Latitude E6400 / NVidia Thermal Issue ***SOLVED***

After landing one of these last year, and restoring it with working parts, I learned about the thermal issues firsthand.

At this point, I can ***DEFINITIVELY*** state that the problem had nothing to do with motherboards, as some stated Dell was supplying replacement motherboards when these were still being supported.

The problem is the poorly made heatsink assembly, and the fact that the GPU was not contacting the heat pipe, but just a shelf on the frame for the heat pipe, and even then only via a thermal pad.

This heatsink assembly requires manual adjustment to get it to lay flat on the dies.  The GPU section ***REQUIRES*** a copper shim.

Doing the above will improve temperatures, but the real key is in the Thermal Interface Material chosen for the job.

I was still getting hot, but improved, temperatures after doing the above, and using Arctic MX-4 for the TIM.  This is what got me suspecting the heatsink assembly itself.

I this past weekend, redid it as follows:

Northbridge: Phobya Ultra thermal pad (5W/mk, 1mm thick).

CPU: Coollaboratory Liquid Ultra

GPU: painted CLU on die, placed lapped copper shim on die, painted top of copper shim with CLU.

The end result is in Prime95 Large FFT (max heat generation), the CPU cores on a T9600 never exceed 82C (I still suspect it isn't perfectly flat on the die), and the GPU never exceeding 81C, while running a SETI@home enhanced (Lunatics) cuda workload on the GPU simultaneous with the Prime95 workload on both CPU cores.

On a more mundane workloads, averages are in the 60's with both the seti workunit crunching on the GPU, and DivX Author converting Captain America 2.

This same laptop would produce mid-upper 90's, with the previous configuration using Arctic MX-4 with good contact on the CPU, and copper shim.

This definitively shows that it was poor heatsink design.  Once the same heatsink is properely shaped, and thermal barriers removed, there is nothing wrong with the motherboard.

So far after several days, there is no degradation of the "shelf" on the heat pipe frame for the GPU.  I will keep an eye on that, and report back after a month if it still has no degradation.  I would think that if it was aluminum, it would have degraded by now.