Just to put a cat amongst the pigeons, meteorologists measure the water 
vapour content of the air as dew point (I point out, not Dewey!) as 
being the only accurate method. This done with a thin metal mirror 
thermally attached to a Peltier junction. A beam of light impinges on 
the mirror at 45° and is reflected into a photodiode. A current passes 
through the Peltier junction to heat it until the detected light reaches 
a threshold value, then the current is reversed and the mirror is cooled 
at, say, 2°C/min until the light beam diminishes due to dew forming on 
the mirror and its temperature is accurately measured with a Pt100 
resistance thermometer attached to it. This method is very accurate. 
Conversion to RH is done through tables or an algorithm.

The next most accurate method is through a wet and dry bulb thermometer, 
using Pt100 thermometers, with the disadvantage that daily maintenance 
is required to ensure the reservoir is full of pure water and that the 
sock is clean and moist.

Then comes capacitive sensors, consisting of two electrodes separated by 
a hygroscopic dielectric. As water has an enormous dielectric constant 
(permittivity), the more moisture absorbed, so the value of the capacity 
increases.

The second least accurate method, and the one most commonly used for 
electrical measurement, you have the resistive sensor, where the 
resistance of a piece of plastic doped with a fairish quantity of an 
organic hygroscopic substance is measured. This drifts after a couple of 
years or so.

The least accurate and cheapest is the hair hygrometer where the length 
of a hygroscopic fibre (originally human hair but now synthetic) acts 
directly on the needle of a dial. This is not reliable or accurate.

I have a semi-pro weather system that uses capacitive sensors for both 
inside and outside. The dew point is claimed to be accurate to ±1.5°C 
between -76°C and +54°C, translating to ±3% from 0% to 100%.

No matter which system is used, it must be placed in an enclosure which 
allows free access of the air being measured with no risk of a current 
of air exceeding, say, 0.05 m/s. This is why double- or triple-louvred 
housings are commonly used for professional use. For the electronic 
types, these are usually shared with the temperature sensors, to avoid 
errors from direct solar or other heat source radiation and they are 
moulded in bright white plastic. I can say that, as I write at 10:00, 
the outside temp is 15.6°C ±0.5°C and the dew point is 7.9°C ±1.5°C, 
giving a RH of 58% ±3%. Fully professional weather systems would halve 
the accuracy tolerances and would cost 20 times more! Similarly, cheapo 
amateur systems can be obtained for 1/5th cost of mine and their 
accuracy tolerances are never mentioned! The ratio of 100:1 from a <$100 
amateur weather station to a >$10,000 professional one is very eloquent. 
This ratio can also apply to humidity sensors.

As has already been pointed out, the unit for absolute humidity is g/m³ 
and the conversion AH>RH is temperature dependent. For applications we, 
as production guys, are likely to need and encounter, the unit is 
percent relative humidity at a given temperature.

Brian

On 08/03/2013 18:23, Robert Kondner wrote:
> Hi,
>
>
>
> As I start looking at humidity measurement devices I started to wonder if
> IPC humidity numbers are Relative or Absolute? Does anyone know off hand. I
> need to go buy some specs.
>
>
>
> I would have thought Absolute but the cardboard indicators are temperature
> sensitive as are the desiccant drying systems in dry cabinets.
>
>
>
> If the temperature is high, say > 100C then a part will "Bake Out"
> regardless of the absolut humidity, is that right?
>
>
>
> Bob K.
>
>
>
> ______________________________________________________________________
> This email has been scanned by the Symantec Email Security.cloud service.
> For more information please contact helpdesk at x2960 or [log in to unmask]
> ______________________________________________________________________
>

______________________________________________________________________
This email has been scanned by the Symantec Email Security.cloud service.
For more information please contact helpdesk at x2960 or [log in to unmask] 
______________________________________________________________________