TPS65320D-Q1: TPS65320D-Q1 power transient performs not very good

Hello,

The output voltage deep, during load transient, is due to the following:

1- Measurement, the scope probe/ground should be very close to output capacitor terminal especially for relatively high current

2- Low value output capacitance

3- Slow loop response time (low bandwidth)

I'm analyzing the circuit and will post the result today.

Regards

Hello,

I have calculated the bandwidth and phase margin. The bandwidth is around 100KHz and phase margin is around 80 degree. The voltage deep formula is

1/2 x Iout/Cout x 0.35/Fc

Iout = 1.5A

Cout = 47uF

Fc = 100KHz

Voltage deep is 56mV.

There are tolerances due to parasitic capacitor on FB pin that can lower the bandwidth and effective output capacitor (lower than 47uF). This can cause higher voltage deep than the calculated value.

By increasing the output capacitance to 2x47uF and the compensation resistor to 42K, the bandwidth stays the same and voltage deep will be the half.

Please provide the following info:

1- Do you measure the output voltage across the output capacitor?

2- Is your system has problem or just observation?

3- Can you please send scope picture by zooming in at the first step load transient?

4- Did you measure the loop gain and phase?

Regards 

Hello, Mahmoud Harmouch

Thanks for quick response and information, I have attached my test result accoding to your required, It seems different compensation resistor effect the load transient performance, The TPS65320D-Q1 transient test result we do not accept especially when the LDO 1V8 is enbaled ,The 3v3 output voltage deeps is unsatisfied.Also the 3v3 ripple output is not good. So is there any good idea to decrease:

1, 3V3 Output deep voltage especially when LDO 1V8 enabled;

2, 3V3 output ripple (Our 3V3 output ripple required is 3.14V~3.47V)

Thanks and any quesitons please let me know.

Wait for your feedback!TPS65320D-Q1 Load Transient_0930.docx

Hello,

I understood from your previous post that the Buck voltage deep is due to load transient when you turn on the 1.5A load. The second post says that the voltage deep is when the LDO is turned on. In this case the original compensation loop is OK and the voltage deep is most probably due to 20uF LDO output capacitor.

Please reduce this cap to 10uF and try.

You mention output ripple. Is the load connected? If there is no load, the device is in low power mode and it is normal to see higher ripple.

Regards 

Hello,

First , According to datasheet, The LDO output(1.2 V ≤ V(LDO_OUT) < 3.3 V) capacitor need required from 20 to 40uF, We tried to update the LDO output capacitor to 10uF and test the result,When the LDO(1.8V) enbaled the DC-DC(3.3V) transient still do not get much improved (Around 200mV). 

Then, When we test the DC-DC ripple the load is connected. We test the voltage ripple from the load side very close.

Finally, The TPS65320D-Q1 load transient performance is not very good, Is there any similar type power IC that can be shared with us?

Thanks!

BR

Hello BR,

The Buck output drop of 200mV is due to the LDO output capacitor. The LDO output capacitor is charged from the Buck output capacitor. I believe the voltage deep can be lowered by increasing the Buck output capacitor. In the attached document, the schematic shows 1x47uF but the statement says 2x47uF. Have you done the test based on 1x47uF or 2x47uF? what is the voltage deep specification?

Regards

Hello Mahmoud,

Thanks for your feedback, The schematic we design for 1x47uF, But we do the test based on 2x47uF (As we find 1x47uF output capacitor test the voltage deep reasult is much worse),  As you explained the buck voltage deep maybe caused by the LDO output capacitors inadequacy. If we increase more output capacitors the reasult will be better, But this will increase BOM cost.Anyway this deep voltage is just occured on the power up sequence that we can ignore, There is another matter we are care about is the voltage riiple from the buck voltage (3V3), This ripple we would like to be more lower,the better( within 150mV).Currently we test the buck voltage ripple can up to 300mV, Comprehensive consideration we believve this power IC  performance is not very satisfied with our requirement.We are trying to search for one better power IC that can replace the TPS65320D-Q1. Is there similar type POWER IC can be suggested from your side? 

Thanks 

Perry

Hello Perry,

I believe the 300mV output ripple is at light load and when the ripple is lower when load is connected. This is due to pulse-skip at light load. The output voltage ripple at light load depends on the inductor peak current. It will be lower by increasing the inductor value. Would you please change the inductor value from 2.2uH to 4.7uH and measure the ripple?

I believe TPS65320 is the most compact solution for your application. TPS43330-Q1 dual back controller can be a solution but the BOM is higher.

Regards