to maintain output voltage stabilization. A normal LLC converter can tolerate about +-10% input voltage change.

I had originally thought that the Power Stage Designer Tool was intended for a non-resonant half bridge, but when you click the "info" button, it brings up a link to the UCC25600, which is a resonant controller!

OK - so I don't have any experience with resonant circuits - I've previously read through some of the app. notes for the UCC25600, but I'm missing a few things... does the resonant circuit require the capacitors across Q1, Q2 like a hard switched half bridge?  Also, when I adjust LmLr "m" value, to 3, 4, 5 or any value, and get the corresponding Q figure, I still can't get the spreadsheet to calculate the current sensing/protection figures...

The calculator spreadsheet will not determine the exact curves, but there are some example curves, you can use.

Although in my opinion the XLS has some errors:

- for example you determine m value, which is Lm/Lr, but if you see Lm and Lr, their ratios are not m (there is a different version of the XLS in an other thread, but that has other erros)

- it falsely detemines turn ratio: turn ratio should be chosen as n=Vinmax/(2*Vout)

- it falsely determines impedancia seen by the primary network from the secondary, because of falsely determined turn ratio

The design procedure should be the following (i did it in the following way for my designs):

1. determine the turn ratio. If you plan not using burst-mode then choose a slightly higher value, so the IC could maintain regulation at low loads (+0,1-1% should be added), if you plan to use burst-mode, then you can use a slightly lower values, to increase idle power. Let this new "n" be used later.

2. Determine the gain needed to maintain regulaton at minimum input voltage: Mmax=2*n*Vout/Vinmin

3. Choose m value, if you choose it too high, the switching frequency to maintion regulation at low input votlage and high load can be really low, and can cause high B in the core. If you are using burst mode and use a slightly lower n, then you can bring "m" up to 10, otherwise stay below 6.

4. The needed quality factor of the LLC network should be the function of the previous values: Qr=1/m*sqrt((1+m*(1-1/Mmax2))/(Mmax2-1).

5. Determine minimum switching frequency! Since it is not an input parameter, but the function of other input parameters. Usually this operation point is where core losses are at maximum, so should be looked at carefully. (nominal swithcing frequency equals to the resonant network higher resonant frequency, fswnom=fr) fmin=fswnom/sqrt(1+m*(1-1=Mmax2))

6. Determine resonant network: Rac=8*n2*Rload/Pi2,Lr=Qr*Rac/(2*Pi*fr), Cr=1/(2*Pi*fr*Qr*Rac), Lm=Lr*m

If you found my answers helpful and mathematically correct, please respond.

Laszlo - thank you very much for your help.  It seems to be correct what you are saying, but to tell the truth, I am not yet at the level of experience with resonant systems where I can 100% verify your math.  I will be looking at this in much more detail in the next week or two and will hopefully have a much better understanding by then.

For your information, I used an "IR2110" driver (HS/LS NMOS driver) for the MOSFETs. It runs sweet! 

I managed to get a maximum efficiency of 97% at 515W output at +/- 83V from 400V input from a PFC (UCC28061) :-)