Cooling of N-FETs

[luudee]

Why in the world are the N-FETS located on both sides of the board ???


Has anybody actually considered users that want to draw a bit more
current and need to cool the N-FETS ????

If they all would be located on one side of the PCB, I could comfortable
mount it under a motor/propeller, with a small heat sink attached, and
the normal propeller air flow would do the job.

With N-FETs positioned on both sides of the PCB, it is virtually impossible
to create a decent environment for cooling the N-FETs.

Sometime it helps to think about the design choices we make ...

To bad for me the ESC32 are useless ...

luudee

[Trnquill]

Hhm, I wouldn't say impossible...





My heat sinks are way overkill, with shorter fins the whole package would be more streamlined. I haven't tried but I have a feeling heat sinks are totally useless in my application. Current draw is above "use without cooling" recommended upper limit, but installation position and direction pushes so much air past the ESCs I think they would stay cool even without any sinks.

But yeah, having all the FETs on one side would make things SO much easier!

[JussiH]

What kind of currents are you running - I have run 20A/16V continously and the ESC was barely warm.

Mounting them under the motors will secure enough airflow to keep them well cooled for most motor/prop combinations, even at higher currents. The whole PCB acts as a heatsink and all copper layers are 70 and 105 uM, to increase current handling and cooling capabilities.

You cant compare ESC32 with regular ESC´s. The dedicated Fet drivers make them much more efficient and they will stand up to much higher currents than traditional transistor driven ESC´s, before even beginning to get warm.

Actually, they are more likely to heat up under small loads, because the efficiency is lower at low effect.

N-fets are mounted the way they are to increase EMI resistance and to seperate power and ground planes for logic and driving circuits.

[Trnquill]

[kinderkram]

The factory default for the current limiter in the firmware is 20A continuously. Going beyond that limit will require attention to proper cooling and temperature of the ESC.

The sentence you quoted means: you can put it anywhere, even if you put it in an environment where it heats up.
10A is a no-brainer.
I´ll add an entry there to make it clearer.

Regarding 25-30A you should be ok with simple aluminium pads to distribute the heat as long as it gets prop wash.
No Babylon Tower needed.

[Trnquill]

[kinderkram]

Thx for pointing at it.

We always try to keep the Wiki up to date but sometimes we simply overlook or forget to change something that is outdated or obsolete. That´s why we have several threads dealing with Wiki issues: viewtopic.php?f=27&t=371

It helps us if you point at the flaws of the documentation so we can iron out misleading or outdated entries.

[luudee]

Trnquill, that's a really nice holder there. Where did you get it ?


Jussi, my calculations show just a tad under 30A. I am going by
the ESC description that anything above 10A should have heat sinks.
I am in an early planning stage for a new multi rotor. Since all
the HW design is not public, it is difficult to make independent
judgments. Don't mean to be rude, but the statement "N-fets are
mounted the way they are to increase EMI resistance and to separate
power and ground ..." sounds like a sad excuse ... Hire some
more experienced PCB designs ... lol ...
Could you please publish the N-FET part numbers ? And perhaps the
worst case timing model, so that we can do a proper analysis of
the worst case power dissipation ? That should be the proper way
to deal with this type of situations. Most of us had enough
physics education in high school to be able to figure out the
thermodynamics of this baby ...

Other than that I think you guys are doing a great job ! Please
keep up the good work. Let me know if you need any help with the
HW design, lol

Cheers,
luudee

[bn999]

[luudee]