LP5907: Output 4.5V no load, light load output ripple high

Hi Gabriel, 

Thank you for sharing this information. 

Is the question related to why there is ripple at the output of the LDO? 

Does this ripple appear when you test using the battery only, or also when there is a 5V DC supply? 

Best, 

Edgar Acosta

Dear Edgar

Yes , the output of LDO ripple is abonormal .

Both 5V DC and battery test output ripple is high ,and the failre rate is about 16% .

Hi Gabriel. 

Is there a GND plane under the LDO? Also, Is it possible to get a sartup up waveform of Vin and Vout in a same scope shot, in addition to a board that is behaving normal? It is odd that there is not even a 4.5V rail. From the scope shots I see the voltage going below zero from the no load and light load , correct me if I'm wrong. Lastly, what is the voltage rating for C4? 

Best, 

Edgar Acosta

Dear Edgar

1. In the middle of the LDO chip, the copper floor is laid, and the next layer is also complete.

2. The input and output startup waveforms are shown as follows:

3, oscilloscope noise measurement is through a separated straight amplification front circuit, and the oscilloscope adopts AC coupling, so DC4.5V can not be seen;

4, C4 pressure is 10V:

Hey Gabriel, 

Thank you for sharing this. I wasn't sure if the scope shots provided were AC or DC coupled and wanted to make sure. 

After looking at the capacitor, it could be that once it derates, it is causing some stability issues: 

Here is the capacitance change with a DC bias of 4.5V, and its right at 0.7uF, which is the bare minimum needed for stability. This does not take into account any temperature change, or the +-10% tolerance of the part. 

Is is possible to place another cap in parallel just to verify this? 

Best, 

Edgar Acosta

Dear Edgar

The output capacitance has been changed during the actual test.

1, one more 1uF capacitor in parallel

2. Replace the output capacitor with 4.7uF ceramic capacitor

3. Replace the output capacitor with 10uF ceramic capacitor

The output is still problematic

In addition, will the stability of LP5907 have such strict requirements on the output capacitance?

Hi Gabriel, 

The LP5907 does have a very narrow requirement for the output capacitance. However, if the recommended values stated at the data sheet are followed, then the device should behave as expected 

I was looking at the output waveform and noticed that the oscillations occur ~46Hz very close to the 50Hz from the AC line. 

If the issue still persists by changing the output cap, it could be that noise is coupling in from another source.

There are several ways to verify this. Disconnect any source near the board, i.e. power supplies, Multi Meters etc. except for the Oscilloscope needed. Test using a battery and a shield box (if available). 

You mentioned a separated straight amplification front circuit? Has it been verified that this provides a clean output using other sources? 

Also, when doing the test, has Vin been looked at as well? This will help us verify that Vin does not show oscillations as well that might be coupling to Vout. Lastly, probing directly at Cin and Cout (AC coupled) will be very helpful as well if possible. 

Best, 

Edgar Acosta

Dear Edgar

1. The test environment has been verified and there is no problem. The test single channel is built separately and directly tested at the input and output capacitors;

2. If battery is used as input test, the output will still have abnormal basis. At the same time, according to the performance index in LP5907 specification, even if the input ripple is present, the high PSRR of 5907 and the ability to suppress input voltage transients should not lead to such a large ripple voltage output;

3. The earliest 46Hz noise frequency in the attached figure is the 5907 output test noise frequency with normal output (the power frequency is introduced due to the influence of the connecting line, but the test voltage ripple is only 45uV at this time). The abnormal output frequency has voltage ripple 2.5mV@176Hz at no-load. The abnormal output frequency at light load 50uA voltage ripple is 2.5mV@3.44kHz;

4. The following figure shows the waveform after the input and output AC coupling;

5. It is found in the test that the output increases the load. When the load exceeds 0.5mA, the ripple will be significantly smaller, and when the load continues to increase to 1mA, it will drop to dozens of uV

Hi Gabriel, 

Thank you for sharing this information. 

When you mentioned earlier that the IC was changed and the issue is gone, do you refer to an ABA swap? Also, when doing the ABA swap, did the suspect IC got retested on a new board and the issue was gone, and/or the new IC that was installed on the suspect board showed that the issue was gone as well? 

When changing the output capacitors, was the same frequency and amplitude observed? It would be expected that changing the capacitors would change these values. 

Furthermore, can you describe more what the load is? How is the 50uA being measured, i.e at Vout, expected value from Data Sheet, supplied current? 

Is the No load and 50uA load a requirement that is causing issues on the system, or is this something that was observed while testing the part? 

Can the customer use TPS7A20 as a short term solution? This is the next generation of LP5907 and should offer better performance. 

Best, 

Edgar Acosta