[time-nuts] Room temperature control (was: Holdover, RTC for Pi as NTP GPS source)
namichie at gmail.com
Wed Nov 1 18:14:18 EDT 2017
You would be amazed by the effectiveness of installing a small fan, mounted parallel to the wall,
to create a slow whirlpool circulation in the room. Just a 10 w computer fan.
If the air velocity is below 0.3 m/s it is hardly perceptible. That takes about 30 seconds to get
around the room. The room becomes a well mixed volume of air. Heat sources distribute their burden
of heat with only small local temperature rise. The room contains a 50 to 100kg mass of air and provides
an averaging effect on perturbations.
I have constructed 7 labs using this principle and have little difficulty keeping temperature swings below
1.0 degree C. The temperature control sensor must be very small (fast), exposed to the air flow, have zero
hysteresis and be located on the wall that is opposite to the fan (and the AC unit).
The temperature sensor directly controls the AC unit, with the overriding logic that although a 0.1C deviation
from the set point can switch the compressor on immediately (not even a second of delay), when switched off
the compressor can not be switched on again for about 2 minutes, the time needed for the gas pressure in the AC
The natural cycle of about 2 - 3 minutes of on/off is attenuated by the integrated mass of the stirred air in
the room. 1 Kw of heating or cooling is 1 kJ/s, air has an enthalpy of about 1kJ/Kg per degree C, so 100Kg
of air heats/cools at 1/100 degrees per second. The room temperature gently cycles by a fraction of a degree
as the AC cycles.
If anything, the service life (10 years or more) of the AC unit is longer than conventional hysteresis sensor
All of this occurs because you stirred the air in the room!
> On 2 Nov 2017, at 6:32 AM, Poul-Henning Kamp <phk at phk.freebsd.dk> wrote:
> In message <20171101190841.366058d77e544d256b86248c at kinali.ch>, Attila Kinali w
>> The best we can do today is to have a well insulated room (no windows
>> with whith unknown power flows) and measure the temperature at a few
>> strategically choosen points. Then control the heat influx and
>> outflux using an approriate control loop. This will still result
>> in deviations of 1-2°C when somone walks in.
> It is a matter of energy balance.
> Humans emit heat on the order of a hundred watts and the only way
> to have that not affect the room temperture, is to "wash" it away
> in airflow with much higher energy content, as is typically done
> in clean-rooms.
> That still leaves you with the thermal radiation imbalance from the
> higher temperature of the human skin, which is why "nano" laboratories
> sometimes are kept at an uncomfortably warm temperatures.
> Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
> phk at FreeBSD.ORG | TCP/IP since RFC 956
> FreeBSD committer | BSD since 4.3-tahoe
> Never attribute to malice what can adequately be explained by incompetence.
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
More information about the time-nuts