UCC28950: some questions about controller

Hello Vladimir

1/  P40. The input current to the power stage has two major components. One is the output inductor current, reflected through the transformer turns ratio and the other is the transformer magnetizing current. The magnetizing current therefore acts as a slope compensation ramp because it adds to the current sense signal. This equation was generated to insure that the ramp due to the magnetizing current was not more than half the downslope of the output inductor current. The inductor current downslope happens during the Toff interval which is . The output inductor sees a frequency of 2*F_SW . D_TYP was chosen because it was central between D_MIN and D_MAX. It may be necessary to add a further slope compensation ramp using the RSUM resistor as explained in paragraph 7.3.11 of the datasheet SLUSA16D.

2/ The 0.5 factor for ILOUT is used because we want the slope of the magnetizing current to be less than 50% that of the output inductor slope.

3/ You can use the UCC28950 controller in a step up system - A single winding secondary makes sense because the voltage stresses on the secondary switches (SR or Diode) are half those in a centre tapped secondary. Since you are not using SRs, you should simply leave the OUTE and OUTF pins open circuit. It's not necessary, but you can also connect the DCM pin to VREF and this will prevent the OUTE and OUTF signals from being present at the pins - maybe less noise in the system.

Thanks for the link to the article. I'm afraid I can't really say why the authors of that paper used the different formula but there are some things you may like to consider.

The TI example assumes peak current mode control and slope compensation is necessary for stability at large duty cycles (>50%). the formula in the DS (eq 30) is based on making sure that the magnetizing current ramp is less than 50% that of the output inductor. The component values (600uH and 100uH) given in the article by Hassanzadeh et al gives a reflected output inductor current slope of about 4.2A per us and a magnetizing current slope of about 1.2A per us. - They would have to add some slope compensation to ensure stability. However the article doesn't make clear whether the system uses current mode control of voltage mode control. Voltage mode control is more commonly used in high power systems like the one they describe (50kW) because it is less sensitive to noise. If they are using Voltage Mode Control, then slope compensation is not necessary.

Their paper is a clear explanation of the loss mechanisms in this topology. I think that their calculation of copper loss is probably optimistic - it does not take account of skin and proximity effects which are difficult to quantify but which can make the ac resistance of the winding significantly higher than the dc resistance they have used. the article by our collegues in Infineon has some more detail on the differences in the energy needed to achieve ZVS on the Left (AP) leg versus that needed to switch the Right (PA) leg.

Finally, I'd suggest that you use the UCC28951 rather than the UCC28950. It's a 100% compatible, drop in replacement with better performance at wide duty cycles in current limit.

Regards

Colin