TPS23861: TPS23861 inconsistent start up and working

Hello Mikha.

Have you placed the adapter through SIFOs testing and/or UNH-IOL certification? Also, is it connected to a PD load which has undergone similar testing?

Without additional information, waveforms, or certification testing, this is a difficult situation to help trouble shoot.

If wanting to learn more about the PoE logo program, please check out our E2E FAQ on the topic here --> e2e.ti.com/.../724427 .

Thank you for your consideration of PoE products from Texas Instruments,

Mikha, It has been a few days since your original posting. Do you still need assistance here or may I close the thread?

Thanks, Thomas

Hi Thomas,

Sorry for the delayed response. Which waveforms do you need?

Mikha, I assume that the TPS23861 adapter has not been put through SIFOs testing? We highly recommend that as the first level of debug in situations like this one.

With that said, I will re-assign this thread to PSE application engineer Penny Xu. She will be able to give more detailed guidance as we try to help you debug this situation.

Thank you for considering PoE products from Texas Instruments,

Hi Mikha,

When you have issues with power up the PD, can you capture port voltage and current waveform? And dump all register's status? I need register 0x00,0x02,0x06,0x08,0x0C,0x0D,0x0E and 0x0F status at least to know what happened in PSE. Thanks.

Best regards,

Penny

Hi Penny,

I have attached the Register Readout values for all registers as well as the waveforms for Port Voltage(Vpwr to VDrain) and the input Vpwr signal(Vpwr to Gnd). 

Note that the after Power On- Reset values for registers 0x00, 0xA0 and 0x2C to 0x3F changed to 0x00.

VPwr to Vgnd

VPwr to VDrain:

The conditions while capturing the readings are as follows:

Vpwr Voltage: 54V

VDD: 3.0 to 3.6

Load: No Load.

Please let me know if you need any other data.

Thank you.

Regards,

Mikha.

Register_status_readout.docx

Hi Mikha,

I have a couple comments regarding the register settings:

1. The I2C address was changed to 0x00 in register 0x11 which is an invalid I2C address for TPS23861.  You need to set the I2C address to something other tan 0x00, 0x0C and 0x30. You will also need to make sure bit 3 in register 0x11 matches the A3 pin status in hardware design. If you need autonomous mode(auto mode after power cycle), you will need to set AUTO bit in register 0x11 to 1.
2. Since you cleared AUTO bit in register 0x11, the device will be in OFF mode after power cycle, you will need to do the following configurations to make the device start negotiating with the PD:
   1. Set device in semi auto mode or auto mode in register 0x12
   2. enable DC disconnect in register 0x13
   3. enable detection and classification in register 0x14
3. If you configure the device in auto mode in register 0x12, the PD should be powered up after being connected. If you configure the device in semi auto mode, you will need to initiate power on command in register 0x19 after checking classification and detection status in register 0x0C ~0x0F(depending on which port you connect the PD to).

Thanks.

Best regards,

Penny

Hi Penny,

I was able to do the following:

1. Write new slave address to the TPS23861 as well as set the Auto bit.

2. Set all ports in Auto Mode through the 0x12 Register

3. Set the DC disconnect enable through the 0x13 Register

4. Set all the Detect and Classification fields for all Ports in the 0x14 Register.

But even after all this process  all the registers set to the written values there was no Detection cycle visible at the Drain pin.

And also all the values written to their respective registers specified above got cleared after a Power on Reset except for the AUTO bit and the newly written slave address which were retained even after the Vpwr and VDD power cycle.

And one thing I forgot to specify the last time is that I didn't program the slave address 0x00 into the 0x11 register it changed automatically.

Is it possible that there is a hardware fault that is causing the values of these registers to be cleared.

Thank you.

Regards,

Mikha.

Hi Penny,

I have attached the schematic.

Yes, I followed the procedure for programming the slave address into the 0x11 register. The value written to the 0x11 register is 0xF7. The new slave address is 0x77 with A3 connected to Gnd. 

Yes, I was able to verify that the value of 0x11 is 0xF7.

Before programming the new slave address I floated the AIN pin(pin no. 25).

Note:

1. D8(LED) is not populated.

2. I have also connected a 10uF capacitor across ~Reset and Gnd for following the Power On Sequence of TPS23861.

Thank you.

Regards,

Mikha.

P3_schematic.zip

Hi Mikha,

One thing I noticed on your schematic is that you should float the AIN pin all the time. Can you cut the trace on AIN pin and try again?
Can you capture a waveform including Vpwr, VDD and RESET? I want to check your power on sequence. Thanks.

Best regards,
Penny

Hi Penny,

I have floated the AIN pin. About the startup waveforms, I only have a two channel oscilloscope so I was able to capture only two startup waveforms at a time.

So first I captured the waveforms for VPWR and VDD simultaneously.

And then captured the waveforms for VDD and ~RESET simultaneously. There are two waveforms for VDD and ~RESET startup, one is for no capacitor connected between ~RESET and GND and other is when there is a 10uF capacitor placed between ~RESET and GND.

Thank you.

Regards,

Mikha.

P3_Startup_Waveforms.zip

Hi Penny,

I'll look into the VDD Power supply issue and get back to you.