TPS552882: Output Current Limiting at 3.8A with Shunt Resistor Configured for 5A

Hi Paulo,

Thank you for reaching out. Can you change R49 to 36.6k and do the test again?

Regards

Mulin

Hello,

I tested with a 20kΩ resistor on R49, and the current still limits at 3.8A.

Hi Paulo,

1. With 60W output, recommend to set 400kHz switching frequency and use 4.7uH inductor.

2. Can you provide waveforms of inductor current, Vin, Vout, SW1, SW2? We can check if there is any abnormal behavior.

Regards,

Mulin

Hi,

1. I replaced R37 (FSW) with 49.9kΩ and the inductor with 6.1µH (the closest value to 4.7µH that I have available). However, this caused instability in the circuit, preventing the output voltage from starting up. Below is the waveform for VIN, VOUT, SW1, and SW2:

I believe the issue might be related to the control loop. I am changing the values of R51 (RCOMP) and C40 (CCOMP) to match those used in the EVM:

• Current R51 = 15kΩ | EVM R51 = 56.2kΩ
• Current C40 = 10nF | EVM C40 = 4.7nF

Once I have the results with this new COMP R and C configuration, I will post an update on the forum.

2. Attached are the waveforms:

a) No-load condition (0A):

b) 3A load condition:

c) 3.8A load condition (Current limiting):

Regards,

Paulo

Hi Paulo,

To improve loop performance, you can reduce R51 and increase C40. For example, 5k+15nF.

Can you check if there is any soldering problem?

Can you provide layout file and output capacitor part number?

When you test the waveforms with 3.8A load, did you see any oscillation when the time scale is zoomed out?

Regards,

Mulin

1) To improve loop performance, you can reduce R51 and increase C40. For example, 5k + 15nF.
R.: The value I initially set for R39 (FSW) to achieve 400kHz was incorrect, which is why the power supply output did not start. After correcting the value, the output properly regulated to 12V, but the current still limited at 3.8A, both with a 1µH inductor and a 6.1µH inductor.

2) Can you check if there is any soldering problem?
R.: I have three PCB samples, all assembled by automated machines at a company that manufactures multiple PCBs for us. All three boards exhibit the same issue.

3) Can you provide the layout file and output capacitor part number?
R.:

• Capacitor C38 (P/N 25SEPF56M) - Panasonic Aluminum Organic Polymer Capacitor, 25V, 56µF, 30mΩ ESR
• Capacitors C22, C25, and C6 (P/N CL10B105KB8NQNC) - Samsung Electro-Mechanics Multilayer Ceramic Capacitor (MLCC - SMD/SMT, 1.0µF ±10%, 50V, X7R, 0603)

The PCB design was created using Altium Designer. The project files are attached.
PCI_Backup.rar

4) When you test the waveforms with a 3.8A load, did you see any oscillation when the time scale is zoomed out?
R.: No, the switching remains continuous.

I performed a comparison of the voltage drop between ISN and ISP on the EVM and PCB to assist in the analysis.

Hi Paulo,

 I think layout is not good compared to our guideline. 5556.TPS55288 Layout Guideline (Rev B).pdf

If you still have chance to modify the design, recommend to follow layout guideline.

Since the ISN and ISP voltage limit 50mV, it makes sense that ISN-ISP is already 47.55mV when Iout is limited to 3.8A. We need to figure out why the current sence circuit has such large error.

1. You didn't add RC filter in the output current sense trace. Add a 0.1uF cap in parallel with the sense resistor.

2.Move the EVM current sense resistor to your PCB to see if the sense resistor error is too large.

3.

Paulo Piai said:

What issues could arise if I do not use a shunt resistor and rely only on the ILIM resistor for current limitation?

There is no issue if you don't use the shunt resistor. You can choose to remove the resistor.

But the layout is not good for me, especially power loop. There will be some spike issue. Recommend to modify the design.

Regards,

Mulin

1. You didn't add RC filter in the output current sense trace. Add a 0.1uF cap in parallel with the sense resistor.

Paulo Piai said:

I added a 100nF capacitor near the ISN and ISP pins, but the issue persists.

2.Move the EVM current sense resistor to your PCB to see if the sense resistor error is too large.

R.: When using the EVM shunt resistor, the current started to limit at 4A. There was a small difference, but I was still unable to reach 5A.
However, I realized that I am using the shunt resistor incorrectly in my design. The shunt resistor I chose for the project is a four-terminal type (L4T20FR010FER), and I mistakenly connected the V and I terminals together on the PCB. The correct approach is to use the I (Current) terminals for current flow and the V (Voltage) terminals for measuring the voltage drop across the resistor. I will modify the PCB and manufacture a new board for testing. Unfortunately, it is not possible to rework the current board.

But the layout is not good for me, especially the power loop. There will be some spike issues. I recommend modifying the design.
R.: Would it be possible to highlight the critical points in the layout I sent? That way, I can try to correct all the layout errors in this new design.

Thank you.

Hi Paulo,

Glad to hear your finding. You can modify the layout according to the guideline I sent you above and send me to review before gerber out. Main points: Vin and Vout power loop, VCC, AGND and PGND isolation, sense trace, drive cirtcuit.

Regards,

Mulin