Dose anyone know how to cool a lot of DPAK devices effectively ?
I have multiple Diodes / FETs that will disapate several dozen watts of heat . Now the footprints don´t include thermal vias so i cant mount a heatsink on the back of the PCB.
Putting every DPAK device on a seperate alumin(i)um breakoutboard would be an option but a Very expensive one.
And you can’t modify the footprint to add them?
Post an example foot print of what you have and one of what you want.
This is KiCad forum so the first thought is you are asking how to add these vias to footprint.
Is it your question?
Or you ask what to do if PCB is as is.
I think the key information is how many watts per DPAK?
One DPAK diode dissapates 70W of energy.
for the other devices its less.
I doubt this will be an acceptable design. 81 watts MUST be connected directly to a heatsink, and then it may not be enough. To put things in perspective, my computer processor dissipates approx 64 Watts. It has a huge heatsink with a fan to keep it cool. Even if you run your MosFets hotter than my processor, you will still need an aluminum heatsink (may or may not need a fan, depends on your enclosure etc).
An example: using a typical DPAK
The temperature rise of the junction to the case of the DPAK ~ 5 °C / watt = a rise of 5 * 81 = a 405 °C rise over ambient. Max junction temperature should be about 130 °C
So not only can’t you get the heat through vias on the board, but you cannot get that much heat out of a DPAK device.
1 Like
So I would have to use to-220 or simular?
What? Is this a joke or some AI cr*p?
Best case junction-to-case thermal resistance is around 1.8 K/W, which would give you a temparature rise of 126 K.
The diode has a forward voltage of 1.5v and I planned on rectifying 45a with it.
In some cases mosfets with active rectifier circuits are used to lower the dissipation from diodes. I don’t know your circuit or whether that would be a viable option in your case.
1 Like
The diode is used to rectify the output of a transformer. I don’t know how to use an active rectifier IC.
I understood: In total for all parts so I asked: how many watts per DPAK?
???!!! 
I never dissipate more than 0.5W in DPAK.
Its datasheet says it is 45A diode?
What is this diode type?
So 130-405=-275°C.
It is less than 0°K.
I bet it won’t work.
I am currently responding from my phone since I’m not at home but it’s a littrlfuse diode.
Should be 130 + 405 + 273 = 808 °K
No if you stay with using a rectifier with no mosfet then you will need a very high power package. Perhaps a stud mount or something like a VS-100BGQ015.
Would a to-247 diode something like c4d… Be sufficient?
Find the data sheet for the device you are asking about.
Add the Θ junction to case + Θ case to sink junction to case + Θ case to sink. In post 5 that would be 5 + 1.7 = ~7 this will be the temperature rise of the junction above the temperature of the heatsink measured at the device.
The units are °C(rise)/watt.
so for the device in post 5 … 7 * 81 = 500+ °C.
Now you can lookup some typical heatsinks and find their Θ (heat)sink - ambient. Add this number to the Θ junction to case + Θ case to sink. Calculate the temp rise. It should not exceed 130 or 140 °C when added to the ambient temperature
A typical hobbyist soldering iron only dissipates 25W to 35W. Maybe you could keep it from unsoldering itself from the board if you submerge it in a chilled liquid coolant?
I got the the temperature to which you need to cool case down to allow junction being not too hot for work.
What is the interpretation of result you got?
I understand your line of thought. Missed it the first time.
My calc was Tj in °K
I have no idea what you’re trying to do. The only things I can think of that would dissipate that much heat are massive switch mode power supplies or an H bridge. Both are extremely difficult to design and require professional attention.