LM5176: Stability issue

/resized-image/__size/960x500/__key/communityserver-discussions-components-files/196/5141.comp.png

/resized-image/__size/960x0/__key/communityserver-discussions-components-files/196/8765.fb.png

/resized-image/__size/960x0/__key/communityserver-discussions-components-files/196/8765.slope.png

/resized-image/__size/960x0/__key/communityserver-discussions-components-files/196/8765.ss.png

/resized-image/__size/960x0/__key/communityserver-discussions-components-files/196/7217.sw1.png

/resized-image/__size/960x0/__key/communityserver-discussions-components-files/196/7217.sw2.png

/resized-image/__size/960x0/__key/communityserver-discussions-components-files/196/7217.sw2_5F00_1.png

Hi Balazs,

the links in the first email are corrupted and the first 4 lines only link to the quickstart tool but does not provide the schematic.

What i can see from the scope plot is that the device continuously restarts which might be due to detecting over current events.

This could be due to a to short Softstart time or due to issues with the peak inductor current sensing.

Best regards,

 Stefan  

Hi stefan,

thanks for your reply. Sorry for the links. The SS cap is 68nF, aiming ~10ms rise time. I will try to modify the SS and measure the current sense circuitry, but I doubt it will show anything strange.

/cfs-file/__key/communityserver-discussions-components-files/196/6763.lm5176_5F00_sch.pdf

Hi Balazs,

schematic looks OK so far.

A few comments questions on the current sensing part of the layout:

- is the Filter for CS/CSG close to the LM5176?

- do you have a clean Kelvin connection from the CS/CSG to the sense resistor (and only a GND connection at the sense resistor

- Do you have a Sense resistor with terminals on the wide side?

- Can you try to use two shunts in parallel (e.g. with double the resistance) to reduce the parasitic inductance of the Sense Resistor

Best regards,

 Stefan

Thank you Stefan,

it looks like the V_CS voltage actually exceeds the cycle-by-cycle limit in buck mode. My PCB design unfortunately lacks a proper Kelvin connection for the current sense resistor, but the input filter is placed as close as possible to the LM5176.

Could increasing the filter capacitor value help suppress the V_CS spikes and prevent the OVC event? I've attached an image of the affected PCB area for reference.

Thanks in advance for any insights!

/resized-image/__size/960x0/__key/communityserver-discussions-components-files/196/cs_5F00_csg.png

/resized-image/__size/680x0/__key/communityserver-discussions-components-files/196/cs_5F00_csg1.png

/resized-image/__size/906x860/__key/communityserver-discussions-components-files/196/csg_5F00_pcb.png

Hi Stefan,

thanks for pointing out the PCB issues. That said, I’d still expect the circuit to at least stabilize at the target voltage under no-load conditions.

I tried increasing the filter capacitance, but unfortunately it didn’t improve the situation. Interestingly, when I apply a lower input voltage (around 8–10 V), the LM5176 operates properly in boost mode and delivers a stable 12 V output. However, as I increase the input voltage (10–14 V), it starts to misbehave and can no longer regulate properly.

Any thoughts on what might be causing this?

Best regards,
Balazs

HI Balazs,

again it tends to the layout and parasitic inductance of the Sense resistor as for Buck Mode the over current level has a lower value.

Best regards,

 Stefan