LM25085: Vout dropping when loading the buck regulator

Hello

Here are some thoughts:

1. The connection from C7/R12 must be made directly to the source of the FET for current sensing.

You could try to cut the connection you have now, and connect a short wire to the source.

2. In the end, you still may need to adjust R12 to get the current limit you need, if that is the issue.

Since the FET Rdson is variable.

3. Double check you design with the quick start calculator tool on our website.

4. If you need well defined current limit, then the alternative method of using the sense resistor would be more accurate.

Thanks

Thanks for the quick response Frank.

We are not concerned about accurate current limit, our load will take an absolute max of 8A, usually at most 5A. We don't need the current limit to be accurate. Even if the current limit is 15A, It is fine.

For 1. the only issue that will occur with this is that the small length of copper pour resistance would be added to Rds(on), right? This would add 2 mOhm or so at most, right? Even with this the system should work.

For 2. we changed R12 from 2.7k to 4.7k without any change in the behaviour.

We have more observations now of the output voltage in different load conditions:

- no load: 12.002V

- 1k: 12.051V

- 100E: 12.629V

- 15E: 11.697V

- 7.5E: 10.894V

This is very strange because there are situations where the Vout is increasing and decreasing according to the load. Any idea why this could be the case?

Thanks, Prithvi

Hello

It is possible that the device is unstable.

Check the switching wave-forms during the load testing.

Also, try disconnecting the trace from the R17/C8 node and

connected directly to the (+) side of C9/C14.

Thanks

Hi Frank, thanks for getting back to us.

We did as you suggested - getting the 12V out with a wire from C14 + terminal to the R17 feedback divider. No change in the behaviour

We also removed R25 resistor that is added to remove the voltage ripple to see if there is any issue it is causing with the feedback but no change in behaviour.

Here are the waveforms of the switching node in red and the PFET gate voltage input from the regulator (yellow) for different loads

7.5 Ohm load

15 Ohm load

100 Ohm load

1k load

No load

Please note the difference in time base in no load condition. Can you please let us know the reason for the oscillation? Here's the bottom layer with the top layer components. Would switching to a continuous GND layer in bottom layer help? How will the thermal dissipation from the switching node be in that case?

Thanks, Prithvi

Hello

The ringing you see is just the result of the inductor current ramping to zero and the inductor and parasitic capacitance ringing.

I am not sure that it is a thermal issue.  However, using more ground plane on the bottom would help with thermals.

I will ask the application engineer for this device to see if there is anything else we can suggest.

Thanks

Hello Frank,

Thanks for the response. How would we resolve the drop in voltage at output ?

Suggestions so far :

1. The connection from C7/R12 must be made directly to the source of the FET for current sensing. Modify the design

You could try to cut the connection you have now, and connect a short wire to the source.

2. In the end, you still may need to adjust R12 to get the current limit you need, if that is the issue.Modify the design

3. Double check you design with the quick start calculator tool on our website. We generated the circuit using webench

4. If you need well defined current limit, then the alternative method of using the sense resistor would be more accurate.Modify the design

Or Do you suggest any more tests to identify the issue ?

Thanks for your support. Any further support to resolve the issue quickly would be very much appreciated as we running against time for a demo.



Regards,

Darshan

With the exact values suggested in the Excel calculator, it is working much better. With 3.5A load the output voltage is dropping to only 11.7V, though there are still spikes of +/- 2V in Vout.

Hello

The spikes may be caused by the ESL or ESR of the Al. output cap.  Or maybe the ground routing.  The output cap grounds

should be near the diode ground with a short route.  Also, the spikes may be coming from your scope probing.  Be sure to probe

directly across the output capacitor.

Thanks