TPS23861: Manual Turn-on to device w/o Signature

Hi Thomas,

What's expected the port current for the CAT5 device and coax device? For keep a TPS23861 port on, during the inrush stage (60ms after port is on), the port current should be lower than 425mA and during normal operation, the port current should meet valid MPS signal(a minimum dc current of 10 mA or a 10-mA pulsed current for at least 75 ms every 325 ms). Can you please take waveform capturing port voltage(Drain-Vpwr) and port current(current gun at the pairs having current going through) and check TPS23861 fault registers(0x00, 0x02 ~0x0B )? I can help with the analysis. Thanks.

Best regards,
Penny

Hi Penny,

Thanks for your quick response.  I'll get some waveforms, but I can answer some of your questions now.  The normal operating current for both units is about 30 to 35 mA.  The Start/ILIM Event Register (0x09) does show a start fault for the ports in question.  I should mention that nearly identical units that have a signature chip start up without any problems.

Thanks again,

Thomas

Hi Penny,

Here's a shot of the (very brief) turn on waveform.

Thanks,

Thomas

Hi Thomas,

I can't tell what's happening with the limited information. Can you please take waveform capturing port voltage(Drain-Vpwr) and port current(current gun at the pairs having current going through) and all register dump(0x00 to 0x0B)? Thanks.

Best regards,

Penny

Hi Penny,

The image I sent earlier is Vpower relative to the drain voltage.  Now I'm attaching the voltage across the current sense resistors (0.255 ohms).  I'll send the register dump later.

Thanks,

Thomas

Hi Penny,

Here's the register dump.

Address	Value
0x00	0xC1
0x01	0xE4
0x02	0x01
0x03	0x01
0x04	0x00
0x05	0x00
0x06	0x00
0x07	0x00
0x08	0x01
0x09	0x00

The port in question is port 1.  Ports 2 & 3 are powering devices with signature chips and working normally.  Port 4 has a different problem.  I think the FET is shorted on.  That's not my concern now.  Port 1 has the waveforms I sent earlier.

Thanks,

Thomas

Hi Penny,

The following register dump might be more useful.

Address Auto Man1 Man2
0x00 0x88 0xC3 0xC1
0x01 0xE4 0xE4 0xE4
0x02 0x00 0x45 0x01
0x03 0x00 0x45 0x01
0x04 0x0F 0x00 0x00
0x05 0x0F 0x00 0x00
0x06 0x00 0x00 0x00
0x07 0x00 0x00 0x00
0x08 0x00 0x01 0x01
0x09 0x00 0x00 0x00

The second column lists the register values after the port was in automatic mode. The third column shows the first readings after setting the port in manual mode. The last column shows the values after repeated readings in manual mode.

Thanks,

Thomas

Hi Penny,

I just realized that I hadn't included all the registers you requested, so here's an updated register dump.

Addr	Auto	Man1	Man2
0x00	0x88	0xC3	0xC1
0x01	0xE4	0xE4	0xE4
0x02	0x00	0x67	0x01
0x03	0x00	0x67	0x01
0x04	0x0F	0x00	0x00
0x05	0x0F	0x00	0x00
0x06	0x00	0x00	0x00
0x07	0x00	0x00	0x00
0x08	0x00	0x01	0x01
0x09	0x00	0x01	0x01
0x0A	0x30	0x30	0x30
0x0B	0x00	0x00	0x00

The columns are the same as my earlier post.  Any ideas?

Thanks,

Thomas

Hi Penny,

The current appears to be well below 425mA. The second waveform I sent shows a voltage of around 20mV across the 0.255ohm current limit resistors. Am I missing something?

Thanks,

Thomas

Hi Thomas,

Can you please connect your load to TPS23861EVM-612 and hook up the GUI? I want to isolate the problem. Thanks.

Best regards,
Penny

Hi Penny,

We are now convinced that the problem is that the current is limited to less than 100mA initially and power limited until the voltage rises enough  (over 27V).  The foldback during inrush curve (figure 57 in my version of the datasheet) shows the limiting.  Our devices try to turn on (and draw well over 100mA) at fairly low voltages (well below 20V).  Is there a way to change the turn on to allow higher current at lower voltage?

Thanks,

Thomas

Hi Penny,

Thanks for your response.  The devices at the load side are established products that we can't change.  We're hoping to use the TPS23861 to replace a chip that'll soon be discontinued.  We expected that in manual mode the TPS23861 would just turn on without trying to meet IEEE requirements.  We have found a possible workaround.  If we reduce the current sense resistors enough, the chip reads a lower current and turns on our load devices.  We haven't settled on the final value, but it'll probably be between 0.1275 and 0.170 ohms.  We can adjust our user interface to give the correct current readings.  We are not concerned about limiting the power sent to the load devices.  We have current limiting fuses on each output port, so we aren't concerned about overloading the TPS23861.  Are there any reasons we should not use the TPS23861 this way?

Thanks,

Thomas