• State Resolved
  • Locked Locked
  • Replies 17 replies
  • Answers 1 answer
  • Subscribers 64 subscribers
  • Views 2528 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.

LM43603EVM / Short-circuit test

Tsukui Yusuke
Guru 21930 points
Other Parts Discussed in Thread: LM43603

Hi,

Could you tell me about EN pin status of LM43603EVM in short-circuit test?
EN pin of the status is Low in the short and over current test.
Will you confirm an attached file?

LM43603EVM.pdf

I want to know the cause for this behavior.

Best Regards,
Yusuke / Japan Disty

  • 0
    TI__Mastermind 19891 points
    Hello Yusuke-san,

    We can check this on the EVM and confirm.
    Did you make any modifications to the EVM (i.e. BOM changes)?

    Regards,
    Denislav
  • 0
    TI__Genius 11085 points

    Hi, Can you attach the EN, VIN and VOUT waveforms during an over current (OC) or short circuit (SC) test please?

    The LM43603 does not have condition to pull low EN pin internally. On the EVM, the EN pin is connected to VIN through a resistor divider pair. Please monitor the VIN voltage during the OC and SC tests. If the current limit of VIN supply is not high enough for OC and SC tests, VIN will droop low and pull low EN at the same time.

    The LM43603 has hiccup OC protection. In OC or SC conditions, the Vout will shutdown and restart after about 6ms. It is not because of EN status, but the protection mode itself.

    Regards,

    -Yang

     

  • 0 in reply to YangZhang
    Guru 21930 points

    Hi

    Thank you for your information.
    EVM does not change setting.
    And, I check for Vin power in SC test.

    I attach a waveform that added the VCC and VBIAS below.
    Will you confirm an attached file?
    LM43603EVM ver2　20150609.pdf

    I think that there is suspicious switching of Bias mode.
    When BIAS mode is switched, EN will be Low?
    Could you give me your advice?

    Best Regards,
    Yusuke / Japan Disty

  • 0 in reply to Tsukui Yusuke
    TI__Genius 11085 points

    Hi,

    From the scope shots, we can see that the output voltage rings to about -2V during the SC test. When Vout is tied to Bias, Bias voltage is pulled below zero a lot as well. When Bias is pulled below zero, substract parasitic diode will be turned to conduct current and cause all kinds of wierd behavior. The key is to prevent Bias voltage being pulled to negative voltage < -0.5V.

    A few ways to do that:

    First test can be done is to tie Bias to ground, instead of Vout, and try the SC test again. EN behavior should be correct.

    It would be the best to reduce the Vout ringing magnitude duing SC test. Please reduce the wire length from the output caps to the short circuit point. Long wiring adds parasitic inductance to the short circuit current transition path, causing Vout to ring a lot.

    Bias tied to ground is not ideal for efficiency. It is still better to tie Bias to Vout. A series resistor can be added between Vout and Bias to create a low pass filter with Cbias. This will reduce the negative voltage magnitude seen by Bias. But it might still be there if the ringing is too strong. Or a small Schottky diode can be added between Vout to Ground, to clamp the negative voltage on Vout. The Schottky can be very small, since it only conduct during harsh load transient or SC for a short period of time.

    Hope these helps.

    Best,

    Yang

  • 0 in reply to YangZhang
    Guru 21930 points

    Hi Yang,

    Thank you for your support.
    I will contact the customer to process as not a negative voltage.

    Let me ask you one more thing.
    Customer want to know the cause of EN Low when Vbias=negative voltage.
    Please tell me why the EN is Low when Vbias is a negative voltage.

    Best Regards,
    Yusuke / Japan Disty

  • 0 in reply to Tsukui Yusuke
    Guru 21930 points
    Hi Team,
     
    Can I get the information ?
    If you need more information or if  I should  clarify my comment ,
    please let me know.
     
