I am working on thermal design and validation for a system using the Jetson Orin NX 16GB module.
While the Thermal Design Guide provides excellent steady-state thermal resistance information, I need additional data (or guidance) to improve transient thermal simulations and model correlation:
Is there any information, or does anyone have any experience with thermal design of such modules, and if so is there any estimate for effective specific heat capacity of the Jetson Orin NX module or an approximate material composition of the module (e.g. SoC, LPDDR, PCB, copper planes, shielding, connectors – approximate percentages) for me to be able to make an estimate?
Other suggestions for lumped thermal mass models or simplifications for system-level simulations would be much appeciated.
I am pretty sure you were referring to right collaterals which is “Jetson Orin NX Series and Jetson Orin Nano Series Modules Thermal Design Guide” and Table 3-1 for 40 W, Table 3-2 for 25 W or Table 3-3 for 15 W operational Mode. These are the worst case power that Jetson Orin NX 16G can generate and user can use these for their thermal simulation. and the mode can be set by using nvpmodel.
Let me know if you have any further question on the thermal simulation application note or the thermal map.
These are all that NV can share with Jetson users and
My main concern is that in the “Jetson Orin NX Series and Jetson Orin Nano Series Modules Thermal Design Guide” the components are presented as heat sources with the respective thermal resistances Rj-c and Rj-b, but I need to consider their thermal mass as well, as I need to model their temperature change over time, for which I need an estimate of the specific heat.
From the percentage composition of the module I could estimate the specific heat, and I have also seen some application where the properties of aluminium have been used, though they were rough estimates, and I would like something more accurate. If there is a model you know of that may model heating of electronic modules please let me know so I can give it a look.
Either way, if what you mentioned is all NV can share I may try to get an experimental value for the specific heat.