CSD19532Q5B,Excessive surface temperature

user5805170

Other Parts Discussed in Thread: PMP11186, CSD19532Q5B

I found a reference design of LM25037QMT on TI's official website.The title is PMP11186 schematic. With respect to the design of 250w push-pull converter reference design for automotive.
There are four mosfet named CSD19532Q5B in this reference design.I found that the reference design did not follow the heat sink.I want to know if I need the radiator mounted on it.
Because I found that at 30W, the surface temperature is 100 degrees Celsius.I don't know if that's normal and I want to know the temperature of the MOSFET surface when the reference circuit works at 250W.
If this is not normal, I wonder why the temperature is too high
ths!

over 5 years ago

0 Ryan Manack over 5 years ago

TI__Intellectual 2610 points

Hello, if I understand your question correctly, the four MOSFETs on the primary side are reaching 100degC at 30W output. This is not normal. The push-pull topology is sensitive to the transformer construction, particularly the leakage. If the leakage is too high, you will see excessive voltage spikes on the primary MOSFETs, potentially causing them to reach avalanche breakdown and consequently heat up. Can you measure the Drain to Source voltage of the primary FETs with Tip and Barrel method? training.ti.com/engineer-it-how-test-power-supplies-measuring-noise

The designer of PMP11186 is out of the office for the week. I will try to support you in the meantime.

Ryan

0 user5805170 over 5 years ago in reply to Ryan Manack

Prodigy 110 points

First, I found that the drive output voltage would be pulled down,At high levels it's about 5V,So I connected the VIN with the VCC pin,The temperature problem has improved a little.Just a little.

Below is the waveform about the Drain to Source voltage of the primary FETs

pl300.pdf

0 user5805170 over 5 years ago in reply to user5805170

Prodigy 110 points

After a day of research.I found several problems.

First,I changed the value of R13,R14 to 27 ohms and in addition to the D8,D9.

This will improve the overshoot of the waveform.as shown in the following figure:

But there was no improvement in temperature

Second,I didn't add inductance on the output before,Only the capacitor.I added a 22uh inductor and a 1000uf capacitor

This will improve the overshoot of the waveform.as shown in the following figure:

The temperature problem has improved.But after a while, it's still going to be 100 degrees

0 Ryan Manack over 5 years ago in reply to user5805170

TI__Intellectual 2610 points

OK, glad you found the issues. The output inductor is needed in the Push-Pull topology. It is the energy storage element, as the transformer stores no energy (or very little in the leakage). 27ohm resistors in series with the primary FET gates may be okay. I would start with 1ohm and work your way up (5ohm, 10ohm, etc.) to get the ringing profile that you need. However, the higher the resistance, the slower the switching speed, and higher the loss.

Ryan

0 user5805170 over 5 years ago in reply to Ryan Manack

Prodigy 110 points

The high temperature is caused by the RC absorption loop.It's not mosfet.
- How should this be resolved

0 Ryan Manack over 5 years ago in reply to user5805170

TI__Intellectual 2610 points

The snubber loss is a function of the capacitance. Please check the TI power management library and look for the paper by Philip C.Todd on snubber design, from the 19193 Unitrode Power Supply Seminar, SEM-900. www.ti.com/.../slup100

Thanks,
Ryan

0 user5805170 over 5 years ago in reply to Ryan Manack

Prodigy 110 points

ok, thanks!

Power management

Power management forum

CSD19532Q5B,Excessive surface temperature