[ale] [OT] how do I monitor the "weather" in my computer room

[Alex Carver]

On 5/30/2013 12:13, Michael H. Warfield wrote:

> If you want to "roll your own" these are some decent DTH22 1-Wire
> temperature / humidity sensors that will interface directly to a
> Raspberry Pi or Arduino and claim an accuracy of .1 (presumably C) and
> only set you back $11 per sensor.  Doesn't say what the accuracy of the
> humidity part of the sensor (hygrometer) is but the precision is down to
> +- 1%.  Relative humidity is a real iffy thing to measure and calculate
> and a "sling psychrometer" is still the gold standard.  In most cases,
> the accuracy of your temperature sensor will bear a direct impact on the
> accuracy of your hygrometer.

You are very correct about the psychrometer being a gold standard.  The 
NWS uses automated wet bulb/dry bulb systems in their official weather 
stations to make the RH measurements.  Ultra precise laboratory 
measurements are made using a chilled mirror hygrometer.  But I don't 
think anyone here is going to rush out and buy one of those just to 
measure a living room. :)

Commodity RH sensors typically use capacitive sensors covered in polymer 
to detect moisture.  Some are resistive but similar operation, directly 
detect moisture.  They rely on the temperature sensor to complete the 
calculations.  They do drift because the polymer encapsulant can change 
with age and with extremes of temperature and humidity.

The one-wire devices are pretty nice for rough measurements like this. 
They do suffer an internal heating issues which is why the accuracy 
isn't as nice as the higher end sensors.


> Champagne taste, beer budget.
>
> Seriously, +- %5 is not bad considering what you're measuring.  You
> could move from one end of a room to another and vary by as much as that
> just due to airflow and temperature differentials.  You're not dealing
> with chemical engineering here.  Most of the time, if you're not drawing
> static sparks or condensing moisture, you're probably in the ball park
> on that one.

Michael is right about 5% being good enough for most applications.  The 
sensor I have is +/- 2%RH in the middle of its range (10-90% RH) with a 
1% hysteresis.  I can watch the humidity in the room change by 5% or 
more if I walk into the room right after taking a shower.  You don't 
need too much more than a general "wet/nice/dry" reading.

Now in my lab I've got a sensor that will tell me the temperature to 
within 0.01 Celsius but the price of that sensor is more than the 
reciprocal of its accuracy. :)