I'm back on the Enermax solution, didn't have quite enough of the TC-5121 compound to refill crevices and anoint for a third go at re-seating the Enermax.
I ordered a 20g syringe of MX-4.
After my initial experience, in which I am chasing a few thorough tests already linked in this thread, the word that less-is-more is right on. Like minimalist art theory, looks like less compound = more temp reduction. And quite likely, the center of the processor is where the heat action is.
With my eye on subjective variables that any builder should be aware of, "once more unto the breach" I go with thermal compound, anti-static device and long handled screwdriver...
Here's what I got. I pulled the T-40 and documented the compound spread, which was maybe a 1/2 degree more successful than my first effort, based on p95 blend test results, see previous posts.
Enermax base
CPU top
As you can see, significantly less spread area and material in general, this was still the TC-5121 compound packaged with my Enermax cooler. This application gave me maybe .5C degrees improvement in heat dispersal over my first effort, as tested by p95 blend test at 26.6C ambient.
Not a great-looking spread pattern, but less overall material and perhaps just enough in the right place or places, whatever they are, yielded a .5C improvement in one blend test. Center-processor might be the key, which confirms the small-dot practice already evaluated by experts. Keep in mind the heat pipe cooler construction and crevice-filling technique
So, with a ginormous syringe of MX-4 in hand, (about $15 after coupon at newegg), I removed and assessed the T40 cooler. After cleaning, I noticed a nick in one of the center copper pipes. There is a corresponding ding in the surface of my processor, hmm, I didn't see that before now. Never mind the arc-shaped factory grind marks. I have read about lapping, and already purchased a pack of silicon carbide sheets at Harbor Tools for, damn, $2.99, are you kidding? 220, 500, 1000, 1200 grits, 2 each, wet/dry. Now I have a reason to try lapping the base of the cooler, there is an obvious nick in a center pipe area, I can feel it with my finger tip.
nick
It's time to try lapping the base of the cooler, since I am out of the 5121, whatever, science is really just a matter of writing stuff down so here goes. Dry sanding for starters.
after hitting it with 500 grit
1000 grit
1200 grit
and now, crevice filling and double dash anointment with MX-4
Then, I ran the blend test for about 1.5 hrs
26.6 ambient temp
Max 72 71 71 69
for an average of 70.75
Hmm, that's my science for the day!