4
\$\begingroup\$

I recently completed, had manufacturer, and assembled my (second attempt at a) wide input buck regulator using the LMR16030SDDA regulator ICs. This time around I used the Texas Instruments WEBENCH Power Designer to give me a base circuit to base my design on. The criteria given to the Power Designer were 12V out, 2A load, 14-50V in. I am testing it at 24V in, and at 1.5A load the diode gets hot enough to hurt (sorry, I don't have an accurate thermometer to test it with). When running at 14V in, it still functions, and the diode doesn't get nearly as hot (though probably hotter than it should). While it does seem to work it's getting a lot hotter than I want it to for longevity sake and I know something is wrong, but I cannot figure out what. At 50V I'm guessing it would probably get hot enough to burn me, though I have not yet tested.

My schematic is spot on with the recommended schematic in WEBENCH and all of the components are the same values and spec as what they called for. Most are in fact the exact part number they gave to use.

Any advice or tips as to what might be causing the heat is greatly appreciated. I am very new to SMPS design and so I have limited abilities in troubleshooting... I could use some advice from some experts. I have assembled the circuit twice using different parts, I've tried adding input and output capacitance, and I've tried replacing the diode with a similar one (60V 3A vs original 100V 3A) with no effect.

I know my layout is probably not perfect (open to suggestions there as well), but the output does seem to be stable so the heat is really my main concern. enter image description here

\$\endgroup\$
1
  • 3
    \$\begingroup\$ Depending on your sensitivity, an object at 140F may be painful to touch, but that temperature is also often within the safe operating region of many components. I would expect a diode could be too hot to touch, but fine for operating. \$\endgroup\$
    – Math Keeps Me Busy
    Jun 13 at 23:54

1 Answer 1

Reset to default
4
\$\begingroup\$

Objects at or above 45°C are already painful to touch. Your diode can operate at up to 150°C according to its datasheet, and it's not a problem at all for it to operate at 100°C continuously. This means that it's totally fine for it to get hot enough to burn you.

Let's run the numbers real quick: At 1.5A and 125°C, the diode typically drops about 0.55 Volts. Let's call it 0.6 Volts for some margin. The power dissipated in the diode is therefore 1.5A * 0.6V = 0.9W. Let's call it 1 Watt for some more margin. The datasheet lists its junction-to-ambient thermal resistance as 77°C/W typical, which means that the diode will heat up 77°C above ambient when you run the converter at full load and maximum input voltage. With 40°C ambient temperature, that's 117°C on the junction, leaving some comfortable margin to its 150°C maximum operating temperature.

Given that your PCB has a large ground plane and via stitching, the actual temperature will likely be even lower.

The diode is fine, and your PCB layout is pretty good as well.

If you want to make extra sure that it's not overheating, you can place a tiny drop of water on the surface of the diode after it's had some time to heat up. If the water doesn't boil, there's nothing at all to worry about.

Oh, and don't touch power semiconductors to feel if they're overheating - you'll just burn yourself.

If you want to keep the diode cooler, you can add a ground plane in the large empty area between the input connector and the diode. Copper pours on PCBs can act as heatsinks for connected SMD components.

\$\endgroup\$
4
  • \$\begingroup\$ As an aside, that diode's giving you a moderate hit to your efficiency. For 24 W out, 1 W in the diode is a 4% efficiency hit, which might be significant; if you can reduce that power consumption (a lower forward-voltage diode, perhaps, or switching to a synchronous converter instead of a diode-based one) you can improve the efficiency by a few percentage points. \$\endgroup\$
    – Hearth
    Jun 14 at 1:59
  • \$\begingroup\$ Thanks for the advice both of you. Looking back at Power Designer again after reviewing your replies, it occurs to me that I should have paid more attention to the circuit specs in PD. It actually tells me the expected power dissipation and heat. The Diode's expected Tj is 95C, and the IC's expected Tj is 107C! And that's probably based on their own PCB layout which is considerably larger than mine with much more surface area. While both of these temperatures are in spec, It's definitely a bit too hot for my liking. So, what is my best course of action then? Find a different IC? \$\endgroup\$
    – KpilotRCHelis
    Jun 14 at 5:14
  • \$\begingroup\$ @KpilotRCHelis There are buck controllers with synchronous rectification (e.g., LMR38020, LM76002). You could tell the Power Designer to prioritize efficiency when searching for ICs. \$\endgroup\$
    – CL.
    Jun 14 at 7:53
  • \$\begingroup\$ @KpilotRCHelis As a quick and dirty solution, you can place a 2mm ultra-soft(!) thermal pad on your PCB and strap a little heatsink to it (with a zip tie or two). The pad insulates the PCB from the heatsink and provides thermal transfer. As a more permanent solution, you can enter a higher ambient temperature in Webench Power Designer to force it into generating a design with less temperature rise on the semiconductors. \$\endgroup\$
    – Jonathan S.
    Jun 14 at 15:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.