[ntp:questions] Optimising NTP build for Raspberry Pi 2

Neil Green ncguk at hotmail.co.uk
Wed Mar 11 20:10:29 UTC 2015


On 10 Mar 2015, at 19:26, Mike Cook <michael.cook at sfr.fr> wrote:
> 
>> Le 10 mars 2015 à 19:13, Neil Green <ncguk at hotmail.co.uk> a écrit :
>> 
>> 
>> Thanks to everyone who gave answers to my question. David, as the GPS I use sits directly over the Pi I’ve just started a cron job to collect the CPU temperature every 10 minutes, which I’ll run for 24 hours or so, but was wondering if you (or anyone) could offer suggestions with regards to keeping temperature constant in a normal home environment (room heated but only as needed and not overnight, computer and GPS placed near a window through necessity (GPS antenna is placed on the window sill)). Any advice much appreciated.
> 
> Quick and cheap test to see if you get the results you need. Put the combination in a closed cardboard box to protect it from air currents and if you can find something suitable, surround the r-pi and receiver in the box with some thermal mass, or even foam to lessen temperature gradients. You may have to poke a few holes in the top if the cpu temp gets too high.  If you find you need better, prettier, you can use the pi itself to control a fan to regulate box temp. Plenty of sources on the web.  I myself am just using an insulated cardboard box without fan for one. It is fine for NTP. 

I’ve been putting off getting into the temperature control side of things thinking it would be complicated, but the cardboard box idea is something that had simply never occurred to me. I need to refine the box size and insulation side of things but just a basic box and a bit of foam has lowered a CPU temperature swing of 5 degrees to approximately 1 degree. This has been reflected in the pool.ntp.org <http://pool.ntp.org/> graph (which I know is not a serious benchmark but is representative). Brilliant stuff, thanks.


More information about the questions mailing list