Author Topic: Question about QFN-heatsink  (Read 336 times)

0 Members and 1 Guest are viewing this topic.

Offline paulscout550Topic starter

  • Newbie
  • Posts: 6
  • Country: de
Question about QFN-heatsink
« on: January 20, 2025, 09:55:51 pm »
Hello,

I am currently designing a PCB that includes a MAX17526A as a current limiter. The circuit is intended to output 20V with an adjustable current limit ranging from 2A to 3.5A. While the simulation seems to work like intended, I’ve noticed that the IC is dissipating too much power. According to the datasheet, the continuous power dissipation is rated at 2758mW (for TA = +70°C, with TQFN derating at 34.5mW/°C above +70°C). However, in my simulation, the IC dissipates around 3.5W and it would go in thermal shutdown.
Would it be sufficient to add a heatsink like this one https://www.ebay.de/itm/153188634825?_skw=QFN+heatsink&itmmeta=01JJ2T224JT5CBY3XA42ZVAPVT&hash=item23aac110c9:g:7CwAAOSwPlFeahBg&itmprp=enc%3AAQAJAAAA0HoV3kP08IDx%2BKZ9MfhVJKlqeaRHBO2gQ4wdYYISdpqgV1C7%2BMzmraTT4ouJIKZaq0j%2BctjIqKNZBa6RW0Qx%2F0K1RqiQaTCT5XgUbUA0ts3xHlMjNiSlH4qV3d9anwg9JTGqWqdjQMCaLqbiesc4w4B5DgBEpowuL2xNIgq3hdmbm7NsS5vYh9imwaOaEwVD6oKQFeyEbbDh6%2FmrmCDqCr57JbJC7pf7zjUyz%2Bp%2BXmLbh%2BUBScApvpdiZBGydPxW40YH8ZBKW2juz1hFP6KrAG0%3D%7Ctkp%3ABk9SR8CiiNqQZQ with a thermal pad, even though the IC doesn’t have a thermal interface on top?

Thank you  :)
 

Offline thm_w

  • Super Contributor
  • ***
  • Posts: 7663
  • Country: ca
  • Non-expert
Re: Question about QFN-heatsink
« Reply #1 on: January 20, 2025, 10:11:10 pm »
Would it be sufficient? No idea, probably depends on local airflow, seems like it might be small: https://celsiainc.com/resources/calculators/heat-sink-size-calculator/
Do you have the center ground pad tied to a plane with vias?

Thermal pad is ok but either you need mechanical way to hold it in place or adhesive thermal tape, which I've never found to be that great. Thermal epoxy is another option.
Profile -> Modify profile -> Look and Layout ->  Don't show users' signatures
 
The following users thanked this post: paulscout550

Online Siwastaja

  • Super Contributor
  • ***
  • Posts: 9490
  • Country: fi
Re: Question about QFN-heatsink
« Reply #2 on: January 21, 2025, 01:31:40 pm »
Mounting a heatsink on a top of an SMD component not designed to be cooled from the top is pretty iffy. If you are lucky, then the plastic case is thin and there is some thermal coupling through the plastic. Maybe your main cooling route through the exposed pad on the bottom to the PCB planes is then almost enough, and the poorly coupled top heatsink adds just enough dissipation to matter. But usually these are just cargo cult used by Chinese "ebay/aliexpress module" manufacturers.

It's hard to design because manufacturer does not give any data about the thermal resistance from junction to top of the package, so this leaves testing (and having to somehow measure Tj). Better idea is to make sure your design cools enough through the normal route (exposed pad on the bottom), spread the heat over a large copper fill on the opposite side of the board, and use a thin thermal interface material to couple a bigger heatsink on the PCB, opposite to the chip.
 
The following users thanked this post: paulscout550

Offline ahsrabrifat

  • Regular Contributor
  • *
  • Posts: 50
  • Country: pk
Re: Question about QFN-heatsink
« Reply #3 on: January 22, 2025, 12:15:01 pm »
As you are designing your own PCB, I suggest you include thermal vias in your PCB. Place thermal vias directly under the exposed pad of the MAX17526A. It will let the heat transfer more efficiently. Here are some tips:
https://www.pcbway.com/blog/2/PCB_layout_tips_for_thermal_vias.html

Distribute the vias evenly across the pad area for uniform heat dissipation.
 

Offline paulscout550Topic starter

  • Newbie
  • Posts: 6
  • Country: de
Re: Question about QFN-heatsink
« Reply #4 on: January 22, 2025, 08:54:46 pm »
Thank you for all the suggestions! I've decided to go with a different IC, the MAX17616A, which has a maximum continuous power dissipation of 3470mW. Additionally, it includes a feature where, if the junction temperature gets too high, it limits the current via foldback until it cools down sufficiently. But I still plan to include some thermal vias.  :-+
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf