The right design for a relative humidity sensor system
Optimizing the response characteristics and accuracy of a humidity sensor system
To make the right choice when selecting a relative humidity sensor for an application, it is important to know and to be able to judge the deciding factors. In addition to long-term stability, which is a measure on how much a sensor changes its properties over time, these factors also include the measurement accuracy and the response characteristics of the sensor. Capacitive humidity sensors are based on the principle that a humidity-sensitive polymer absorbs or releases moisture as a function of the relative ambient humidity. Because this method is only a spot measurement at the sensor location, and usually the humidity of the surroundings is the desired quantity, the sensor must be brought into moisture equilibrium with the surroundings to obtain a precise measurement value. This process is realized by various transport phenomena (cf. the section titled "The housing effect on the response time"), which exhibit a time constant. Accuracy and response time are thus closely dependent on each other, and the design of a humidity measurement system becomes a challenge.
The term measurement accuracy of a humidity sensor is understood primarily to refer to the deviation of the value measured by the sensor from the actual humidity. To determine the measurement accuracy, references, such as chilled mirror hygrometers, whose own tolerance must be taken into account, are used. In addition to this trivial component, humidity sensors require a given time for reaching stable humidity and temperature equilibrium (the humidity is a function of temperature and decreases with increasing temperature; a difference between sensor and ambient temperature leads to measurement errors). This response time thus has a significant effect on the value measured by the sensor and thus on the determined accuracy.This time-dependent characteristic is explained in more detail in the following.
3Response characteristics and response time
The response characteristics are defined by various parameters. These are:
●The actual response characteristics of the humidity sensor at constant temperature.
(1) How quickly the sensitive polymer absorbs or releases moisture until equilibrium is reached (intrinsic response time)
(2) How fast the entire system reaches humidity equilibrium (housing effect)
●The thermal response characteristics of the humidity sensor at a non-constant temperature
(3) The thermal mass of the sensor
(4) The system's thermal mass, which is thermally coupled to the sensor (e.g. printed circuit board)
(5) Heat sources in the direct surroundings of the sensor (electronic components)
(1) and (3) are determined entirely by the sensor itself, (1) primarily by the characteristics of the sensitive polymer.
(2) and (4) are primarily determined by the construction of the entire system (shape and size of housing andreadout circuitry).
(5) is determined by heat-emitting electronic components.
These points will be discussed in more detail in the following.
The intrinsic response time (1)
Qualitatively, the response characteristics of capacitive humidity sensors look like the following (Fig. 1).
Fig. 1: Typical and idealized response characteristics of capacitive humidity sensors (schematic)
Because these response characteristics are especially pronounced at high humidity values,
an isothermal humidity jump from 40% to 100% was selected here for illustration. The desired ideal behavior of the sensor is indicated in blue. In practice, however, the sensor behaves according to the red line, approximately according to:
Here, the time span 1 is usually very short (typ. 1 – 30 min.), in contrast, the time span 2 is very long (typ. Many hours to days). Here the connection of measurement accuracy and response characteristics becomes clear (t until RH=100% is reached). The value at t4 (Fig. 1) is considered to be an exact measured value. However, this assumes that both the humidity and also the temperature remain stable during this entire time, and that the testing waits until this very long measurement time is completed. These conditions are both very hard to achieve and unusual in practice. For the calibration, there are the following two approaches, which both find use in practice (cf. Fig. 2):
1.The measured value at t2 (Fig. 1) is used as a calibration reference.
●The required measurement time for reaching the end value (in the example 100%) is
clearly shortened,corresponds to practice, and achieves an apparent short response
time of the sensor (cf. Fig. 2).
●If the conditions are similar for a long time (e.g., wet periods in outdoor operation),
the sensors exceed the correct end value (in the example 100%) undesirably by up
to 10% (cf. Fig. 2).
2. The measured value at t4 (Fig. 1) is used as a calibration reference.
●Even for similar conditions over a long time (e.g., wet periods in outdoor operation),
an exact measurement result is obtained (cf. Fig. 2).
●For a humidity jump like in Fig. 1, the sensors very quickly deliver the measured
value at t2, but reaching a stable end value (about 3-6% higher) takes a long time
(apparent longer response time)(cf. Fig. 2).
In order to take into account both approaches optimally, the measured values at t3 (cf. Fig. 1) are used as the calibration reference by Sensirion AG.
Fig. 2: Response characteristics of different humidity measurement systems
The housing effect on the response time (2)
Here, two types of transport phenomena play a deciding role:
●Convection: For this very fast process, the air, whose humidity is to be determined,
is transported to the sensor by means of ventilation.
●Diffusion: This very slow process is determined by the thermal, molecular
self-motion of the water molecules. It occurs even in "stationary" air (e.g., within a
housing), but leads to a long response time.
In order to achieve favorable response characteristics in the humidity measurement system, the very fast convection process must be supported by large housing openings and the slow diffusion process must be supported by a small housing around the sensor (small "dead
volume") with "stationary" air reduced to a minimum. The following applies:
Thermal effects (3), (4), and (5)
Because the total thermal mass of the humidity measurement system (sensor + housing)
has a significant effect on its response time, the total thermal mass must be kept as low as
possible. The greater the total thermal mass, the more inert the measurement system becomes
thermally and its response time, which is temperature-dependent, increases. In order to
prevent measurement errors, the sensor should not be mounted in the vicinity of heatgenerating components.
