I'd like to use the LTC4269-1 (LTC4269IDKD-1#TRPBF) in a design, but I'm seeing current spikes that are too high on the eval board (DC1335B-C). How can you explain this and what can I do to reduce them below the peak current threshold (684 mA)?
Let me confirm, you're testing LTC4269-1, DC1335B-C and you're seeing large input current spikes while testing with the auxiliary input? Your auxiliary supply is set to 44V and have a 1 Ohm series resistor to monitor current?
First, the converter is not seeing 44V. There is the current sense resistor voltage drop, about 0.6V, and the Schottky diode drop, another 0.7V. That will throw off your power calculations slightly.
Second, IEEE 802.3at does not specify auxiliary mode current limits. The PD must not backfeed power onto the ethernet cable, but it can draw as much current as required from the auxiliary supply. It also does not specify auxiliary supply input capacitance, so you can add a bulk capacitor between the AUX+ and AUX- terminals. I would recommend this if you're using a 1 Ohm series current sense resistor.
How are you probing the sense resistor? Long leads to easily inject noise into this measurement.
The reason for using AUX input and resistors for the load was to simplify the example. This way you could verify the results using your evaluation board.
To be more specific, the 44.1 V I measured initially was between GND and Aux+ using a handheld multi-meter at max load. When the current was flowing, the 0.6 V drop you mentioned would definitely be reflected in the scope measurement. That's why I measured it using a multi-meter but forgot to mention that.
The probe I am using the measure the sense resistor is a Tektronix P6245 probe connected to a Tektronix TDS 684B oscilloscope.
Here are scope shots without Aux input. I see the current spikes when I have a PSE (MAX5971A) connected to SPARE1, SPARE2 instead. This time the input voltage is 56 V and the load is not simply resistors. It's a custom PCB running from the 12.1 V PD output.
Since I am still seeing current spikes using a 802.3at PSE, could I increase the values on the Pi filter at the input of the transformer? L2 C5 on the DC1335B-C?
Problem solved. Measurement (user) error. I was getting false readings on the current because the probe was too close to the eval board. I moved the sense resistor R5 40 cm away from the eval board, and retested. Now I don't see the current spikes. (They never existed.)
I'm very pleased with the performance of the circuit. I also took some measurements to see what kid of efficiency I'm getting.
Test 1, Low input voltage
Vin 45.0 VIin: 0.557 AVout: 12.073 VEfficiency: 91.6%
Test 2, High input voltage
Vin: 56.7 VIin: 0.443 AVout: 12.045 VEfficiency: 91.0%
Thank you for your support and quick response time. I'm designing in the LTC4269-1. This is better than the Silicon Labs part is was evaluating (Si34061).