LM5118: Buck-boost design, low efficiency at low Vin high loading condition

Hi Neo,

thank you for using the E2E forum.

I assume the efficiency goes down gradually. Do you have some more data points also at the levels above? Or where does is start to get less efficient?

First it would be good to find out which component is contributing to this loss.

Can you try to find out which component is contributing to the loss? This should be measurable by a temperature increase.

You can also check the voltage drop across the MOSFETs with differential probe?

Best regards,

 Stefan

Hi Stefan,

There should be some test setup issue with original test data. The updated efficiency is as below, which looks pretty good.

However, when P54V output current is monitored by current probe, the current waveform is swinging like sine wave. Even with e-load current added, it still looks like sine wave with higher amplitude. Is this reasonable or not? 

24Vin, P54V output current, 0A-eload

24Vin, P54V output current, 0.5A-eload

Here is where the current is measured and e-load added.

Thanks

Neo

Hi Neo,

Thank you for the update.

When going through the new questions I do not understand the first point:

"24Vin, P54V output current, 0A-eload"

-> How can there be any current oscillation with no current

Another point to check:

Can you check the output voltage with a scope probe. If this is stable and the eload should give a constant load, something with the eload or the current probe is not OK.

Best regards,

 Stefan

Hi Stefan,

Please find my update.

"24Vin, P54V output current, 0A-eload"

-> How can there be any current oscillation with no current

--> The system loading is about 7.9W. The current waveform is the system current (e-load connected in parallel with system but set as 0A)

Another point to check:

Can you check the output voltage with a scope probe. If this is stable and the eload should give a constant load, something with the eload or the current probe is not OK.

--> The output voltage is stable, the switching node waveform looks ok. I can provide the output voltage after capture it.

With current probed on e-load only, it looks constant at ligh load (<0.5A) but goes to swing gradually after e-load increase.

i will double check the current probe and e-load.

Meanwhile, if you have any other comments, please kindly let me know.

Thanks

Shisong

Hi Neo,

as mentioned, if the output voltage is stable (keep in mind we do voltage regulation) the issue is needs to be the eload side.

May you can try another one or use a shunt resistor as load to check.

Best regards,

 Stefan