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.

TLV702: instability of the LDO with higher capacitance on the output

Expert 6280 points
Part Number: TLV702
Other Parts Discussed in Thread: MULTIPKGLDOEVM-823

Team,

When LDO is enabled (from +3.3V) and experiences input voltage drop (from 4.3V to ~2V), it sometimes goes into instability and remains in it.

LDO then has still 3.3V (no load) on output when disabled and has ripple voltage up to 0-3.3V when load of 5-100mA is applied. When it's enabled, instability seems to exist for 400-500ms and then disappears. The drop from 4.3V->2V happens when circuitry that has a lot of capacitance is being turned on. Also in this mode, LDO cannot be turned off.

What we found was instability starts when LDO has more than 12uF of capacitance on the output.

Who can help us to resolve this (would share schematic and layout offline).

Thanks
TI Customer

  • Hi Bart,

    In order for us to help debug, please provide a multichannel scopeshot showing Vin, Vout, Ven, and Iout. This will help us understand the relationship between the rails.

    From your description, you mention Vout is oscillating from 0 to 3.3 V when disabled. Is this the load or the LDO that is disabled? If the LDO is disabled, this implies that you have a leakage path onto the output from somewhere else in the application.

    The voltage drop on the input rail sounds like a load transient for your upstream supply caused by the inrush current of the circuitry with a lot of capacitance. Due to the drop in input voltage, the TLV702 will go into dropout and may even experience reverse current if Vout should exceed Vin during this time. Additional input capacitance local to the TLV702 should reduce the input supply's voltage droop.

    Very Respectfully,
    Ryan
  • Hi Ryan,

    Ryan Eslinger said:
    From your description, you mention Vout is oscillating from 0 to 3.3 V when disabled. Is this the load or the LDO that is disabled? If the LDO is disabled, this implies that you have a leakage path onto the output from somewhere else in the application.

    Our first thought was that this is leakage from other sources, but when we disconnect everything at LDO output, we can still see oscillation. From that we conclude, that this is LDO problem. This is LDO, that is disabled has some varying voltage on output and can not be disabled.

    Ryan Eslinger said:
    The voltage drop on the input rail sounds like a load transient for your upstream supply caused by the inrush current of the circuitry with a lot of capacitance. Due to the drop in input voltage, the TLV702 will go into dropout and may even experience reverse current if Vout should exceed Vin during this time. Additional input capacitance local to the TLV702 should reduce the input supply's voltage droop.

    *We have solved the drop problem by limiting capacitor inrush/charge current. This also helped with LDO oscillation, LDO is not going into that mode anymore, but we are afraid, that this might still happen in application, when external power supply is not stable (in old vehicles). So we also lowered output capacitance to 1uF (as oscillations appear at ~11uF). *

    Ryan Eslinger said:
    In order for us to help debug, please provide a multichannel scopeshot showing Vin, Vout, Ven, and Iout. This will help us understand the relationship between the rails.

    We can do such measurements, but i'm not sure how to measure output current. At attached picture is Vin (blue), Vout(yellow), InRushcurrent Enable(Orange) - (it's a bit spikey, because i was using a wire shorting it to ground) and LDO enable(green). After this moment LDO goes into oscillation and can not be disabled. Maybe this graph is enough? Output current can be measured, but we also saw this with no load at the output, so i doubt this would be of any assistance. After LDO is disabled, LDO still has voltage on output, but voltage is dependant upon load, we have seen higher voltages also, up to 3.7V.

  • Hi Bart,

    I am glad to hear that you solved the voltage drop on Vin. The scopeshot you provided in your latest post confirms, that when the orange signal turns on, the inrush current on that rail is causing the shared input rail (blue) to drop. The LDO (yellow) is therefore forced into dropout until Vin rises again.

    The LDO output appears to be properly regulated whenever you only have the LDO enabled (the orange signal is in the disabled state). When the orange signal is in the enabled state, the LDO output always appears to be biased higher than 3.3 V. Traditional LDOs do not sink current. As such, when the LDO output is biased high, the LDO will simply disable the pass FET until the load can pull the output voltage back into regulation.

    You mention that you can isolate the output of the LDO from the other circuitry. Are you able to provide a scopeshot like the one above with the LDO output isolated (R20 is unpopulated)? Also, please confirm that R1244 is unpopulated during your tests.

    Very Respectfully,
    Ryan
  • Hi Ryan,

    thank you for your feedback.

    • R1244 unpopulated as requested.

     Green - Ldo output, Yellow - LDO enable pin.

    Green ldo output, Yellow - LDO input voltage:

    2. R1244 and R20 unpopulated:

    Green - ldo output, Yellow - LDO enable signal. Ldo has 3.94V on output with no load, also when disabled. We have checked 2 devices.

    Looking forward for your feedback.

    Kind regards,
    TI Customer

  • Hi Ryan,

    do you have some updates/feedback for us?

    Thanks & regards,
    Bart

  • Hi Bart,

    The LDO output being higher than its regulation voltage indicates that either a higher voltage rail is leaking onto the LDO output rail or that the LDO itself is damaged. The most telling scopeshot is your last scopeshot where you have isolated the LDO output from all other signals. As such, it appears that the LDO is damaged. This could be confirmed further by removing the LDO from the application altogether and placing it on an EVM such as MULTIPKGLDOEVM-823.

    Very Respectfully,
    Ryan
  • Thanks Ryan.
    We are going to check it as you advised on the EVM.

    -What are further steps if we find that LDO is damaged?

    -What are the steps if it will work fine on evaluation board?
  • Hi Bart,

    If the LDO is not operating correctly on the EVM, then we need to determine if the application damaged the LDO or if the LDO was damaged before it was in the application. If it was damaged in the application, the most likely causes would be an over voltage, over current, or reverse current. These events will most frequently occur during the power up or power down of various rails in the application.

    If the LDO is still operating on the EVM, then we need to look for additional leakage paths that would bias the output voltage high.

    Very Respectfully,
    Ryan
  • Hi Bart,

    Since it has been a week without a reply, I assume that this thread is resolved. If this is not correct, please feel free to post again.

    Very Respectfully,
    Ryan