TPS23755: Schematic reivew

Hi Denny, 

Is this the customers own design or did they base this off of a reference design? This doesn't look familiar to me. 

The TPS23755 is recommend for primary side regulation (PSR) approach and it looks like they are using secondary side regulation. 

For secondary side regulation, I recommend taking a look at PMP8803 which uses the TPS23753A in 12V output. 

Otherwise if PSR is of interest, I recommend taking a look at TPS23755EVM-894 which uses the TPS23755 in 12V output. 

Please let me know if there are any other questions about this. Thank you! 

Best Regards, 

Brett Colteaux 

Hi Denny, 

I have some concerns about this design as it currently is. 

The transformer dot convention is incorrect. 

I also don't recommend the ferrite beads on the secondary. 

I also don't recommend the power diode on the low side, I recommend on the high side of the output. 

All in all, I am worried about the design, and I recommend they copy the TPS2755EVM-894 as this is known working solution for 12V. 

Also of note is that the AUX_D opto feedback circuitry is not required for designs that do not require adapter priority. It looks like they are diode ORing the 12V adapter on the output directly, and if that is the case, the opto feedback components can all be removed. 

Let me know if you have any question about this. Thank you. 

Best Regards,

Brett Colteaux 

Hi Denny,

The circuitry circled below enables the adapter to have priority over PoE:

In the customer schematic, they can add the 12V adapter input across C2775, similar to the TPS23755EVM-894 schematic shown above. This will enable the adapter to have priority over PoE. 

Alternatively, a more simplistic way to enable adapter priority is to diode OR the 12V adapter on the output. For the adapter to have priority, they would need to make sure the PoE output voltage is always lower than the minimum voltage of the adapter.

Please also see my feedback below based on the schematic:

• For D29, you can keep this on the board, but this may impact the detection current because the off-state leakage current is 5uA. May end up needing to DNP. 
• Recommend to change C2017 to 1nF.
• Must connect the node located at the anode of D67 to VSS.
• I assume what is labeled GND1, GND2, GND3, AND GND4 are the A1, A2, A3, A4 Anchor pins on the TPS23755. These are all connected to PGND but that is not correct. I recommend:
  • A1 needs to be connected to RSNS.
  • A2 needs to be connected to VSS.
  • A3 needs to be connected to VDD.
  • A4 needs to be connected to DRAIN.
• I assume what is labeled GND5 is the PAD. If so, I recommend this be connected to VSS.
• The VPD/VDD pin connections are incorrect. For instance, D6 can be removed and C2272 can also be removed. R2264 should be connected to VSS. Please reference TPS23755EVM-894 for how to connect these pins and the components. 
• Recommend to add 2kV caps to secondary ground from pin 2 of transformer as shown in TPS23755EVM-894.
• Recommend to not use BAV20WS-7-F for D68 and to instead use similar diode as in the TPS23755EVM-894 (voltage rating, current, and reverse recovery time) otherwise it could damage the FET.
• Recommend to use 680 pF for C1991.
• I recommend to add a 0ohm resistor in parallel with the output inductor L1524 as this may help with regulation. The inductor may make the regulation worse.
• Recommend to remove D63.
• Also of note is that the IC pinout isn't correct, recommend to correct this. 

Let me know once these changes are made and I can do another review. Thanks!

Best Regards,

Brett Colteaux

Hi Denny,

Please find answers to your questions below:

1. This is the data transformer which is used to extract the data sent over the Ethernet cable.
2. There are a part of the Bob Smith terminations, you can learn more about this in the app note here: https://www.ti.com/lit/an/slua469/slua469.pdf?ts=1613488928783&ref_url=https%253A%252F%252Fwww.google.com%252F

Thank you.

Best Regards,

Brett Colteaux

Hi Denny, 

I have reviewed the sections in green and it looks ok to me.

Thank you. 

Best Regards,

Brett Colteaux 

Hi Denny, 

I recommend to change the resistor dummy load to 0.5W (285 ohm and rated for 1W) so as to not dissipate too much power and I also recommend this resistor be moved to the right of the output inductor. 

Of note is that majority of customers real application load is able to supply the 10mA required MPS current without the need for a dummy load. I'm not sure the dummy load will be needed in the end but customer can always DNP the dummy load if they can meet MPS with their real load.

Best Regards,

Brett Colteaux

Hi Denny, 

The Bob Smith termination should never be connected to the primary side ground. 

In applications where there is no true Earth ground, Bob Smith can be connected to secondary side ground. 

If this post answers your question, please indicate so by marking this thread as resolved. Thank you. 

Best Regards,

Brett Colteaux 

Hi Denny,

1. In the schematic you posted, VPD and VDD are connected with a 0 ohm resistor, so electrically they are the same node. 

2. Yes that is what these pins connect to. Think about the primary power path, this is the positive side of the primary power.

Additionally, we would greatly appreciate it if you could please post new questions in new threads. This helps us keep track of questions and answers, and know if a schematic question got resolved, while a layout question is being answered. Thank you. 

If this post answers your question, please indicate so by marking this thread as resolved. Thank you.

Regards, 

Michael P.

Applications Engineer

Texas Instruments 

Hi Brett and Michael 

I appreciated your great support.

If the customer have another questions, I will create a new post.

Many thanks

Denny