Case to heatsink thermal resistance for Wolfpack GM3 module

Duclair

Hello,

I would, please, like to know what is the recommended value for the case to heatsink thermal resistance (Rth,ch), to use in speedfit or PLECs when simulating a wolfpack GM3 module with pre-applied TIM.

Thanks.

Forum_Moderator

Thank you for your post, it has been approved and we will respond as soon as possible.

AustinC

Hi Duclair,

The WolfPACK modules (F and G) include the case to heatsink thermal resistance in the values provided on the datasheet, which is the thermal resistance from junction to heatsink, RTH-JS, and includes the thermal interface material. This is measured via a thermistor installed into the coldplate 2 mm below the top surface as shown in the image below. To complete the thermal stack in the simulation, you will only need to add a heatsink to ambient, RTH-SA, value. If you need a starting place for the RTH-SA value for the GM3, you can utilize ~0.13°C/W to 0.16°C/W for a liquid coldplate or ~0.25°C/W to 0.30°C/W for an air-cooled heatsink.

Best Regards,

Austin C.

Duclair

Hi Austin,

Thanks for your reply.

Sorry, I do not see RTH-SA on the online speedfit simulator. do you mean "MOSFET (grease) resistance Rth,ch", or the heatsink to ambiant thermal resistance for the variable temperature heasink model which is marked as "Thermal resistance Rth,ha" ?

Best.

AustinC

For the WolfPACK modules, SpeedFit will have the RTH-JS value built in, and you'll add the thermal resistance from heatsink to ambient.

Duclair

Ok, thank you.

My understanding at this point is that an example of approach here would be to leave Rth,ch (case to heatsink) as 0, since the device comes with TIM and the Rth-js built-in speedfit already includes the default case to heatsink impedance when using Wolfpack GM3 with pre-applied TIM. and then specify a value of ~0.13°C/W to 0.16°C/W for Rth-ha to model a small cold plate with variable temperature, or assume a fixed heatsink temperature and no Rth needs to be indicated in speedfit. Is this somehow correct?

AustinC

Hi Duclair,

You're correct; you can leave RTH,CH at 0, and fill in the RTH-HA value to model the coldplate.

Thanks!

Austin C.

Duclair

Ok, thanks.