From: blockedofcourse@foo.invalid   
      
   On 11/29/2025 8:28 AM, Martin Brown wrote:   
   > On 28/11/2025 23:35, Don Y wrote:   
   >>>> Do I need that much more "resolution" than simply imaging (sky and soil)   
   >>>> with an Ir camera?   
   >>>   
   >>> The idea is to have a two pixel imaging system consisting of a pair of   
   >>> thermometers with low thermal inertia one facing upwards shielded from the   
   >>> ground and another that is ground facing and shielded from the sky.   
   >>   
   >> The thinking being that "change" is quick and sudden?   
   >   
   > It is essentially a fairly pure measurement of radiative heat loss to the sky   
   > which will under most circumstances be a very good guide.   
   >   
   > Once you have a few temperature curves for typical conditions that's good   
   > enough. High cirrus can make quite a difference to thermal losses to the sky   
   > although nothing like as much thick low cloud does.   
      
   You continuous revise your models as that addresses changes in the   
   sensing technology. You don't ever want to have a special "recalibrate"   
   event.   
      
   >>> The temperature difference and its rate of change gives you the extent to   
   >>> which the situation is likely to allow a frost and/of fog to form.   
   >>>   
   >>> You also need windspeed and relative humidity. This is probably the   
   simplest   
   >>> kit configuration to predict risk of frost of fog.   
   >>>   
   >>> Dew point was always very important to astronomers and highway traffic   
   >>> monitors so you can get purpose built kit for this sort of thing.   
   >>   
   >> We used to used chilled mirror sensors for dewpoint.  But, were concerned   
   about   
   >> *actual* dewpoints as we were regulating a thermodynamic process.   
   >   
   > We were mainly concerned with not getting condensation onto the optics or   
   mast   
   > head amplifier depending on the type of receiver/sensor.   
      
   We were actually monitoring and controlling the mass of water in the   
   process.  Adding moisture in one place and waiting for it to be   
   extracted (dried) in another.   
      
   So, you conditioned the external air that was being fed into the   
   process using refrigeration to dehumidify with a control loop   
   around that.  Then, adding heat to the air with another control   
   loop.  Then, moisture (carried by materials that you wanted to   
   deposit).  The, extracting the moisture to ensure the deposited   
   materials had dried, using a final loop.   
      
   Of course, you wanted this to be as quick as possible.  But, more   
   importantly, wanted the results to be as CONSISTENT and PREDICTABLE   
   as possible.   
      
   >> I think I can cut corners, here, as I am looking for "correlations" and not   
   >> "numerics":   
   >>   
   >> "When I see this set of conditions, the following has proven to be true..."   
   >   
   > I think what we have proposed might be OK for that and fairly robust minimal   
   &   
   > cheap too.   
      
   The advantage of "non-quantitative" measurements is that they don't require   
   calibration against known standards.  The disadvantage is you have to do a   
   lot of watching and have a deep history.  Not the sort of thing you can usually   
   do with open-ended approaches.   
      
   I'll set something up and just start collecting data and logging results.   
   Winter tends to give us some "weather" that is different from out summer   
   weather.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)