Re: Recommendation for Observatory PC


Thomas Swann
 

Please, where did I say -10C is extreme?

People should be aware what the rated temperature range is instead of
just cavalierly thinking I'll just use an SSD and it'll be fine.

10 C is a significant percentage of the rated temperature range ,and
there will be people here would like to operate a remote observatory in
temperatures even lower than -10 C.

Having worked on enterprise SSD firmware for quite a while I just wanted
people to be aware that there will be additional problems at the low end
of the temperature range.  NAND flash is inherently unreliable and
requires a large number of error correction bits just to make the
technology usable.  Modern (cheap) devices are using TLC or QLC NAND
which further exacerbates the problem as it reduces the read margin in
the chips.

On 2/27/2020 5:25 AM, Dale Ghent wrote:
-10C is hardly extreme. Get ahold of yourself.

On Feb 26, 2020, at 22:58, Thomas Swann <thomas@aperturefever.org> wrote:

I would be careful assuming that an SSD is better than a hard drive
particularly in a very cold environment. While the SSD has no moving
parts, NAND flash erases and writes are more problematic at low
temperatures. If the data was written at a high temperature, data
retention at low temp will improve, however.

The NAND technology makes a difference too. A SLC (single layer cell),
more expensive device will operate more reliably over a larger range of
temperatures than a MLC or TLC device.

Manufacturers do make SSD's that are spec'ed to operate at much lower
temperatures (-30C or lower) for more extreme environments, but you'd
have to source it at Digikey or Mouser instead of some place like
Newegg. Because the SSD isn't the only consideration it would be best
to look for a complete PC which is designed to operate in extreme
temperatures.

On 2/27/2020 1:15 AM, Dale Ghent wrote:
If you use solid state drives, you'll be fine.


Join main@ap-gto.groups.io to automatically receive all group messages.