LMR14203 Buck Converter issue [Help]

[Ahmad_k]

I am developing a tiny device that can handle DC input from 8 to 30V 100mA max, so my choice went to LMR14203 as I see it a lot of devices and it is working great.
I replicated the same PCB mentioned in their application note but I have weird result. The falling time of switching signal is high.
That's not a problem when my input is below 13V. Increasing it will make my signal near SAW TOOTH format and the chip starts heating even on 50mA load. Does anyone know what factor am I missing ? can be saturated coil or what ?
The falling time on a working LMR14203 is so small compared to my result

[sadman]

I am developing a tiny device that can handle DC input from 8 to 30V 100mA max, so my choice went to LMR14203 as I see it a lot of devices and it is working great.
I replicated the same PCB mentioned in their application note but I have weird result. The falling time of switching signal is high.
That's not a problem when my input is below 13V. Increasing it will make my signal near SAW TOOTH format and the chip starts heating even on 50mA load. Does anyone know what factor am I missing ? can be saturated coil or what ?
The falling
 time on a working LMR14203 is so small compared to my result

I think for fixed voltage output you need to recalculate the swicthing ferquency CAP (CB ) and L1 or use values that used in demonstration board user guide.

https://www.ti.com/lit/ug/snva502b/snva502b.pdf?ts=1744980548862&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FLMR14203

sadman

[dennis78]

1) Under what conditions (Vin, Vout, Iout...) is waveform?

2) waveform you show is voltage at U5.6 - GND  or U5.5 - U5.6?

3) Is your IC's genuine (many asian clones can be problematic)?


@sadman about swicthing ferquency CAP (CB ) - I think CB(8?) isn't swicthing ferquency CAP, it's bootstrap cap for driving internal mosfet?

[Ahmad_k]

I think for fixed voltage output you need to recalculate the swicthing ferquency CAP (CB ) and L1 or use values that used in demonstration board user guide.

https://www.ti.com/lit/ug/snva502b/snva502b.pdf?ts=1744980548862&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FLMR14203

sadman

I did follow the datasheet and the application note, those values in the schematic are just for reference, I did try a lot of different pairs or Cboost and Inductor

1) Under what conditions (Vin, Vout, Iout...) is waveform?

2) waveform you show is voltage at U5.6 - GND  or U5.5 - U5.6?

3) Is your IC's genuine (many asian clones can be problematic)?


@sadman about swicthing ferquency CAP (CB ) - I think CB(8?) isn't swicthing ferquency CAP, it's bootstrap cap for driving internal mosfet?

Vin was 10v, Vout 5.1V I=50mA Only.

Now I start believing that maybe it could be a clone. I spent all the day trying different values. I will order another batch from different seller. Those chips I have now can be used for input below 13v without any issue.

First I thought that my ceramic capacitor was 16v only and I said that's it, but it was 50v

Posted on: April 18, 2025, 06:24:19 18:24 - Automerged

2) waveform you show is voltage at U5.6 - GND  or U5.5 - U5.6?
@sadman about swicthing ferquency CAP (CB ) - I think CB(8?) isn't swicthing ferquency CAP, it's bootstrap cap for driving internal mosfet?

between SW and GND

[dennis78]

Clone IC's can be problem in most cases (they usualy sell it as LMR14203 but real chip work only for low voltages - many users will use it for 12 to xx converters so no problems

@between SW and GND -> im not sure but...that possible means that conveter work in dcm mode. It can produce problem for bootstrap voltage because C8 can't charge and it produce many other problems for this type of converter (did you try for C8: 10nF + 0.1uF-1uF?).

1. Check with scope voltage on C8.

2. What inductor you use? Did you try with inductor for much bigger current (22-47uH, 1A or etc...)?

[Ahmad_k]

For C8 I didn't try any value below 100nF as the datasheet recommended a Cap of 150nF.
For the inductor I tried 22uH, 47uH, 68uH, 220uH. Higher the inductor value, higher falling time.

I will check the waveform on C8 tomorrow

[dennis78]

Check waveform voltage on C8. C8 shouldn't be under 100-150nF (C8 is for supply driving circuit in active cycle so it must not be too low). I thougt add to them parallel 1-10nF (I not expect too much from this but try).

For inductors you used, what is declared current? (I ask beacuse of possible saturation. Bigger inductance for same size mean lower saturation current...).

I suspect on IC, but you check all other (beacuse of very high switshnig freq these ics are sensitive so layout also can cause different problems)...

[Ahmad_k]

Check waveform voltage on C8. C8 shouldn't be under 100-150nF (C8 is for supply driving circuit in active cycle so it must not be too low). I thougt add to them parallel 1-10nF (I not expect too much from this but try).

For inductors you used, what is declared current? (I ask beacuse of possible saturation. Bigger inductance for same size mean lower saturation current...).

I suspect on IC, but you check all other (beacuse of very high switshnig freq these ics are sensitive so layout also can cause different problems)...

1A max current coil. So it should be safe. I still suspect the IC as you said because I noticed the switching frequency was 1.75Mhz which is outside the limitation.
I will order from trusted reseller and report back in a week or two

[Sideshow Bob]

This is me just spitballing. But are you sure the inductor is up to snuff current and frequency way. The datasheet mention that saturating the inductor may cause the device going into thermal shutdown mode and the possibility of damaging the inductor. See the datasheet page 7-8

[Ahmad_k]

I am sure I am using 1A coil. also if the coil was saturated it would heat at least a little. the coil still cool even if the chip heated above 13V

[dennis78]

Ahmad_k: "I noticed the switching frequency was 1.75Mhz" It's convincing evidence in the direction of suspicion regarding ic. They often just overwrite labels on similar pin compatible  ics or sell pass failed of quality control or bad clones. My experience with digital ics from China is in most times ok, but analog/mixed and similar ic's is not good even I use it significantly below declared characteristics and for simple purposes.

If you have any wishes you can try with <22uH but with open core option (Of curse, Vin will be limited but >12V as now ). That inductors are generaly more suitable for higher freq. Be careful when connecting to keep parasitic capacitance to a minimum.

If not, we expect report for 15 days

[Ahmad_k]

Yes clearly it is fake chip. One thing I noticed also when I increase Vin is that the switching frequency is slightly decreased to 1.5Mhz which is  another sign of fake chip.

the attached image shows a genuine chip on a device drawing 50mA VS my PCB drawing 50mA, same Vin and same components value, only the chip is different

[Sideshow Bob]

Yeah the datasheet state max frequency should be 1.5MHz

[Catcatcat]

Please pay attention to such modules LMZM23601SILR...TPS82150SILT... from Texas Instruments.
https://www.ti.com/lit/ds/symlink/lmz21701.pdf
https://www.ti.com/lit/ds/symlink/tpsm861252.pdf
https://www.ti.com/lit/ds/symlink/tpsm861252.pdf
....
In terms of cost to size ratio, DCDCs are very attractive.

[Ahmad_k]

Please pay attention to such modules LMZM23601SILR...TPS82150SILT... from Texas Instruments.
https://www.ti.com/lit/ds/symlink/lmz21701.pdf
https://www.ti.com/lit/ds/symlink/tpsm861252.pdf
https://www.ti.com/lit/ds/symlink/tpsm861252.pdf
....
In terms of cost to size ratio, DCDCs are very attractive.

Interesting but outside of my requirements. it cannot handle 30V input.

[Catcatcat]

They can even have higher voltage
LMZM2360xxx Vin - 4V to 36V
TPSM5601xxx  Vin - 4.2V~60V