Other Parts Discussed in Thread: TPS2378
We are seeing occasional detection failures with the TPS23880 on our board.
The device is wired in 4-pair mode. Device is configured to run in auto mode.
The problem occurs with IEEE 802.3af and 802.3at PD devices that have dual diode bridges. See for example the typical application circuit shown in the TPS2378 datasheet. I believe this circuit assumes that detection and classification will only occur on either the A or B pairs, but not both. The circuit allows the PD to be powered using either the A or B pairs.
It seems the TPS23880 incorrectly measures the 25 K Ohm signature resistor with the above configuration. It records both low and high values. When the high exceeds Rgood max, the device appears to stop the detection process. It restarts if one disconnects and reconnects the cable.
We never see detection failures with one diode bridge.
Here is a summary of the state of the TPS23880 after a detection failure:
OM1G-PSE: PoE Status and Configuration
Port PWRG PWRE RClass AClass PClass Detect R(Ohm) V(mV) I(mA)
P1A No Off Unknown Unknown Unknown Ropen
P1B No Off Unknown Unknown Unknown Rhigh 33398
Port Mode DCDE CLE DETE DCUT MBP 2xFB ALTFB ALTIR
P1A Automatic Yes Yes Yes No 0 No No No
P1B Automatic Yes Yes Yes No 0 No No No
Port AutoClass Signature MAC AAC ACPWR(W) Pcut2P(W) Pcut4P(W)
P1A No Unknown No No 16.0 Auto Auto
P1B No Unknown No No 16.0 Auto Auto
Port NLM NCT PPCT DCDT 4PW Power-Allocation
P1 No No No No No 2-Pair 30W
Port DISF PCUT ILIM STRT Power-on-Fault
P1A No No No No No Fault
P1B No No No No No Fault
VPWR(V) Temp(C) TSD VDUV VPUV PCUT12 PCUT34 RAMFLT HRev FRev
53.619 68.672 No Yes Yes No No No 21 02
As can be seen, the detected resistance is 33398 and it is too high.
Here is the status for a device with a single bridge:
OM1G-PSE: PoE Status and Configuration
Port PWRG PWRE RClass AClass PClass Detect R(Ohm) V(mV) I(mA)
P1A Yes On Class 1 Class 1 Unknown Rvalid 25195 53798 40
P1B No Off Unknown Unknown Unknown Ropen
Port Mode DCDE CLE DETE DCUT MBP 2xFB ALTFB ALTIR
P1A Automatic Yes Yes Yes No 0 No No No
P1B Automatic Yes Yes Yes No 0 No No No
Port AutoClass Signature MAC AAC ACPWR(W) Pcut2P(W) Pcut4P(W)
P1A No Unknown No No 16.0 4.0 127.5
P1B No Unknown No No 16.0 Auto Auto
Port NLM NCT PPCT DCDT 4PW Power-Allocation
P1 No No No No No 2-Pair 30W
Port DISF PCUT ILIM STRT Power-on-Fault
P1A No No No No No Fault
P1B No No No No No Fault
VPWR(V) Temp(C) TSD VDUV VPUV PCUT12 PCUT34 RAMFLT HRev FRev
53.491 78.452 No No No No No No 21 02
My sense is there must be something like parallel detection going on. Is it possible this can be fixed in firmware or by a configuration change?
We don't want to lose the plug and play aspect of the current setup (e.g., by configuring for 2-pair ALT A, 2-pair ALT B, 4-pair).