• State
  • Locked Locked
  • Replies 2 replies
  • Subscribers 63 subscribers
  • Views 927 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

LM22674MR-ADJ Step-Down Voltage Regulator Power Up (Transient) Problem

Justan Steichen
Prodigy 20 points

Problem:   On +24V power up, an LM22674MR-ADJ Step-Down Voltage Regulator internally shorts itself to ground.  The fault does not happen on every printed circuit board.

Circuit Description & information:  I have utilized two LM22674MR-ADJ on one printed circuit board which produce separate DC outputs - one for 5V and one for 7V.  (Please see attached for schematic and board layout).  The PWRIN ranges from 8V to 40V.  Both regulators really only need to output upto 50 mA, although these are sized for 500mA (for future use).  The typical resistances are as follows:  PWRIN to ground = >222K, +5V to ground = >5Kohm, and +7V to ground > 6Kohm.

I initially used the WEBENCH tool to design these circuits and have built the circuit with the recommended BOM and layout.  On the second page of the attachment, you can see an example of the damage to one of the LM22674s that is caused during the power up.  It should be noted, that the circuit works at 24V steady state.  Also, the problem idoes not consistently occur on one circuit or the other... sometimes the +5V regulator circuit blows, and sometimes it is the +7V regulator circuit.

Initially I tought the problem was related to the power supply providing 24V, but different supplies proved me wrong.

If anyone can help, please do!  I am stuck.5543.Power Supply.pdf

  • 0
    TI__Mastermind 18940 points

    Hello Justan, 

    Do you have any oscilloscope waveforms of the input voltage across C8 and C13 as the regulators are starting up?

    Try to add some bulk electrolytic input capacitance (i.e. 10uF or more) and see if the problem goes away. Make sure the capacitor is rated for the input voltage. Ceramic capacitors de-rate their capacitance under DC bias. With 24V applied, these 50V 1uF input capacitors may be at half of their value. 

    Also, I would suggest using X5R or X7R input capacitors instead of Y5V. The Y5V types can have very wide capacitance variation under temperature. 

    In terms of the board layout, I would suggest rotating the input capacitors (C7 C8) and (C11 C13) by 180 degrees so that the GND terminal of the input capacitor is tied to the anode of the schottky diode. This is a high di/dt path and should be as short as possible to mitigate voltage spikes. Also, I would suggest rotating the bottom feedback resistors (R17 and R20) by 180 degrees so that they are tied closer to the GND pin of the IC. 

    I hope this helps. 

    Best Regards, 

    Denislav

  • 0 in reply to Denislav Petkov
    Prodigy 20 points

    Denislav Petkov said:

    Hello Justan, 

    Do you have any oscilloscope waveforms of the input voltage across C8 and C13 as the regulators are starting up?

    Try to add some bulk electrolytic input capacitance (i.e. 10uF or more) and see if the problem goes away. Make sure the capacitor is rated for the input voltage. Ceramic capacitors de-rate their capacitance under DC bias. With 24V applied, these 50V 1uF input capacitors may be at half of their value. 

    Also, I would suggest using X5R or X7R input capacitors instead of Y5V. The Y5V types can have very wide capacitance variation under temperature. 

    In terms of the board layout, I would suggest rotating the input capacitors (C7 C8) and (C11 C13) by 180 degrees so that the GND terminal of the input capacitor is tied to the anode of the schottky diode. This is a high di/dt path and should be as short as possible to mitigate voltage spikes. Also, I would suggest rotating the bottom feedback resistors (R17 and R20) by 180 degrees so that they are tied closer to the GND pin of the IC. 

    I hope this helps. 

    Best Regards, 

    Denislav

    Thank you for your feedback.  Attached I have included a oscope waveform.  C1 (yellow) is VIN on U7.  C2 (red) is the 5V output  C3 (blue) is the 7V output.  This waveform was taken when VIN only went to 12V and I do not have readily available a 24V VIN oscope plot.

    Since Friday, I have made a couple of component changes to my circuit.  I replaced R15 with a 2.49Kohm resistor and replaced R16 with a 1.65kohm.  This sets up a voltage divider on the enable pin of the 7V regulator circuit.  The goal of this change was to ensure that the 5V rail is up and stable before the the 7V supply starts.  However, this did not work!  Indeed I measured ~2.1V on the EN pin of the U7, however the circuit output nothing.  Not sure why.  Did I need to have 1.9V + 0.6V hysteresis?  So for right now, I have removed both R15 and R16.  (Now the 7V circuit will power up without any EN control.) (This may not be such a bad thing in my application.)

    The next change I made is to replace the D10 schottky with a resettable fuse(mfn pn: MICROSMD035F-2).  The goal here is to limit the current as the regulator are powering up.  The original purpose of the schottkey diode was to protect the regulators in the event of a mis-wiring on PWRIN and Ground.  I suspect the resettabel fus will also help protect against this scenario - but I do need to test.   (FYI:  MICROSMD035F-2 is what I selected because it is what I coveniently had in stock - no other reason!)

    The next change is to bump up the capactiance C7 and C11 to 4.7uF each.  This increases the bulk capacitance and may help with the high current transient startup of the load. 

    These component changes are being made of more boards.  A sample size of one worked out ok but I need more to verify.

    I will most likely need to change the artwork (another collegaue recommended putting in a higher rated SW schottkey diode) and I will consider your suggestions.   Thank you again for your feedback.

    Kind Regards,

    Justan