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.

TPS38700S-Q1: Interrupt behavior of the component

Part Number: TPS38700S-Q1
Other Parts Discussed in Thread: TPS38900X-Q1, , TPS389006-Q1, TPS389006, TPS38700

Tool/software:

Hello,

I am a developer using tps389006s-q1 with the sequencer. I observed that whenever I create a forced an interruption the sequencer does not behave and reset the power_sequence. 

The datasheet is not clear about the behavior of tps38700s when there is an interruption. Could you explain me how it should behave?

I thought that it would reset and shutdown the activated ENx.

Regards,

Tom LEVISSE

  • Hello Tom, 

    Can you please clarify what type of interruption is being forced. Is it an interruption on the ACT pin when a power up or power down sequence is occuring? Can you please provide a oscilloscope screen capture if possible, this will greatly help me understand the issue. 

    Regards, 

    Oscar Ambriz 

  • Hello, I am doing the software development, so I need to test everything. So on purpose during ACT 0 -> 1 transition I had disconnected one MONx pin to provoke an INT_UVLF and INT_SEQ_ON interruption.

    I am sorry, I don't have an oscilloscope with the screen capture option 

  • Hello Tom, 

    Is there a schematic you could share with me? I assume the sequencer is being used with its TPS38900x-Q1 pair based on the description provided?

    Regards, 

    Oscar Ambriz

  • Hello Oscar,

    Yes I am using the TPS38700s-q1 sequencer with the TPS389006-q1 voltage monitor.

    I have reproduced the schematic below as written in the datasheet. 

    In order to test the behavior of both ICs, I had disconnected one ENx pin to simulate a power supply fault. 

    But then, The tps38700s-q1 seems to be blocked in its last state before the interruption of the tps389006-q6. 

    Is there any solution to control the tps38700s-q1 after that and make the ENx pin down in case of interruption?

    Can you explain how it shall behave in case of an incorrect power sequence and interrupt signaled by the tps389006-q1?

    Regards,

    Tom LEVISSE

  • Hello Tom, 

    Thank you for the explanation, I will consult will my colleagues on this query. Hope to provide an update by 5/31. 

    Regards, 

    Oscar Ambriz 

  • Hello Tom, 

    Based on your description my understanding is that power up or down sequence was interrupted purposefully. This action was done by disconnecting an ENx from one of the converters, meaning an output voltage is never applied to the designated MON pin. When this happens TPS38700s-Q1 will pause its sequence because it has not received the command from TPS389006-q1 to continue. TPS389006-Q1 will send a pulse through the sync pin to TPS38700s-Q1 when it sees that the voltage at the MON pin in question has crossed the UVLF point, if the voltage was never applied then TPS389006-Q1 will not send the required SYNC pulse to TPS38700s-Q1 and the sequence is paused. 

    Is this the behavior you are seeing? 

    Regards, 

    Oscar Ambriz

  • Yes it is exactly the behavior I am observing. But then the NIRQ pin is pulled low by TPS389006-Q1 and the TPS38700s-q1 stay blocked, even if ACT goes from 1 to 0.

    Could you tell the behavior of the TPS38700s-q1 in case of fault during a power sequence ?

    regards,

    Tom LEVISSE

  • Hello Tom, 

    Understood, what is going on here is TPS389006-Q1 was expecting a voltage to be present at its MON pin since no voltage was provided NIRQ asserted. NIRQ will not de-assert unless a write one to clear is performed on the register in question and voltage is provided. If the voltage is still not connected then the NIRQ will again report the fault. 

    TPS38700S is not blocked due to the NIRQ signal being low. It is actually blocked because it never received a pulse on its sync pin form TPS389006. The reason TPS389006 did not provide a sync pulse is because the voltage at the MON was never connected. You can force the TPS38700S device to continue its sequence by pulsing the sync pin low externally is you wish to test functionality. 

    The intended use of TPS38700S is to stop the sequence when it does not receive the needed sync pulse from TPS389006. This is a safety measure to ensure the system stops enabling voltage sources when an error has been recorded. If you wish for the sequence to not be dependent on the sync pin then i would recommend the TPS38700 device (no sync pin), this device will complete the full sequence when ACT is enabled and not wait for the SYNC pin handshake. 

    Regards, 

    Oscar Ambriz 

  • Hello Oscar, 

    In case of a wrong Voltage read by TPS389006 (making an interruption) , the TPS38700s shoud desactivate the power supplies ? 

    Regards,

    Tom LEVISSE

  • Hello Tom, 

    It is not due to the NIRQ (interruption) asserting low, it is due to the voltage connected to the TPS389006 MON pin not crossing the UV threshold and causing a SYNC pulse to happen. TPS38700s will only enable a sequencing pin when it receives a pulse on the SYNC pin. Since the pulse never happened TPS38700s is in loop. An example of the SYNC operation is shown at the end of the datasheet as follows: 

  • Hello Oscar,

    Could you tell me a way to turn off ENx pin when an interruption is generated during the sequence ?

    Regards,

    Tom LEVISSE

  • Hello Tom, 

    I assume the sequence we are talking about is the power down sequence. One option would be to force a downward pulse on the SYNC pin of the device, this will signal to TPS38700s that the ENx pin is ready to power down. Again the downward pulse should have come from the TPS389006 device but since the voltage at the MON pin never crossed the UV threshold (never powered down) the pulse was never given. There is no software method for forcing the ENx pin down. The intended use of the TPS38700s device is that the user wants to stop the sequence when a sequencing point has misbehaved. Once the error has occurred it is up to the user to determine how they would want to place their system in a safestate. For example the user can use TPS389006 to determine where the sequencing error happened and performed a reset of the device in question to see if the error is resolved. 

    If you would like a device that is not dependent on the SYNC pin I recommend TPS38700 (no S) which will completely sequence down when the ACT pin is triggered low. 

    Regards, 

    Oscar Ambriz 

  • Hello, Oscar,

    I want to shut down all ENx pins if an interruption is generated during the sequence PWR_ON.

    For now it latches until the power supply is correct, but in case of short-circuit the ENx pin that causes the problem stays up even if it causes a short-circuit.

    Do you know how I can shut down the EN pin to avoid any destruction?

    Regards,

    Tom LEVISSE

  • Hello Tom, 

    As i mentioned prior the safety concept here is that the sequence should be halted if an error occurs, such as a power supply not being correct. This way no other power supplies are enabled avoiding further harm to the system. In the case of TPS38700s the only way to force all ENs low would be to disconnect power to the device. Our TPS38700 device does feature an Emergency Power Down Pin which will force a power down sequence to commence but this feature is not currently available in the TPS38700S configuration. 

    Regards, 

    Oscar Ambriz 

  • OK thank you for your time. 

    Maybe you should add some diagram in the datasheet describing the behavior of the TPS38700s-q1 in case of bad sequences.

    It would clarify a lot of things...

    Regards,

    Tom LEVISSE

  • Hello Tom, 

    I agree with your feedback, I will close the thread if no further questions.

    Regards,

    Oscar Ambriz 

  • Hello, Thank again for your time ! 

    Do you know when the next revision of the datasheet will be released ?

    Regards,

    Tom LEVISSE

  • Hello Tom, 

    Due to ongoing datasheet efforts I do not expect this datasheet update until September 2024. 

    Regards, 

    Oscar Ambriz