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.

[TPS2384] the behavior that can not be understood.

Other Parts Discussed in Thread: TPS2384

Hello~

One of my customer is making the NVR products, and He meets the problem that can not be understood.

The customer is using to TPS2384 with Alteranative A & Automode (MS pin = '0').

And, additionally, AC_LO pin is connected directly to drive the ALT_A/B pin.

Test environments, conditions & situations of the customer are as below.

1. Turn the power OFF the NVR system.

2. It is connected between NVR's PSE port and PD (IP-Camera) via Ethernet cable.

3. Turn the power ON the NVR.

4. ALT_A/B pin is '0'.

5. No communication between Host and TPS2384 by I2C.

6. TPS2384 is supply the power to PD, normally.  Below is the waveform of N-pin at this time.

7. While NVR's power is ON, user remove the Ethernet cable from PD.

8. User re-insert the cable.

9. But, TPS2384 is NOT supply the power to PD.  Below is the waveform of N-pin at this time.

10. Through the I2C communication, the Host is control the TPS2384, and the AC_LO pin is set to '1'.

11. ALT_A/B pin also is set to '1', because AC_LO pin is connected to the ALT_A/B pin.

12. After Back-off time (2 seconds), TPS2384 is re-supply the power to PD, normally.  Below is the waveform of N-pin at this time.

13. He claims to use the IEEE compliant PDs, but it was not confirmed yet.

14. In case of Other vendor's IP-Camera (It means another vendor's PD), it does not happen this kind of behavior.  TPS2384 is always supply the power to the PD even if user removes the Ethernet cable and re-inserts the Ethernet cable, while NVR's power is ON.

He wants to know Why does it do this kind of behavior?

Thanks... C.W. :)

  • It appears that the detection signature has failed on the PD after re-plug. Please read the port status register, bit D7 (discovery status) during the issue.

    The appearance is that the PD has some residual capacitor that remained charged and this is causing port discovery fail. When the back off time is increased to 2 seconds, then the PD can reset.

  • Hello Eric~

    Thanks for your kindness answer and sorry for late reply.

    I have one more question from your opinion.

    According to your opinion you referred to the residual capacitor.

    If the main reason is the residual capacitor, I think as below.

    if user removes the Ethernet cable and re-inserts the Ethernet cable after long time then the TPS2384 should be detect the PD, while NVR's power is ON.

    However the TPS2384 has failed to detect the PD after re-plug with long time.

    So my customer doesn't think that the residual capacitor is root cause.

    Through the changing the mode, the TPS2384 can succeed the discovery operation.

    However my customer would like to receive the clear explanation of the upper phenomenon, because the buyer of my customer requests the explanation.

    Thanks... C.W. :)

  • Would you be able to confirm whether the PD controller which fails is a type 2 per the IEEE 802.3at-2009 standard? As part of the mutual identification between PSE and PD types (now there are two types; 1 and 2 for 15.4W and 30W at the PSE), a type two PD controller must detect two classification pulse events from a type 2 PSE (a type 2 PSE that implements two-event physical classification). As part of this sequence, there are two MARK events and the MARK event voltage falls within the detection event voltage range. When a type 2 PD sees a voltage transition from the classification range to the mark range, the PD will present a resistance signature which is invalid with respect to detection (maybe ~12kohm).

    Here is one explanation based on the above comments:

    1. A type 2 PD is plugged into a TPS2384, ALT A PSE port. The TPS2384 port open circuit voltage is normally ~22V.
    2. When the PD makes contact, the port OC voltage falls from 22V down through the classification range into the detection range. This voltage drop/fall can put the type 2 PD into the MARK resistance range (~12k).
    3. When the TPS2384 is in ALT A mode, there is no backoff time between detection tries (4.4V to 8.8V to 4.4V to 8.8V continuously).
    4. When the TPS2384 is in ALT B mode, there is a backoff time and the port voltage falls much lower than 4.4V. This is probably enough to reset the type 2 PD out of the MARK state and allow it to present the normal 25kohm detection resistance.

    This is one possible explanation of the behavior, but the TPS2384 measured detection registance would have to be extracted from the TPS2384 registers when the failure occurs to be sure.