UCC28740: The major difference between UCC28740 and UCC28742?

Hello Gabriel,

Thank you for your interest in the UCC28740 and UCC28742 flyback controllers.

Both devices are equally suited for the 24-V, 75-W application. One does not have a performance advantage over the other in this regard.

As you noted, the UCC28740 has a larger package and a high-voltage start-up pin (HV), whereas the UCC28742 is in a smaller package without HV start-up. The UCC28740 incorporates a high-voltage JFET to provide about 250 uA to charge up the VDD capacitance relatively quickly, and is switched off after start-up to save no-load standby power .  This additional JFET increases the size of the controller and so the larger SOIC-8 package is required.

On the other hand, the UCC28742 has no start-up FET and so relies on an external resistor string from the bulk voltage to charge the VDD capacitance.  This reduces cost and the allows a smaller SOT-23 package, however it requires a trade-off of start-up speed vs. additional stand-by power dissipation in the start-up resistors.  Their values are determined by the longest tolerable start-up time with a given VDD capacitance at the lowest input voltage.  Once this value is fixed, the power dissipation at high line is always present, even at no-load, hence the higher stand-by power with UCC28742. 

Regards,
Ulrich

Hi Ulrich

Thanks for your comments.

As customer only need no load power consumption less than 0.1W, it seems UCC28742 could also meet their requirement.

I suppose UCC28742 will be more cost effective choice? 

Few further questions for UCC28742:

• In fault condition that MOSFET can't turn-on. Will CVDD keep being charged and exceed the voltage tolenrence of VDD Pin? Will this pin be damaged in this condition?
• In application note " Minimize Standby Consumption for UCC287xx Family" it indicated that X-cap have impact to standby power consumption and could calculated by below formulate. I want to know what Rx mean and how to obtain this value?

Hello Gabriel,

Yes, the UCC28742 should be able to meet the 0.1W no-load stand-by limit and will be more cost effective.

In the case of a MOSFET fault where it cannot switch, there is no danger to the VDD or to CVDD. At start-up, IC bias current is about 1.5uA whereas the start-up resistor will provide maybe 15uA or more to charge CVDD.  When the VDD voltage reaches ~21V, the IC will turn on and try to drive the MOSFET into switching.  The IC bias current go up to about 2mA and this level is much higher than the few uA available through the VDD charging resistance.  If there is no switching, VDD will sink to the ~8V UVLO threshold and bias current will drop to 1.5uA again. Then the start-up resistor will charge VDD back up to ~21V, and the cycle will repeat indefinitely as long as the MOSFET is unable to switch.  There will be no overvoltage stress on the VDD net.

In the "Pdis" formula above, Rx is the "bleeder" resistance across the X-cap, which is needed to discharge any voltage across the cap within 1 or 2 seconds (depending on which safety standard is being used).  The value of the X-cap is chosen based on how much capacitance is needed to help reduce differential mode conducted EMI noise.  Once this value is known, the Rx value is chosen to meet the X-cap discharge time based on the RC time constant required.  Therefore, Rx = tau/Cx, where tau = discharge time constant and Cx = total value of all X-caps.

Note: the higher the X-cap value, the better the EMI performance, BUT the lower the Rx value that will be needed to discharge it within the time limit allowed.  Lower values of Rx result in higher stand-by dissipation at high line (230Vac), so Cx should be as low as possible to do the job.

Total stand-by power = ( no-load input power of UCC28742 converter + dissipation of Rx ), must be < 0.1W.

Regards,
Ulrich