    Best Regards,
    Yusuke　/ Japan Disty
  • 0 in reply to Tsukui Yusuke
    TI__Genius 11085 points
    We will talk to the designers to get more in depth understanding of what was happening. -Yang
  • 0 in reply to YangZhang
    Guru 21930 points

    Hi Yang,

    Thank you for your support.
    I'll be waiting for your ｉｎｆｏｒｍａｔｉｏｎ.
    If direct mail is required, please contact the following.
    My e-mail address is: tsukui-y@clv.macnica.co.jp

    Best Regards,
    Yusuke / Japan Disty

  • 0 in reply to Tsukui Yusuke
    TI__Genius 11085 points

    What happens when Bias is pulled negtive is that a parasitic diode between EN pin and Bias input is forward biased. The current going through the parasitic diode pulls down the EN pin low. Basicly, it is not a good idea to pull a power pin negative in general.  Is it possible with your customer to reduce the wiring inductance from the output of the IC to the short circuit location? Or can they add a schottky diode to clamp the negative voltage magnitude on Vout? I'd recommend the method with a small schottky if the short circuit location cannot be moved.

    -Yang

  • 0 in reply to YangZhang
    Guru 21930 points

    Hi Yang-san,

    Thank you for your support.
    I understand.And,I will advise to customer considerations that I got from you.
    Would you give us some details?
    LM43603 is broken by becoming the EN=Low?
    Because customer EN is connected to VIN, so customer worried.

    Best Regards,
    Yusuke / Japan Disty

  • 0 in reply to Tsukui Yusuke
    Guru 21930 points
    Hi Team,
     
    Can I get the information ?
    If you need more information or if  I should  clarify my comment ,
    please let me know.
    Customer have wanted this behavior to be a problem.
     
    Best Regards,
    Yusuke　/ Japan Disty
  • 0 in reply to Tsukui Yusuke
    TI__Expert 3870 points

    Hi Yusuke-san: is this measurement taken on the EVM? Any modifications to the EVM? How is the EN pin being driven?

    Thank you!

    -JP

  • 0 in reply to Juan Paulo Fung
    Guru 21930 points

    Hi Paulo-san,

    Thank you for your response.
    Customer taken on the default EVM.
    EN is connected to Vin.
    So, Customer feels that problem the phenomenon of EN = Low.

    Best Regards,
    Yusuke/Japan Disty

  • 0 in reply to Tsukui Yusuke
    TI__Genius 11085 points

    If they are worrying about EN tie to VIN directly, they can add a 10k pull up resistor to the EN pin. The KEY is not to allow the Bias pin being pulled to negative <-0.4V. Please advise your customer to protect the circuit from fault conditions like that.

    Thank you,

    Yang

  • 0 in reply to YangZhang
    Guru 21930 points

    Hi Yang-san,

    Thank you for your support.
    I will contact the information to the customer.
    Let me ask you one more thing.

    ・Is it not broken if I add 10k pull up resistor to the EN pin?
    ・Could you tell me if there is a recommended diode?

    Best Regards,
    Yusuke/Japan Disty

  • 0 in reply to Tsukui Yusuke
    Guru 21930 points
    Hi Team,
     
    Can I get the information ?
    If you need more information or if  I should  clarify my comment ,
    please let me know.
     
    Best Regards,
    Yusuke　/ Japan Disty
  • 0 in reply to Tsukui Yusuke
    TI__Expert 4365 points
    Hello Yusuke, I am also on the applications team for this series of products. Allow me to clarify several things.
    1) The recommended connection for the Enable Input is to have a voltage divider of two resistors from Vin with the midpoint connected to the Enable input. Use the equations in the data sheet to establish a system UVLO to be just below the system minimum Vin. For a 24V industrial rail this would probably be in the 16 or 17V range. The factory EVM will be set around 6 or 7V for a 5V out. This should be adjusted higher for the 24V to 5V case.

    2) The waveforms sent show a -2V output excursion when the short is applied. The negative excursion enters both the Switch node via the inductor but more directly through the Bias input which is connected to Vout in applications of Vout greater than 3.3V. If the customer end application truly has negative excursions where an inductive path causes the -2V negative surge then it would be advisable to place a 200 ohm resistor in series from Vout to bias input. This will prevent any damage to the bias input.

    Alan Martin