• State TI Thinks Resolved
  • Locked Locked
  • Replies 11 replies
  • Answers 1 answer
  • Subscribers 64 subscribers
  • Views 818 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

LM74610-Q1: Timing of Vcap does not match the expected value

Inigo Monge
Prodigy 40 points
Part Number: LM74610-Q1
Other Parts Discussed in Thread: LM74610-SQEVM

Hello,

The circuit we are designing is the input to a DCDC module. We have used the LM74610-Q1 in the return path to protect from reverse polarity. As of now the circuit is as it is shown in the schematic picture attached.

When checking the voltage drop across the Vcap capacitor the timing does not match the behavior described in the datasheet. The datasheet says that the fall time should be 5.7 seconds and the rise time should be 0.1175 seconds when using a 4.7uF capacitor. As it can be seen in the picture we get 1.22 seconds fall time and 0.034s rise time. Giving a 97.3% duty cycle instead of the 98% that the datasheet says.

Ambient temperature is 25C and the IC gets around 37C after running 10 minutes with 20A through the transistor (the model is IAUT240N08S5N019ATMA1).

In one of the boards we discovered that the timing was 500ms fall time and 40ms rise time, giving a 92.3% duty cycle. This was solved by changing the LM74610-Q1.

What could be the cause of the wrong timing? What could have caused that the IC broke?

Thanks in advance

Best regards, Inigo

Photos LM.zip

  • 0
    TI__Genius 16305 points
    Hi
    Let me look into this and get back soon.

    Regards,
    Kari.
  • 0 in reply to Karikalan Selvaraj
    Prodigy 40 points

    Hi,

    We have tested a bit more and here are some new things that we have discovered.

    We tested with 28V and 18A going into the DCDC for around 1 hour. The triangle wave became faster and faster and after 1 hour running time it was 400ms faster. The duty cycle was more or less constant the whole time. The temperature of the capacitor and the IC was around 33C maximum. The vertical cursors that appear in the picture mark the inital waveform limits.

    After that we tried switching on and off the Load connected to the DCDC module. The timing of the waveform varies with (18A) or without load (the DCDC draws around 100mA with no load) and it is around 100ms faster with load.

    We have also tried to make both the IC and the capacitor warmer without load. Setup is shown in the photos. In the waveform it can clearly be seen that as soon as the heatgun was switched ON the capacitor voltage behaves unexpectedly. Sometimes it stayed constant and other times it went up. As soon as the heatgun was turned off the waveform quickly became normal. 

    In the waveform "Heat applied" the heatgun was turned on right after the vertical cursor to the right and turned off right when the flat line starts falling to "normal".

    The temperature of both the IC and capacitor did not exceed 40-45 degrees. If heat was constantly applied then the capacitor voltage became constant and then it needed some time to cool down before returning to "normal" operation.

    I hope this new information helps.Load and Heat test.zip

    Best regards,

    Inigo

  • 0 in reply to Inigo Monge
    Prodigy 40 points

    Hi,

    We purchased the evaluation module LM74610-SQEVM and ran the same tests:

     http://www.ti.com/tool/LM74610-SQEVM

    The same behaviour that we had in our PCB could be seen in the evaluation module as well.

    Best regards,

    Inigo

  • 0 in reply to Karikalan Selvaraj
    Prodigy 40 points

    Hi Kari

    Im the project manager for the project where Inigo has this problem.

    Do you have any update?

    BR

    David

  • 0 in reply to Inigo Monge
    TI__Genius 16305 points
    Hi

    Regarding the first question, Duty cylce is ~98% and it can vary part to part around 98% mark. But turn on/off timing should match the datasheet values unless the IC is damaged. Possibility is that the IC could be damaged.

    Regarding the second question, when the load current is about 20A, MOSFET can be hot due to the turn on/off of the MOSFET by LM747610-Q1.
    At 20A, MOSFET will dissipate 20A*20A*2mohm *0.98 (duty cycle) = 0.784W. When the MOSFET is off, power dissipation = 20A*0.9V*(1-0.98) = 0.36W.
    Total average power dissipation is 0.784+0.36 = 1.144 W. This can increase the temperature around MOSFET and increase case temperature of LM74610-Q1. Is it possible to measure the MOSFET temperature?

    Also, are you using a capacitor having high temperature variation?
    If temperature increases, the will vary the value capacitor and also the IC's leakage current will increase (during turn ON duration) and will reduce falling time of waveform. Capacitor variation will also add to this resulting reduced fall time and rise time.
    However duty cycle will remain same.


    Regards,
    Kari.
  • 0 in reply to Karikalan Selvaraj
    Prodigy 40 points

    Hi,

    Thank you for your reply.

    Regarding the first answer hinting that the IC could be damaged. This has been tested in 20+ parts in our boards and 3-4 parts in the reference board LM74610-SQEVM. If they were damaged, why were they damaged? Even without current the capacitor shows wrong timings. Could I get pictures of the setup and waveforms of a functioning IC from you?

    When running 20A and using a thermal imaging camera model Testo 875-1i the temperature of the IC, capacitor and surroundings gets around 40C. MOSFET could warm up to 50C. The temperatures are all withing the recommended temperatures in the datasheets.

    We are using a X7R capacitor as suggested in the datasheet.

    Best regards,

    Inigo

  • 0 in reply to Inigo Monge
    TI__Genius 16305 points
    Hi

    Yes we can test this, but will need some time, towards end of this week.
    We will get back by Friday possibly.

    Regards,
    Kari.
  • 0 in reply to Karikalan Selvaraj
    Prodigy 40 points

    Hi Kari

    Any updates on this issue?

    BR

    David

  • 0 in reply to David Karlsson
    TI__Genius 16305 points
    Hi

    I am testing this on the EVM now and will get back soon tomorrow.

    Regards,
    Kari.
  • 0 in reply to Karikalan Selvaraj
    TI__Genius 16305 points

    TestWaveforms.zipHi

    Issue with the probing the VCAPH or VCAPL or Gate Pins. Probing causes additional leakage affecting the ramp down time.

    When it is ramping, ramp down time is decided by the leakage. Internal IC leakge is specified in the datasheet 0.95uA (I_discharge current).

    Now with the probe connected, this leakage increase and affects the cycle time.

    I tested on the LM74610-SQEVM and found that the output (cathode) is switching as per design when i probe only Anode and Cathode.

    Timing matches datasheet calculations.

    However, when i probe VCAPH or VCAPL or Gate, timing is reduced.

    Please refer to the attached waveforms.

    Regards,

    Kari.

  • 0 in reply to Karikalan Selvaraj
    TI__Genius 16305 points
    Hi

    When you have time look into this and let us know if this has resolved your issue.

    If your question is answered, click "this resolved my issue" button at the end.

    Regards,
    Kari.