TPS54360B: Overheating Problem in the power circuit

Hello,

For best thermal performance make sure the exposed pad under the IC connected to good GND copper pour.

See example layout image in section 10.2 of the TPS54360B datasheet.

Your schematic upload did not work, please try in a zip file. If you have layout images please share those too.

Hope this helps,

-Orlando

Hi, 

I did exactly like section 10.2 of the TPS54360B datasheet. Zip file attached for images.

Thanks for the help. I would be very happy if you can help by looking at the images.0844.file.zip

Are there any updates, could you see the screenshots ?

Please let me know. 

Hello,

Your layout looks good, it is a bit tight but overall similar to the TPS54360EVM-182.

Please probe SW node with an oscilloscope, this will help check if the converter is operating stable.

Also what component is getting the hottest? Is it the diode, inductor, or IC?

Let me know,

-Orlando

Thanks for the answer. The hottest component IC and Schottky Diode. Output inductor also hot but not that much. I check the SW node which is stable. 

How can I lower the temperature to 40 degrees and below.

Please let me know.

Hi,

Can you help me please, I'm still dealing with the high temperature problem?

Best Regards.

Hi,

It would be better if you shared your waveform.

From your schematic seems like your desired switching frequency is ~1MHz? or do you have an external clock?

Lower switching frequency will have lower switching losses, lower power dissipation, and lower temperature rise.

You should try switching at 500KHz, however you will have to size up your inductor.

With what I can see your top layer layout is OK from a heat perspective.

You can try to pull out the boot capacitor some to increase the width of GND copper pour around the IC and Diode.

This GND pour to diode and IC is doing much of the heat spreading, so the more polygon area the better.

Just be careful not to pull out the boot capacitor too far as this is critical.

Thicker copper will also improve thermal performance.

Do you have a thermal image? I'm not sure if the GND pour South of the TPS54360 is too thin and bottle-necking heat spread to that direction.

Let me know,

-Orlando

Hi,

Thank you so much for helping me. Yes 1 MHz external clock.

I changed the frequency to 600 kHz. I changed the elements connected to the COMP pin and RT/CLK pin. In this condition temperature is 65 Centigrade. 

Is it necessary to use a large size inductor or one with a small internal resistance?

PCB copper thickness is 2 oz. I will try to increase the width of GND copper pour.

I will add  attach. SW signal picture and temperature photo are attached. 

I'm waiting for your comments. Please help me.

Best Regards.

foto for ti.zip

Your switching looks good. What is your maximum output current?

I meant increasing inductance to adjust ripple current, reducing the switching frequency will cause the peak-to-peak ripple current increasing.

Right now it is ~1Apk-pk at 600kHz, and at 1MHz it was ~0.6Apk-pk. It should be ~30% of maximum IOUT.

Lower peak currents mean lower AC losses, and would help slightly with temperature.

Inductor temperatures are very specific, many datasheets give a temperature rise vs DC-I curve.

Using only DCR or footprint will ignore material properties, packaging, etc. 

I think opening out the boot capacitor and increase that GND plane will help the most. 

I quickly read through AN-1520 it mentions multlayer PCB with internal GND planes help spread heat internally.

For example the TPS54360EVM-182 has two internal GND layers.

Section 6 of that app note summarizes some other PCB layout tips such as as using larger vias under the EP.

Hope this helps,

-Orlando

Hi,

Max output current is 2A. 

What kind of inductor would you suggest to use in this case?

My PCB is 4 layer and 2 internal layers are connected to GND. 

I will check section 6 in the document. I used referance EVM for my PCB. 

I will try opening out the boot capacitor and increase that GND plane in next design version.

Thanks a lot.

Can you tell me more about inductor please ? If you give me example of inductor that would be awesome.

Typically high current inductors will have lower temperature rise at the smaller current so you could try that.

Also I would avoid a short or low-profile inductor. Taller inductors generally do better with heat.

For example these two Wurth inductors are both 8.2µH and in 10x10mm but have very different Temperature rise vs current characterestic curves.

https://www.we-online.com/catalog/datasheet/7447797820.pdf
https://www.we-online.com/catalog/datasheet/7443330820.pdf

I would definitely recommend the taller inductor as the temperature rise would be much less.

Look for Temperature vs Current curve or request one from manufacturer when selecting your inductor.

-Orlando

Thank you very much for the valuable information you provided.

I will order immediately this  inductor and like this.  I will try when the products arrive.

I will write my comment here again.

Best Regards.

Hi,

In my PCB, I changed Lout, Cout 150uF and Schotty Diode. Lout is bigger and much more reliable for current. Cout 150uF must be very low ESR. Sch diode's forward voltage 0.6 and lower choosen. This help a lot to reduce heat on chip. 

Also recommend, chip output pin area sholud be great. 

I order EVM module for check the temperature rise. When I get the EVM, I will write comment again.

With thermal paste and update Lout, Cout 150uF and Schotty Diode, temperature low 40 degree. But I want lower than that. I want to lower the temperature as much as I can.