At first, everything looked pretty good. However, I noticed when I was putting it together that the pad placed on the SSD didn’t have any pressure marks after assembling and then opening it, and it didn’t stick to the lid any harder, which should indicate a low pressure. I then turned my attention to this next and first measured the lid. Caliper and VHX gave the same result. There is a distance of about 3.5 mm between the rim of the shell and the heatsink surface. Let’s make a note of this value.
Next, I measured the counterpart with the SSD inserted. Here we are interested in the initial protrusion of the NAND over the edge support surface, since we need to subtract that from the 3.5 mm when it comes to the air gap the pad should fill. Of course, there are lots of factors for a certain tolerance range, but 0.7 to 0.75 mm is about right. A 1 mm pad is actually perfect, with a thickness of around 0.9 to 1.1 mm. Because one thing is clear, such a soft pad only works if you press it together properly. Simply laying it on doesn’t really help.
And how strong is the pad really? I measured around 1 mm, which corresponds to the usual size. However, this thickness is already somewhat borderline, because only really hard pads build up enough pressure to be able to work within an optimal framework. I consider such a soft 1 mm pad too thin for such a large gap of up to 0.75 mm. But it gets much worse, because the NAND is not the only component that needs to be cooled on an SSD.
The controller also gets quite hot, but its package is much lower. I measured that as well and was equally stunned. With an air gap of 1.2 mm, it has no thermal contact at all. It is one of the basics to determine the largest and smallest air gap for cooling beforehand, and then to optimally size the appropriate pad and its hardness when everything has to be done with one pad. The controller needs at least a 1.5 mm soft pad, which means that the pad used is already 50% too thin.
For fun, I placed the two included 1 mm pads on top of each other and could still just about close the case. However, too much pressure is not good either, so I ended up using a nice 1.5 mm OEM pad from Fujipoly, which fit perfectly. And you can very nicely do the test, tape or secure the case without the outer casing and then test where exactly it gets warm under load. The controller is so happy about the cooling that it recommended the spot directly above itself warmest.
Overall, the case got much warmer with the new pad, which indicated better cooling for the SSD. Which brings us to another shortcoming of the installed NVMe-to-USB controller: The SSD status and SMART values cannot be read out via this bridge solution and thus not the temperatures. Of course, I also checked the SSD, which is part of the process. After scrolling further…
27 Antworten
Kommentar
Lade neue Kommentare
Urgestein
1
Urgestein
1
Urgestein
Urgestein
1
Urgestein
Mitglied
Veteran
1
Veteran
Urgestein
Urgestein
1
Urgestein
Veteran
Urgestein
Urgestein
Alle Kommentare lesen unter igor´sLAB Community →