TPS23754: Using xfmr Coilcraft POE40Q-24ED with TPS23754-hiccups on POE++ 56V

Hi Diang,

Long time no see!

As to the multiple .1 caps I agree-but for the EMI/EMC testing sometimes they are necessary.

If all else fails I will eliminate them during bench testing and see what happens.

I think i will DNP the 3 of them and if I fail conducted or radiated at EMC lab I can repopulate them easily 1 at the time.

We've always used the 3 caps without issue other POE designs.

I'll downsize the 220uf to 100uF.

I agree with your analysis of the diode, I've ordered some to replace the ones we are now using.

Also, I've read in other threads in the forum where TI recommends upsizing the C33 to say a 10uF to slow down the current ramp to allow the supply to stabilize with a deep inrush current. Would this help me keep my 220uF?

l will implement these changes and let you know what i find.

I've already been looking at the PMP22301.

Thanks again!

Ray

Hi Ray,

Hope you are doing well!

The IEEE 802.3 standards require a PD with detection cap 50 - 120 nF. Passive PoE injector is not in standard and it can power on the PD without detection.

Reducing the 220uF cap is not mandatory as well as it won't cause peak power test failure in Sifos standard tests. You can also try to increase the C33 to make the soft start period longer. 

Best regards,

Diang

Hi Diang,

A POE++ Phihong 56V POE60U-1BT-X-R Bt injector I am using (that we currently sell as an accessory) for a a non-related Bt amplifier is rated below:

Over-Current Output #1 <650mA; Output #2 <650mA; Output #1 and #2 <1300mA; Output #1 and #2.

When connected to this supply PoE board causes the Phihong Bt power supply to hiccup, it's not the PoE supply apparently.

As before when connected to a PoE+ switch, it boots up normally, so I do not think there is a "short" or "low impedance" causing the POE++ Supply to hiccup.

My thinking is that the PoE Bt having more power, the initial surge driving the primary winding is causing the Phihong PoE Bt supply to self protect due to over current.

I have another PoE Bt power supply that is also set at 56V that is rated 48 V, 2.5 A which is higher than the Phihong, it does not hiccup with the PoE board.

I have made the current 10K resistor R31 on my schematic 100K but that only seems to delay the boot up of the PoE and delay the Phihong supply hiccuping.

Separately, (not both at the same time) I have also increased the value of the C58 22uF connected to the VC to 470uF and again, just creates a delay.

What can you suggest to lower the initial inrush current of the supply so that I do not trip the PoE Bt Phihong supply?

Or at least a soft start "ramp" of current into the MOSFET driving the primary?

Thanks!

Ray

What about adjusting the blanking period? To allow the current spike to pass?

Or does that only control the inrush to the Mosfet before the Mosfet turns on? If so, then the overcurrent would still trip.

well im up to 220K on that blanking resistor, just takes longer for the Phihong Bt power supply to hiccup.

Hi Ray,

Thanks for your updates.

May I know if a higher voltage rated diode was changed? If so, may I get the part number?

For your issues of hiccup, possibly by inrush or surge current failure, it may be worth to run the test with C622 = 220uF DNP. Meanwhile, sometimes adjusting the feedback RCs can help to address too high surge current response too. You could try with the PMP22301's values. 

Best regards,

Diang