Thermal ISSUE in TPS92314

Hi Anandkumar,

I have read your question and looked at the schematic included. I have a couple of questions. The schematic shows the output voltage to be 48V, but in your question you post the output to be 78 Vs. Is you flyback a current source? Also you show the input voltage spread to be 90 V to 300 Vac. The standard for US and European is 90 to 264 Vmax. Why is your maximum ac input voltage so high? Also the snubber shown in  the schematic does not use an RCD clamp, it has a zener diode in series with a resistor. Is this an error?

A few more comments. You say the MOSFET reaches 80 C in 30 to 40 minutes. Does it stay at 80 C? In other words is it still rising and you shut the converter off before is burns up? If 80 C is the steady state temp on the MOSFET, that is acceptable. Most MOSFETs have a maximum Tj of 125 C and  some 150 C and they have a very low theta jc. 

If you are using an RCD clamp and the MOSFET is getting too hot and the snubber resistor is getting hot you need to re-design the snubber, choose a MOSFET with a high VDS breakdown voltage, or reduce the leakage inductance in your transformer. First you need to measure the leakage inductance. From this you can determine the energy that the Rsnub needs to dissipate. In addition you have to determine the Vsnub. Vsnub is the reflected secondary voltage and the leakage voltage spike. You have to select a maximum clamp voltage for the leakage spike. It's usually about 50V for a US input voltage. So for your application it is probably 100 Vs -that's pretty large. The lower the voltage value of Vsnub, the more power you have to dissipate in your Rsnub. If you make Vsnub larger, you need to have a high VDSon breakdown voltage on your MOSFET. So that is one trade you can make. The problem there is if you increase the VDS breakdown voltage you also increase the MOSFET RDSon and you increase conduction losses. 

The best solution is to reduce the leakage inductance of your transformer. Make sure you use multiple interleaves of primary and secondary windings. This will greatly reduce the energy your RSNUB needs to absorb.

-Chuck

Dear Chuck,

Thank for the reply,

I kept the driver for aging continuously for 3days with normal 230Vac. third day afternoon i found the driver is burn. no component fails only driver IC is not responding. change a new driver then it start working.

I have few observation, while doing design calculation we consider 65Khz as frequency to select transformer primary inductance @ min 80Vac. but when IC is working with 80Vac, the gate pulse is 176Khz. Mosfet on time is 55ns and Toff is 85ns. our pulse period is in micro second. (Is this will be a problem??)

While attaching the schematic, I missed the correct one. I am attaching schematic for 2nd time. please re-check.

I am attaching few waveform at ZCD, ISNS, GATE, COMP,DLY, VCC.

please check for any abnormal.

Temperature rise is slow and remain constant at around 80degree.celcius.

Regards

Anand

6266.Reading.zip