This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

  • TI Thinks Resolved

TPS92310: Ripple output at the output..

Part Number: TPS92310

Hi,

Please find the attached Isolated 7Watts CC circuit. I have a problem at output voltage which is having 3Volts of ripple out put. Tried increasing the Cout(to 470Uf + 220Uf) but still ripple(50Hz , 3Vpp) been seen. 

Rgds

ChandraACTODC_CONV.pdf

  • In reply to Chandrashekhar V9:

    Hi Irwin,

    Attached the xls with calculation for our design, one more observation is from xls we get,

    ZCD pin resistor R10(In our case R65) 643 kΩ(In EVM and our case 66.5K)

    ZCD pin resistor R12(In our case R58) 13 kΩ(In EVM and our case 11K)

    TP9S92310_Calculation_tools__REV1 3a.zip

    Rgds

    Chandra

  • In reply to Chandrashekhar V9:

    Hi Irwin,

    In one of the board we changed the value of R65 to 499 and R58 to 30K and added RCD snubber(R=SMF3W150KJ, C=1812B102K202CT, D=1N0007), now we are able to see constant voltage and no hiccup of OVP , also we see expected current output for 7 series LEDs.

    But now big problem started, We see U15(TPS92310DGS), R55(3 ohm resistor), Q1(STD3NK80ZT4) getting damaged in some of the boards even with the some of working boards also. We are sure rework is good.

    PS: we observer this type of behavior(U15, R55, Q1 getting damaged) even without snubber circuit when we removed R50(Current limiter) and added 0.5 Amps fuse in that place. We removed R50 and tried with fuse because this resistor was frequently getting damaged.

    Rgds

    Chandra

  • In reply to Chandrashekhar V9:

    Hi Chandra,

    The ZCD resistors will not work with the new values you have, 499 ohm and 30 Kohm set the ZCD at 0.36V using a 22 volt output. the transformer you are using has the same turns ratio for the secondary and primary bias so the primary bias will be close to the output voltage. Also, what are your resistor values in series with the LEDs, I saw 50 ohms and someone stated they were changed to 1 ohm? If they are still 50 ohms the primary bias voltage will get too high and possibly damage the LM3424 (I would expect R57 to open before that). Can you measure the output voltage?

    The diode you chose for the snubber needs to be much faster response than the 1N4007. A 1N4007 is a rectifier diode for 50 or 60 Hz. The snubber clamping frequency is above 50 KHz. It's good you chose a 1000V diode but its recovery will be too slow.

    Also, a 30 Kohm will see Vout times transformer 'n' which is 7.615. If Vout plus Vfd = 22V, Vreflected will be 167V across the 30K resistor which is almost one watt (this depends on what Vout really is). You could replace the snubber R with a 200-350V transient supressor and the diode with a 1000V ultrafast recovery which would work (again depending on what Vout really is).

    Regards,
  • In reply to Irwin Nederbragt:

    Hi Irwin,

    Sorry it is 499K Ohm typo error.

    what are your resistor values in series with the LEDs, I saw 50 ohms and someone stated they were changed to 1 ohm? --> I already mentioned that 50 Ohm is replaced with Short.

    Also, We expect the Vout of 23V to 24V becuase 7*3.2V(LED VF) = 22.4 V

    Rgds
    Chandra

  • In reply to Chandrashekhar V9:

    Hi Chandra,

    499 Kohm and 30 Kohm is too close to having issues as well. The ZCD arming threshold is 1.24V (max spec at 1.3V), at 22.4V this voltage from the divider will be 1.27V. There may be issues with part tolerances or thermals as the LED stack voltage will drop with temperature.

    22.4V * 30K/30K+499K) = 1.27V, this is because the primary bias and secondary winding are the same turns ratio.

    Regards,
  • In reply to Chandrashekhar V9:

    Hi Irwin,

    Looks like we nailed down to the issue we had it. 

    1. Wrong calculation of two resistor of Vzcd, Proper values are between 63K~73K and 11K. Attached the sheet for our circuit.

    TP9S92310_Calculation_tools__REV1 3a_chandra.zip

    2. Missing snubber circuit also related to this high current and voltage spikes in primary side. Which is said optional in EVM.

    3. Mosfet miss leaded because chosen FET is not good compared to the one in EVM for SOA. Mosfet SOA and Miller region we need to validate with our circuit and layout. Also found mosfet we are using is not same as been used in EVM, Coss is almost 10 times more in our mosfet, Below is the SOA, for our mosfet not sure about the Id at SOA region for our mosfet because we are at 75Khz switching frequency.

    4. Layout improvements required to increase thickness of traces in certain area.

    Rgds 

    Chandra

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.