4Summary –what should be taken into account when designing a humidity measurement system
In order to achieve error-free operation of a humidity-measurement system with response times as short as possible, the following points should be taken into account especially for the selection of the sensor and for the design of the system.
●The selection of the humidity sensor element. It should
●be as small as possible,
●have a thermal mass that is as low as possible,
●work with a polymer, which exhibits minimal fluctuations in measured values during
the time span 2(cf. Fig. 1); testing gives simple information on this condition,
●provide calibration, which corresponds to the requirements (see above), e. g.,
SHT11/SHT15 from Sensirion.
●The housing design (cf. Formula 1). It should
●have air openings that are as large as possible in the vicinity of the sensor or the
sensor should be operated outside of the housing à good convection!
●enclose a "dead volume" that is as small as possible around the sensor àlittle
●The sensor should be decoupled thermally as much as possible from other components,
so that the response characteristics of the sensor are not negatively affected by the thermal inertia of the entire system.(e.g., its own printed circuit board for the humidity sensor, structurally partitioning the housing to create a small volume for the humidity sensor, see Fig. 3)
Fig. 3: Mounting example for Sensirion sensors SHT11 and SHT15 with slits for thermal decoupling
●The sensor should not be mounted in the vicinity of heat sources. If it was, measured
temperature would increase and measured humidity decrease.
The challenge is to realize a system that operates cleanly by optimally taking into account all of the points in section 4. The already calibrated SMD humidity sensors SHT11 and SHT15 from Sensirion are the ideal solution. For optimum integration of the sensors in a measurement system, Sensirion AG has also developed a filter cap as an adapter aid, which takes into account as much as possible the points in section 4 and also protects the sensor against contaminants with a filter membrane. Fig. 4 shows schematically how the sensors can be ideally integrated into a housing wall by means of the filter cap SF1.
Fig. 4: Filter cap for SHT11 and SHT15
In addition to the advantages mentioned above, there is also the option of building an IP67-compatible humidity measurement device (with O-ring, cf. Fig. 4) with optimal performance. Detailed information is available on the Sensirion Web site.
照“响应时间的壳体(1)效应”这一部分) 。这个湿度平衡过程可以通过各种用时间常数表征的传输现象得以实现 。测量精度和响应时间是如此地接近并且相互依赖,使湿度测量系统的结构设计成为一项挑战 。
湿度传感器的测量精度这个术语主要用于表述湿度传感器的测量值与实际湿度之间的偏差值。要确定测量精度,应参考使用冷镜式湿度计,同时必须考虑冷镜自身的误差范围。除了这个微不足道的部件以外,湿度传感器还需要给定一个达到湿度与温度稳定平衡的时间 。(湿度是温度的函数,湿度会随着温度升高而降低,传感器与被测环境之间的温差会导致湿度测量误差)更详细的说明请见下一节 。
2. 整个系统到达湿度平衡时所需要的时间快慢 。(壳体效应)
3. 传感器的热质量 。
4. 与传感器相关的系统热质量 。(例如印刷线路板的结构影响)
5. 传感器周围的直接热源 。(例如电子元器件发热影响)
1 和3 完全取决于传感器自身特性 。1 主要取决于敏感聚合体的特性。
2 和4 主要取决于整个系统的结构设计。(壳体的形状和尺寸设计以及输出电路设计)5 取决于电子元器件的发热量。
图 1 典型的和理想的电容式湿度传感器响应特性(示意图)
由于在高湿阶段其响应特性表现的特别明显,故选择了湿度从40%RH 到100%RH 的等温阶跃来说明。所期望的传感器的理想响应特性用蓝色虚线表示,而实际的响应特性用红线表示,其近似公式为:
这里,时间段1 通常非常短(大约1--30 分钟) 。相比之下时间段2 是很长很长的, (数小时至数天),测量精度与响应特性的关系在这张图上看得更清晰了(t 延续到湿度达到100%RH 时为止) 。t4 时间所对应的测量值是非常精准的。无论如何,得假定在整个t4测试时间段内湿度和温度都要保持稳定并且等候完成测试所需要的时间也很长 。在实践中这些不寻常的工作条件是很难实现的。实际的校准工作通常用以下两种方法(参见图2):
1. 以t2 时间对应的测量值作为校验参考基准;
2. 以t4 时间(参考图1)对应的测量值作为校验参考基准;
●对于像图1 那样的湿度跃升,传感器会很快在t2 时间达到相应的测量值 。但是,
要到达距其还有3%RH -- 6%RH的终端测量值还需要很长的时间 。(显然其响应
综合以上两种处理方法的优点,Sensirion AG 采用t3时间(参考图1)对应的测量值作为校验参考基准 。
由于湿度测量系统(传感器+壳体)总的热质量对其响应特性有着重要的影响,所以设计时必须尽可能地减小其热质量 。测量系统逐渐变热,受温度影响的的系统响应时间就会增加;系统总热质量越大,其惰性也越大。为了防止附加的测量误差,传感器不要安装在发热的电子元器件附近 。
4 . 总结—设计湿度测量系统时应该考虑的问题
● 具有湿敏聚合体结构;其在时间段2 上的测量值波动极小(参考图1),在测试条件下能提供简单的测试报告。
● 能提供校准以对应上述的各种需求,就像Sensirion SHTXX 系列传感器所能提供的那种校准 。
● 传感器周围空气流通口的尺寸要尽可能地大,或者直接将传感器臵于壳外 → 对流性最好封入套管内时传感器周围的“死区”要尽可能地小,→ 传导性最小 。