[time-nuts] Re: poor mans environmental control for time servers (with pictures)
Eric Scace
read at scace.org
Mon Feb 7 14:46:17 UTC 2022
My brother, who worked on gas impurity metrology at NIST, tipped me off to another useful temperature control guideline: 2-3 cm of styrofoam adds about an order of magnitude to the time needed for external environmental temperature change to penetrate into a chamber.
E.g., If one sees an environmental 1º rise resulting in ½º rise internally after 10 minutes, adding that layer of styrofoam will roughly stretch that time out to 100 minutes⦠or 1000 minutes with double the thickness.
> On 2022 Feb 07, at 07:23 , Attila Kinali <attila at kinali.ch> wrote:
>
> On Mon, 7 Feb 2022 13:55:23 +0100
> folkert <folkert at vanheusden.com> wrote:
>
>> https://vanheusden.com/texts/usb-asic-miner-heater/
>>
>> What it is about: I've put my time-servers in a box and now them up
>> using usb bitcoin miners (- they have no use anymore because they are no
>> longer profitable) to 35 degrees celsius. As expected/visible in graphs,
>> this stabilizes the clocks of the time-servers a bit.
>
> Nice use of recycled electronics :-)
>
>
> Some small additional notes for those who want to do the same and
> use it to stabilize oscillators (OCXO, rubidium and the like)
>
> 1) If you are building a control loop, have a temperature sensor
> with more than 0.1°C resolution. 0.1°C is quite a large dead-band
> for a control loop in a case like this and might lead to oscillations
> in the control loop. I can recommend using a BME280 break-out board
> as a temperature sensor. It's relatively cheap (at least it was
> before 2020), delivers temperature to 0.001°C with very little noise
> (noise floor is below 0.005°C) and gives you humidity as a bonus.
> Its long term stability is better than I can verify with the equipment
> I have, so good enough :-)
>
> 2) Use a heating element that you can control with at least 10 bits.
> It doesn't really matter whether it's some PWM with 5minutes loop time
> or some heater element controlled with a DAC. Going below 10 bit will
> give too little control to hold the temperature stable and thus can
> lead to oscillations as well.
>
> 3) If you pass cables through the lid, tape the lid shut. Most of the
> heat will be lost through convection through the holes in the lid
>
> 4) It's not explicitly written in the article, but the small fan you
> see in there is cruical. There is convection even in these small boxes
> and it is very slow. You want to mix the air inside the box well
> so that you don't get temperature gradients and hot/cold air bubbles
> that move around.
>
>
> Attila Kinali
>
> --
> The driving force behind research is the question: "Why?"
> There are things we don't understand and things we always
> wonder about. And that's why we do research.
> -- Kobayashi Makoto
> _______________________________________________
> time-nuts mailing list -- time-nuts at lists.febo.com -- To unsubscribe send an email to time-nuts-leave at lists.febo.com
> To unsubscribe, go to and follow the instructions there.
More information about the Time-nuts_lists.febo.com
mailing list