• State Resolved
  • Locked Locked
  • Replies 16 replies
  • Subscribers 32 subscribers
  • Views 977 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.

TPS2553: Fault operation and short protection

Hirotaka Matsumoto
Guru 27525 points
Part Number: TPS2553
Other Parts Discussed in Thread: TPS25221

Hi all

Would you mind if we ask TPS2553?

<Question1>
Please refer to the attachment file, and could you give us the advice?
2388.20180917_TPS2553.pdf

<Question2>
In case of short, we assume that the infinity current flows FET with  during 2us.
We think  the device doesn't damege, because this is not continious current.
Is our recognition correct?

Kind regards,

Hirotaka Matsumoto

  • 0
    TI__Expert 8980 points
    Hi, Hirotaka-San,

    For question 1,
    1) If Vin of TPS2553 is a DC-DC, Optimize the DCDC control loop, avoid low frequency ripple in TPS2553 Vin from 20kHz to 270kHz.

    2) Suggest placing 0.1uF and 100uF ceramic cap on TPS2553 Input pin, make sure the input voltage ripple < 80mV.

    3) The trace routing the RILIM resistor to the device must be as short as possible to reduce parasitic.

    4) Power pad must be directly connect to PCB ground plane.

    For question 2,
    Yes, you are correct. Part will not damage during this 2us, although the current in this period is very high(even up to 25A). The time is very short, so the heat in this period can't damaged the part, our part will shutdown the power switch and the current will decrease.

    Regards,
    Bob


    --------------------------------------------------------------------------------------------------------------------------------------
    Please verify the answer by clicking the green "RESOLVED MY ISSUE " button in case the reply solved your question.
  • 0 in reply to Bob Ma
    Guru 27525 points

    Bob san

    Thank you so much for your repy.

    We would like to confirm "For question 1".
    If we take in your suggestions (1) to (4), does Fault pin operate normally?
    (With OCP condition, will Fault pin become low?)

    Kind regards,

    Hirotaka Matsumoto

  • 0 in reply to Hirotaka Matsumoto
    TI__Expert 8980 points
    Hirotaka-San,

    If Vin ripple is big and the ripple frequency is low, this ripple will influence internal logic circuit and cause fault not assert at OC condition. If Vin ripple is small, fault will operate normally.

    Regards,
    Bob
  • 0 in reply to Bob Ma
    Guru 27525 points

    Bob san

    Thank you so much for your reply!
    OK, we encourage to our customer to check it!
    We will feedback you later.

    Kind regards,

    Hirotaka Matsumoto

  • 0 in reply to Hirotaka Matsumoto
    Guru 27525 points

    Bob san

    We confirmed the customer as follows;

    1) If Vin of TPS2553 is a DC-DC, Optimize the DCDC control loop, avoid low frequency ripple in TPS2553 Vin from 20kHz to 270kHz.
    ->There is no low frequency ripple in TPS2553 Vin from 20kHz to 270kHz.

    2) Suggest placing 0.1uF and 100uF ceramic cap on TPS2553 Input pin, make sure the input voltage ripple < 80mV.
    ->The customer board has 0.1uF and 100uF actually. The input voltage ripple is ,50mVpp.

    3) The trace routing the RILIM resistor to the device must be as short as possible to reduce parasitic.
    ->The distance between RILIM and IC is 2mm or 3mm.

    4) Power pad must be directly connect to PCB ground plane.
    ->Power pad of customer's board  is directly connected to PCB ground plane.

    However, in case of ILIM=56kohm, the Fault pin didn't turn to Low level.
    So, could you give us advice again?

    Kind regards,

    Hirotaka Matsumoto

  • 0 in reply to Bob Ma
    Guru 27525 points

    Bob san

    We add some information as follows;

    1. Input current
    ->There is no ripple as the attachment file.
    20180920_TPS2553.pdf

    2. Input voltage source
    ->The customer confirmed with stabilized power supply, the result was the same as we mentioned before.

