Temperature strange problem with a switching regulator

Hello guys! I am wondering about the next issue:

I have used the next part of Monolithic Power to design the first stage of my power supply tree: MPQ4470A. Here you can see the schematic design of the part:



When I drive 24V to the input of the MPQ4470A, I get at its output 12V. However when the temperature increases, the output increases step by step until 24V. So when Ta=25 ºC  Vout = 12V. But when Ta=70ºC  Vout = 24V.

I have a PCB V1.0 where the sw regulator works fine. The only difference between my current PCB V2.0 and V1.0 is the output inductor L3. The PCB layout and the layer stackup is exactly the same. I have exchanged the inductors of V1.0 and V2.0 and PCB v1.0 works fine with both kind of inductors but V2.0 does not work fine with whichever inductor.

Finally I got a proper working of the sw regulator on PCB V2.0 making a short circuit L3. Then, the sw regulator works fine at whatever temperature. But I can’t figure out the reason of that behavior, and it is not a good solution because the output ripple is very large.

Anyone has an idea about what could it happen?

Thanks in advance and greetings.

Finally I got a proper working of the sw regulator on PCB V2.0 making a short circuit L3.

Click to expand...

Sorry please explain, you shorted out the buck inductor?

What does your output divider voltage equal at 24v output?

- - - Updated - - -

At the moment the only think i can think is the inductor saturates at higher temperature....also, maybe that mpqxxxx has a tolerance on the peak current limit, and the v1.0 has a lower limit.
When it goes to 24vout...is that only when output is no-loaded?
Maybe you need a minimum load on it....and the tolerance in the chips means v2.0 creeps up in vout when no-loaded

Hi,

You talk about PCBs ... and the layout is critical.

--> show us your PCB layout.

Klaus