TPS23756: Short RTN and VSS (without PSE)

Hello Chentsu,

My previous response: e2e.ti.com/.../720509

I have a couple of clarifying questions & comments about your questions that I will post tomorrow.

Regards,
Tom

Hello Chentsu,

Are you looking to apply your input on VDD-VSS or on VDD-RTN? In our we adapter appnote (www.ti.com/.../slva306a.pdf) we refer to these methods as option 1 or option 2, respectively.

Connecting the adapter to Vdd-Vss, you keep the benefit of using the RTN-VSS pass FET for inrush current limit, over current protection, and thermal shutdown.

To address your questions more directly:

1 & 2. VDD powers the PoE controller circuitry & VDD1 powers the PWM controller. If VDD is no powered, APD does not need to be asserted because the pass FET would be off anyways.

3 & 4. If you are using a 48V adapter on Vdd-Vss, there is no need to short RTN-VSS since you will exceed the VDD-VSS uvlo threshold (35V, rising) which will turn on the pass FET.

Thanks,
Tom Amlee

Chentsu,

additionally, I have revised my answer to the previous post to address the RTN-VSS scenarios.

e2e.ti.com/.../2658295

Thanks,
Tom

Hi Tom,

Thank you! understand that. It's very clear.

But now we connect adapter as Option 2 without PSE (powered from adapter only). And TPS23756 does damaged or getting abnormal sometimes when we short RTN-VSS.

So we are curious about why this cause damage or abnormal. Why we can't short RTN-VSS in this scenario.

Could you help to answer these question below for this kind of scenario, thank you.

1. If work without PSE,  can I short RTN to VSS before start-up? Why?

2. If work without PSE,  can I short RTN to VSS after start-up and PWM is working? Why?

===============================================================================

BTW, Our another customer have tried to use Option 1, input 55V and short VSS to RTN, the PWM work very well.

3. so can I say:no matter using Option1 or Option2, if want to use PWM part of TPS23756 only (without PSE) , PWM part will work as long as VDD1 get the power source?

Thank you.

Hi Chentsu,

For 1 & 2, shorting RTN to VSS is unnecessary. There is simply a pass FET across the two pins. Shorting Vss-RTN is not needed for option 2 ORing, since you only need the PWM controller portion of the IC. In the case of option 2 adapter without PSE (using the PWM controller, only) you can leave Vss floating.

3. Yes, that is correct. Vdd1 must get power.

For your customer, it is not recommended that they short Vss-RTN. Are they wanting to exceed the 850mA current limit on the pass FET?

Thanks,
Tom

Hi Tom,

For 1&2, I know about this, that is unnecessary. But why I ask this is we had tried this before and got IC damaged or abnormal.
We just want to know how this makes it abnormal? PWM portion should be independent with PD portion like what you answered in previous question3, so it should work well. Yes, that is unnecessary, we just want to know the IC better.

Could you help to answer these question below for this kind of scenario, thank you.

1. If work without PSE, can I short RTN to VSS before start-up? Why?

2. If work without PSE, can I short RTN to VSS after start-up and PWM is working? Why?

Hello Chentsu,

To answer your questions:

1. Shorting RTN to VSS prevents the hotswap FET from limiting the current during inrush. During inrush, ff there is no current limiting on the adapter itself, the bulk capacitor will look like a short as it tries to charge itself as quickly as possible. This will likely damage the IC.

2. Shorting RTN to VSS after startup is common. This is implemented when a customer does not want the operational current limiting so they can use higher power levels beyond the IEEE 802.3 standard. This can be seen in PMP8923 that uses a current boosting FET around the RTN-VSS pass FET.

Thanks,
Tom A.

Hi Tom,

Thanks for your responce, I'm appreciate for this. I think this is the last question.

but maybe you misunderstand my question, as what I mantioned proviously, we use adapter as Option2, so the current won't go through hotwap FET. so PD portion and PWM portion should be indepedent to each other, am I right?
 
Accounding to your responce, can I say shorting RTN to VSS won't impact to PWM portion no metter at start-up or normal running if using Option2 connection?