So last night I went out to my loft of my barn where I have a heated insulated office that is only heated when I am out there to turn on the heat vs wasting money heating it when I am only out there maybe once a week or so. It was 27 F in the office. I turned on my computer and it came up with a hard drive failure at boot. After about 3 minutes of letting the system run with this error I  hit CTRL + ALT + DELETE and it booted fine. Looked at the SMART data and the drive says its still healthy which is strange I would have thought that this would have tripped a error counter but it didnt. My assumption is that the drive chilled to 27 F might not have been spinning at the full speed at boot due to it chilled and so after warm booting it it was already to speed. Within about a half hour the temp in the office climbed to a more comfortable 55 F. I should probably go with a cheap SSD in this system after observing this, but I was curious if anyone knows if there is a temp that hard drives need to stay above to function properly or if its just  if it works your golden and if it doesnt in a chilled environment upgrade to a SSD?Hard Drives have operating temperature ranges.

As an example, I have a WD Red 4TB Drive. WD lists the specs here.

In this case, the drive has an operating temperature of 0° C (32 F) to 70° C (158 F), with non-operating going down to -40. The spec can mostly be taken as the range of temperatures the drive is designed to run in, outside of that range and problems like the one you witnessed, or even worse issues (particularly from condensation) aren't unexpected. For HDD drives they use hydraulic bearings, at lower temperatures the hydraulic fluid may be too viscous to allow them to spin at their full speed, or, at very low temperatures, may even be frozen, preventing the platters from spinning at all. It could also be a limitation of the electronic components as well.

For HBoth kinds of drives, it is generally recommended to allow them to get within their operating temperature range before, well, operating them. An SSD doesn't really change this much- the rated operating temperatures on an SSD are the same on the devices I looked at (0-70), with the main difference being a much WIDER range of non-operating temperatures.

SSDs obviously don't have to deal with fluid bearings. Their failure mode at lower temperatures would be due to the design of the PCB itself, since thermal coefficients between the various metals in a PCB as well as contraction between components.To add to/echo BC's post, the biggest issue I've seen with HDDs in cold weather is condensation.  At a previous job we would have PCs awaiting repair sat in a freezing cold warehouse, when we brought them into the much warmer tech room for repair we'd try and let them warm up for a while before powering the machines on and we'd keep an eye on the drives even then, as bringing them in from a 0-4 degree environment at best into a room which could be 20-35C did sometimes leave them with drips of water on - not to mention other surfaces of the machines.
Not something that comes up very often but worth bearing in mind due to water and electronics not mixing well, a possibly more relevant example to most might be a PC or laptop left outside in the boot of a car overnight then brought into a warm house or office.

As BC says, SSDs have different considerations to HDDs but this doesn't mean they're immune to heat or cold.  I'd imagine some enterprise drives, or models for embedded use, may have a wider range of operating temperatures but I haven't Googled this.I was thinking that the computer although exposed to a large swing of temperature, that because the computer isnt temperature shocked  but instead gradually the temperature increases in the room that it probably wouldnt be that bad for it. Such as when the room warmed to 55F I touched the computer case and it was still dry because it took 1 hour for the baseboard heater to warm the room.

I know that electronics brought from cold to indoors where its warm can get condensate quickly on them and cause corrosion and shorting etc. Friend of mine ordered a Dell computer and as soon as UPS guy delivered it from cold truck he had to turn it on and he smoked it because he didnt wait for it to get to temp and dry on its own before power applied.

I have some extra plywood, and plexy glass and I might make a temp controlled cabinet for the computer. A simple 40 watt light bulb, and 80 mm 12 volt fan blowing at the bulb to move the airflow in an insulated cabinet with a thermostat set to a max temp of 65F and a relay to place the 40 watt load across properly rated contacts that can keep the computer into a safer operating temp is probably best for this office as the cheapest and safest to computer components method to have a computer out there kept within normal operating temps.

Was just thinking there is adequate empty space in a computer tower that you could even mount the light at the bottom base of the inside of the tower and have the thermostat and relay internal to the computer case to keeping it all warmer than outside temp and hopefully dry as well. The only problem though is the massive HEATSINK EFFECT of the metal case. So would you then surround the case in blue board insulation to lessen the loss of heat through the tin case. Kidding I am not going to go this route. 

Im gonna go with the cabinet that is temp controlled and make sure the computer doesnt roast in a confined space when operating as the best solution for this. I already have parts on hand to do this except for the relay. But I have a thermostat that I can set to 65F, I have 80mm fan and 12Volt power supply to power the 80mm fan and relay in series with the thermostat for the coil to energize and deenergize the relay to switch the bulb on and off as needed. To prolong the relay contact and bulb life I could even make it shut off the bulb at 70F and not turn the bulb back on until 50F or 55F this way the bulb isnt on and off frequently with a 3 degree temp swing but instead the bulb doesnt turn on to create heat from the filament until the temp drops by 15 or 20 degrees. Which would produce longer CYCLES vs frequent short cycles of hot and cold. 

The computer isnt anything that special for a system so if it dies oh well, but if i can make it last why kill it off early by neglect
Quote

Athlon 64 4850e Dual-Core 2.5Ghz 1MB L2 cache
2GB DDR2 800Mhz Corsair XMS2 (2 x 1GB sticks)
ASUS AMD Radeon HD5450 512MB Video Card
Windows XP Home SP3
40GB IDE Seagate Hard Drive ST340016A
DVD ROM 16x IDE
Rosewill 350 watt PSU

In the barn I do have 2 ducks and 15 chickens that are all egg layers... My office in the upstairs of it