It's not the heat, it's the humidity

From Automotive Test Report, a T&MW publication,
November 2002

Dan Romanchik, Editor, Automotive Test Report -- Test & Measurement World, 11/10/2002

Many automotive tests require a controlled ambient humidity. Vehicles must often operate under extremely humid conditions, such as the Congo or even New Orleans in July.

To test vehicles under these conditions, engineers use environmental test chambers that can control not only the ambient temperature, but also the humidity. This is essential both for tests that require a controlled humidity and also for tests that do not allow condensation at low temperatures. The only way to prevent this condensation is to control the humidity as well as the temperature.

The standard range for most humidity chambers is 10% to 95% RH. Some manufacturers will outfit their chambers with a steam generator or a dehumidifier to allow you to go to 98% RH and down to 5% RH. Accuracy is generally around ±4%, which meets the needs of most automotive tests. You can achieve tighter tolerances, but only over a smaller humidity range.

The Problem
While a specified accuracy of 4% is sufficient for most automotive environmental tests, how can you be sure that your chamber is actually that accurate? According to Ken Durbin, Calibration Manager for Cincinnati Sub-Zero, Cincinnati, OH, humidity is a difficult parameter to measure, and there are several problems with the way test chambers are designed and with the way they are used.

Figure 1 A handheld humidity meter can help you check the accuracy of your test chamber's humidity control system.

First, while most large environmental test chambers may have three or four temperature sensors, they normally have only one humidity sensor. The reason for this is purely economical—thermocouples are relatively cheap, while humidity sensors cost hundreds of dollars. The problem is that if something goes wrong with the single humidity sensor, there is no backup—the chamber's humidity control system depends entirely on the input of a single sensor.

The other big problem test chamber users face is one they create for themselves. Durbin says that when running high humidity test, most chamber users run the chamber flat out. This type of operation saturates the transducer, normally some type of capacitive sensor. Once a test is complete, the technicians may remove the item under test, but normally do not return the chamber to a more normal humidity. This further saturates the sensor and will most likely damage.

Durbin likens this phenomenon to wetting a piece of cardboard. Pour water on a piece of cardboard and it will absorb a lot of that water. You can dry it out later, but that piece of cardboard will no longer have the same shape or physical properties that it did before it was soaked. In the case of the humidity sensor, the sensor will always indicate a humidity that's higher than the actual humidity.

If you're not sure whether your chamber is controlling the humidity properly, one of the first steps you can take is to use a humidity-measuring instrument to measure the chamber's humidity. Several manufacturers make handheld devices that you can take into a chamber (Durbin likes the Rotronic HygroPalm series, shown in Figure 1). Comparing the measurements taken on the handheld to the readings on your chamber's control system will give you an indication if you have a problem or not.

Making Chamber Measurements
If you do have a problem, it's time to calibrate the humidity control system. To do this you must first measure the “as-is” condition. Use a standard humidity probe—which generally will has accuracy of 0.5% or less—and record the humidity readings at several points in the measurement range. If the readings are within spec, then you can proceed to the next step. If they are not, then you need to check the calibration of the sensor.

If the “as-is” measurements are within spec, the next step is to control the system without the sensor. To do this, you connect a simulator to the control system that looks to the system like an ideal humidity sensor. You set the simulator to output voltages that the ideal sensor would output and record the readings on all display devices and the output voltages of the controller.

If these readings are within spec, you're done. If not, then you must make the necessary adjustments and re-run the first set of measurements that you made to ensure that the entire system, including the sensor, is still within spec.

Calibrating the Sensor
If you must calibrate the sensor, there are several different approaches you can take. The first is to simply send it to a lab for calibration. According to Jim Glover, president of Graftel in Rolling Meadows, IL, a good cal lab should be able to calibrate and return your sensor within three working days. By calling ahead, however, you may be able to arrange a shorter turnaround time at no extra cost.

Glover suggests that when looking for a lab, you ask what type of calibration standards they use, and don't choose a lab if they rely on saturated salt standards alone. Instead, look for labs that use either a chilled-mirror hygrometer or a two-pressure generator. He also suggests choosing a lab with a lot of experience in humidity calibration. The reason is that humidity calibration is much more difficult than other types of calibration, such as temperature or weight, and experienced personnel will more likely produce a good calibration.

If you want to calibrate the sensor yourself, there are a couple of methods that you can use. One of the most popular methods is to use saturated salts humidity standards. To use saturated salts, you need to purchase a number of ampoules containing specific salt solutions. You'll also need a small, airtight chamber.

To perform the calibration, break the ampoule inside the chamber, insert the sensor, and then seal it. Once the environment stabilizes—this usually takes 30 to 45 minutes—the relative humidity inside the chamber will be a known value, and you can calibrate the sensor to that value. Salt solutions are available that will generate an RH of 0%, 5%, 10%, 11.3%, 20%, 35%, 50%, 65%, 75.3%, 80%, and 95%.

Durbin notes that one of the drawbacks of using this method is that the salts are corrosive, so once a calibration is complete, you must wash everything very thoroughly. Another disadvantage of this approach is that it's sensitive to both the ambient temperature and to temperature uniformity within the chamber. That is one of the reasons for using a small container as the calibration chamber.

Another method is to use a chilled-mirror hygrometer. These units have a mirror, which is chilled in small temperature increments. A small LED shines on the mirror, while a photodiode receives the reflected light. When water vapor condenses on the mirror, it scatters the light from the LED, reducing the amount of light reaching the photodiode. The temperature at which this happens is the dew point, and using this reading and measurements, or air temperature and pressure, you can calculate the relative humidity.

This is perhaps the most accurate method of measuring humidity—they can measure the dew point from -80º C to +85º C with an accuracy of ±0.15º C. It is, however, the most costly method as well. Chilled mirror hygrometers can cost $8,000 or more.

Whatever method you choose, Durbin suggests calibrating your system every six months. He notes that many companies try to stretch this calibration cycle to one year in an effort to save time and money, but that this is not a wise thing to do. Discovering a temperature chamber that is out of calibration could cause you to re-run a lot of expensive tests, which will end up costing you even more in the end.