• State TI Thinks Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 1 answer
  • Subscribers 63 subscribers
  • Views 410 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.

UCC1804: UCC1804J IC damaged and pulling excessive current at start up prior to reaching UVLO start up

John Sullivan2
Prodigy 40 points
Part Number: UCC1804

This device has been used in our power supply for a number of years.  We have had up to 100 of these ICs fail through those years.  The fundamental problem is generally the converter will not start.  Upon investigation we typically find the UCC1804J is pulling excessive current on the Vcc pin.  The latest failure I was able to apply 10 Volts to just the IC from Vcc to GND pins of the device and verified it is pulling 10m Amps.  As it should be less than 0.1m Amps I must assume the device is damaged.

The IC is being used as part of a bootstrapped 12 Volt power supply that runs from a 28 Volt input.  We start the chip from a current limited transistor off the 28 Volt input.  There is a 1000 Ohm resistor in series with this transistor.  Within about 10mSec after power is applied the chip is then run from an isolated winding from a transformer that is regulated to 11 Volts by the IC.  The transistor is shut off and the supply will continue to run from the 11 Volt winding.  There is a 15uF electrolytic and 0.47uF ceramic capacitor directly across the Vcc and GND pin as close as possible to the IC.

I would like to know what could be damaging the IC.  Another detail is even the damaged IC will start and run normally if given enough voltage to start.  So it appears the only thing damaged on the IC is possibly come internal zener clamp from the Vcc to GND pin.

I would like to know how this could be getting damaged?  It appears all the best design practices have been followed here.  My only concern was that our current limiting resistor at start up will allow about 15mA of current to the chip. But for the short duration it is applied it seems to be well within the data sheet limits for the IC.  I am searching for other answers here.  Or confirmation of what could be causing this?

Below is a scope plot of the Vcc pin at start up.

Below is a excerpt from the schematic it is being used in

  • 0
    TI__Expert 6080 points

    Hello John,

    This is strange behavior because if the start threshold is below 12.5V the controller should require <0.1mA of startup current. 

    Please reattach the images because they did not come through the first time for me to understand the situation better.

    Regards,

    Sonal

  • 0 in reply to Sonal Singh
    Prodigy 40 points

    I agree the device is supposed to pull less than 0.1mA at start up. So the IC has been damaged in some manner.  I was mostly seeking what internal to the IC could have been damaged in this manner and how based on the way we are using it? 

    I will try to re-attach the images, but it would not let me do that earlier.  It would let me attach text but not images.  I will try again.  It still will not let me insert the images.  I tried attaching a file instead.  Not sure if that will work any better.

    5482.Doc1.docx

  • 0 in reply to John Sullivan2
    TI__Expert 6080 points

    Hello John,

    For experimentation purposes do you think we can power the controller externally by a DC source with regulated input current? If the current through Vcc is the issue then we can say that the controller is damaged and I would suggest to replace the unit and redo the same test. 

    It is difficult to pinpoint as to what can be damaged in the IC internally, but a quick way to check is: if any of the pins have a short either with each other or to the GND.

    Regards,

    Sonal

  • 0 in reply to Sonal Singh
    Prodigy 40 points

    While the device was still mounted on the board.  If the voltage was increased through the current limited source until it reached the UVLO start point for the IC.  The chip would start normally and the supply would operate as if nothing were wrong.  The main problem was the supply would not start at low line because the 1804 IC was pulling to much current before reaching the UVLO start point.  This problem first showed up while the unit was turned on during a cold cycle of -40C.

    After the failure was confirmed on the board the chip was removed from the board.  I verified by simply placing 10 Volts through a current limited supply directly at the Vcc pin to GND of the IC that it was actually pulling 10mA when it should have pulling only 0.2mA max.  Indicating the chip is damaged.  In fact the current went up rather linearly at almost 1mA/V or 1000 Ohm impedance.

    We have had 100 incidents of this IC failing over the last few years in a similar manner.  I am just trying to identify what exactly can cause this kind of failure inside the IC. 

  • 0 in reply to John Sullivan2
    TI__Expert 6080 points

    Hello John,

    I understand it now, I would require some more input from you in order to reproduce this failure in our labs.

    1. How many units have you tested and what percentage of those units were faulty?
    2. Does this issue is board dependent or can you reproduce the same issue when the controller is run open loop?
    3. This excess current happens at -40degC: can you also share the line and load condition during this issue?

    Regards,

    Sonal

  • 0 in reply to Sonal Singh
    Prodigy 40 points

    We have had over 100 of these devices fail both during our screening process and at our customer's site.  The last few first showed the problem during the -40C cold test.  We have shipped several thousand of these units through the years.  So it is a low percentage that fails.  And somewhat random to where I can not reproduce it.  All I was trying to find out is what inside the IC could be damaged to cause this failure so I would know what to look for.  The line and load is cycled from low to high line throughout the testing process.  However that should not effect the voltage the IC sees by very much.  At -40C we are typically just running nominal line.  I am thinking the problem shows up first at -40C because the leakage on the Vcc pin is somehow at it's highest then.  So the IC could be damaged earlier in the test cycle and just show up at -40C.

    Again I was really just wanting to know what inside the IC might be damaged and what it would take to cause that kind of damage so I would know better what to look for on our unit to determine the cause of damage.  We had sent some of the damaged ICs out for evaluation but have not heard anything back yet on that.

  • 0 in reply to John Sullivan2
    TI__Expert 6080 points

    Hello John,

    Apologize for the delay, the Vcc pin has an absolute maximum rating of 12V. From the waveform, the pin is already voltage stressed. Also, the datasheet suggests using a current-limiting resistor to power Vcc. Absolute maximum of 12 V applies when VCC is driven from a low impedance source, such that ICC does not exceed 30 mA. This might be the reason why the pin gets overstressed and the controller fails during lower temp conditions.

    If you are really concerned about these failures, I would recommend following the datasheet and using the small resistor to feed the power to the controller.

    But without looking at the failure analysis there is no way to tell for sure what the issue it.

    Regards,

    Sonal