L7987L Not working as expected.

AMens.1 · ‎2024-12-23

Hello everyone,

In the link below I designed a DC/DC Buck configuration with the L7987L chip. We have used this chip many times but I needed to update the EMI performance so I reduced some loops on the layout side. Also removed the separate ground plane. Now there is just one ground plane under the buck circuit.

I was actually surprised that it's not working anymore. Who can help me with this?
It is switching but it is not regulating into the 0.8V feedback for some reason. I tried changing the compensation network but without any result.

HELP !

PS. I tested with different input supply voltages from 9 to 24V. In the document below, I mentioned 2 supply voltages.

PS2. I also broke some chips. It looks really unstable somehow. I think they are broken because they didn't switch anymore.

Schematic, layout and signal images:

https://docs.google.com/document/d/16G_f4qjTBYnygG3ZKfBmmoWEJaGLDoRwkzUj4_T-XSY/edit?usp=sharing

Peter BENSCH · ‎2025-01-02

May I ask you to upload schematics, layouts etc. directly in the thread, as described under How to write your question to maximise your chances to find a solution, bullet 3?

The L7987 is a switching regulator with a buck coil whose strong magnetic field can generate interference voltages in tracks and therefore pins. Unfortunately, you did not mention what kind of buck coil is used. In your layout, however, I can see a long track running directly under the coil that leads to pin 1 (VBIAS). However, as you have not connected the associated capacitor (MLCC?) directly to pin 1, but very far away, switching edges from the coil may well couple back into the regulator and thus interfere it. You can place the (or a 2nd) CBIAS capacitor directly between the via at pin 1 and GND as a test and check whether there is an improvement.

Regards
/Peter

In order to give better visibility on the answered topics, please click on Accept as Solution on the reply which solved your issue or answered your question.

View solution in original post

Peter BENSCH · ‎2025-01-02

May I ask you to upload schematics, layouts etc. directly in the thread, as described under How to write your question to maximise your chances to find a solution, bullet 3?

The L7987 is a switching regulator with a buck coil whose strong magnetic field can generate interference voltages in tracks and therefore pins. Unfortunately, you did not mention what kind of buck coil is used. In your layout, however, I can see a long track running directly under the coil that leads to pin 1 (VBIAS). However, as you have not connected the associated capacitor (MLCC?) directly to pin 1, but very far away, switching edges from the coil may well couple back into the regulator and thus interfere it. You can place the (or a 2nd) CBIAS capacitor directly between the via at pin 1 and GND as a test and check whether there is an improvement.

Regards
/Peter

In order to give better visibility on the answered topics, please click on Accept as Solution on the reply which solved your issue or answered your question.

AMens.1 · ‎2025-01-03

Hi Peter,

Even when there is a ground plane in-between this long wire, (on layer 2 which is not directly visible from the documents I provided) the disturbance was significant enough to prevent the L7987L from working properly.

I placed a second capacitor directly on pin 1 and GND and it worked perfectly.

Thank you so much for this solution!

Greetings,
André

Peter BENSCH · ‎2025-01-03

You're very welcome, great, that it's working now.

If the design is to be produced in larger quantities and therefore cost-optimised, you actually only need one capacitor on CBIAS, so you can remove the one that was placed first.

In order to give better visibility on the answered topics, please click on Accept as Solution on the reply which solved your issue or answered your question.