    Kind regards,

    Hirotaka Matsumoto
    Kind regards,

    Hirotaka Matsumoto

  • 0 in reply to Hirotaka Matsumoto
    TI__Expert 8980 points
    Hi, Hirotaka-San,

    From waveform, the current is just 450mA, but the configured limit point is 500mA. So I think the current limit is not triggered.

    What load are you using for this test? Please use a power resistor to make sure the load current is over limited current. When part current is limited, power FET will work at saturation region, part will rise Rds to decrease load current. This way, Vout will drop. So to make sure current is limited, I suggest you to use a power resistor and measure Vout to judge if current is limited and if there's issue with load.

    For example, if your Vin is 5V, current limit point is 500mA. Choose a R, make 5V/R>500mA. If you choose 5ohm power resistor, the suppose load should be 1A, and the Vout should 0.5A*5ohm=2.5V.

    Regards,
    Bob
  • 0 in reply to Bob Ma
    Guru 27525 points

    Bob san

    Thank you for your reply.
    We checked using TPS2553EVM with the same condition.
    Load is the resistor and FET is used for switching operation.
    As the result, Fault pin became high both ILIM=500mA or ILIM=200mA.
    So, in case of TPS2553EVM, the opetation is correct.

    Therefore, we assume that there is  some problem about the customer's board. 

    Kind regards,

    Hirotaka Matsumoto

  • 0 in reply to Hirotaka Matsumoto
    Guru 27525 points

    Bob san

    Thank you for your support always.

    Now the customer checks using TPS2553EVM.
    We will feedback you it later.

    On the other hand, we would like to FAULT deglitch.
    In case of overcurrent, FAULT deglitch are Min=5ms, Typ=7.5ms and Max=10ms.
    About FAULT deglitch, does it mean continuious overcurrent keeps?  
    So, if there is no overcurrent condition an instant, does the state machine of FAULT deglitch return to initial condition?

    Kind regards,

    Hirotaka Matsumoto

  • 0 in reply to Hirotaka Matsumoto
    TI__Expert 8980 points

    Hi, Hirotaka-San,

    Our parts' over current response time is 2us, fault deglitch time is same as you said. So even though fault is not assert in deglitch time, current is still limit. If overcurrent state is not long enough to deglitch time, fault deglitch timer will initiate.

    Below is a typical OC waveform in datasheet.

    Regards,
    Bob

  • 0 in reply to Bob Ma
    Guru 27525 points

    Bob san

    Thank you so much for your reply!
    OK, we got it!

    Kind regards,

    Hirotaka Matsumoto

  • 0 in reply to Bob Ma
    Guru 27525 points

    Bob san

    We would like to confirm one point just in case.
    1) If Vin of TPS2553 is a DC-DC, Optimize the DCDC control loop, avoid low frequency ripple in TPS2553 Vin from 20kHz to 270kHz.
    ->In case of the noise with below 20kHz(except for 20kHz to 270kHz), can the device operate normally?


    Kind regards,

    Hirotaka Matsumoto

  • 0 in reply to Hirotaka Matsumoto
    TI__Expert 8980 points
    Hi, Hirotaka-San,

    If the noise is beyond 20KHz to 270KHz, the device can overcome higher ripple(Up to ~120mV at least, maybe higher).

    TPS2553 is kinda ripple sensitive. The next generation part of TPS2553 is TPS25221, this part has been released to market. TPS25221 has better filter internal, so it can sustain higher ripple and TPS25221 is cheaper. Just FYI.

    Regards,
    Bob
  • 0 in reply to Bob Ma
    Guru 27525 points

    Bob san

    Thank you so much for your reply!

    Our customer confirmed the ripple with 1Hz or 2Hz at Vin pin.
    In this case, does the device can overcome higher ripple(Up to ~120mV at least, maybe higher) also?

    Kind regards,

    Hirotaka Matsumoto

  • 0 in reply to Hirotaka Matsumoto
    TI__Expert 8980 points
    Hirotaka-San,

    Yes, in this case, the device can overcome higher ripple.

    Regards,
    Bob
  • 0 in reply to Bob Ma
    Guru 27525 points

    Bob san

    Thank you so much for your reply.
    OK, we got it.

    Kind regards,

    Hirotaka Matsumoto