Or just don't use a heatsink, but then most drives will throttle under sustained heavy read/write. It may be possible to overheat the flash if you cool only the controller, but I haven't heard of this happening.
Thanks everyone for coming to my presentation
Me either, and in those cases the controller - which will be cooler and thus more reliably functioning should be aware and throttle due to the overshoot on the NAND temps anyway. I'm sure that could be a problem on uncooled
arrays of SSDs perhaps, but the odds of running into that in a single drive with just general airflow would have to be miniscule (unless you
want to hand out warranty replacements all the time).
Good presentation - where are the coffee and donuts?
Interesting thanks. My Asrock motherboard has a heatsink for the 2 M.2 slots, but due to the ITX smallness, it’s shared with the H470 chipset. I guess the chipset and drive are keeping each other comfortably warm
That's how my ITX board is too (ASUS Strix Z270i). It soaks horribly and is basically useless for cooling at all - in fact, the SSD is much more resistant to throttling with the top
off and no heatsink at all. The PCH runs 2-3C
cooler that way as well, so it's a very confusing use argument. Of course, aesthetically speaking, it looks like utter crap that way - so I use it anyway.
Although I guess it's better than the grossly overpriced disaster in my main workstation (Rampage VI Extreme) where the cover prohibits using a heatsink, is made of
plastic, and completely prevents
any airflow from reaching the SSD. Luckily they also include the DIMM.2 riser for this purpose (and those use the CPU lanes) but it precludes using the slot under the cover for anything other than just basic storage - and even then large transfers will start it throttling in about 30 seconds or so. Fantastic design.