Sensor Specification Statement and Testing Guide

Understanding and testing the specifications of RHT sensors

Abstract
Reading specifications of relative humidity sensors can often be unsatisfying as there is no common agreed
standard for specifying humidity sensors. Statements may therefore be misleading, testing sensors against
specification can require clarification, and comparing specifications of different sensors is often challenging. To
alleviate the issue, this document details how Sensirion humidity and temperature sensors specifications are to
be understood. Furthermore, methodologies for testing the sensors against various statements in the
specifications are given, such that the user can test the sensors accordingly.

Applicability
This document is applicable to all SHTxx humidity and temperature sensors from Sensirion.

Contents
1 Specifications ............................................................................................................................................. 1
1.1 General considerations ................................................................................................................... 1
1.2 Accuracy - Calibration..................................................................................................................... 2
1.3 Accuracy - Hysteresis ..................................................................................................................... 2
1.4 Accuracy - Long-term drift .............................................................................................................. 3
1.5 Short-term stability .......................................................................................................................... 3
1.6 Extreme conditions ......................................................................................................................... 3
1.7 Non-linearity.................................................................................................................................... 3
1.8 Response time ................................................................................................................................ 3
1.9 Supply voltage ................................................................................................................................ 4
1.10 Current and energy consumption ................................................................................................... 4
2 Testing ........................................................................................................................................................ 4
2.1 Preparation ..................................................................................................................................... 4
2.2 Testing RH Accuracy ...................................................................................................................... 5
2.3 Testing T accuracy ......................................................................................................................... 5
2.4 Testing RH response time .............................................................................................................. 5
3 Revision History ......................................................................................................................................... 7

1 Specifications simplification will of course leave some grey zones,

which are not scientifically covered. But the resulting
1.1 General considerations specifications should give the user a reliable tool to
Sensor components detecting relative humidity understand the sensors and integrate them into his
consist of a sensing element – a polymer in most own devices.
cases – which absorbs and desorbs water molecules, Accuracy of relative humidity sensors can be
depending on the surrounding conditions. The characterized by three different, rather independent
process of gaseous molecules entering and leaving terms: calibration accuracy, hysteresis and long-term
solid material is quite sophisticated and bears many drift. Beyond this, short-term stability, extreme
special effects. conditions behavior (i.e. in very cold, hot, humid or
To make humidity sensors applicable to commercial dry environments), and non-linearity are also to be
products, these special effects must be brought to a considered.
simple understanding, which can be tested with
standard, commonly available equipment. Such

© Copyright Sensirion AG, Switzerland 1/8

1.2 Accuracy - Calibration 1.3 Accuracy - Hysteresis
Calibration accuracy is the main component of an The hysteresis value is the difference of measured
accuracy specification. It provides information on values of the same sensor at a certain log point
deviation of the individual sensor readings in accruing from dry environment on one hand and
equilibrium state against a high precision reference humid environment on the other hand – given
at the time of calibration (which is understood to be enough dwell time. In other words, humidity sensors
about the time of sale). Major causes for tolerance carry some memory of conditions experienced in
on calibration accuracy are in-batch variation (e.g. their recent past. Sensors with a dry history carry
homogeneity of conditions within calibration some negative offset while sensors with a humid
chamber), batch-to-batch variation, precision of history carry some positive offset. Hysteresis is due
calibration reference, and stability of sensors. to the composition and design of the sensor element.
Calibration accuracy is measured against a dew
As the hysteresis does not depend on the quality of
point mirror i.e. a high precision reference. Thus, the
calibration but is dependent on the exposure range
user should be able to reproduce the value. Sensirion
while in application, this value is understood to be
specifies calibration accuracy with two different
additional to calibration accuracy.
parameters: the maximal and the typical accuracy.
2.0
Accuracy distribution: For an ensemble of sensors

Deviation vs. Reference (%RH)

1.5 hysteresis
such as a production lot, the variation of the
measured deviation against the reference follows a 1.0
normal distribution. It is characterized by an average 0.5
value μ and a standard deviation σ. 0.0
-0.5
μ±2σ range -1.0
Number of sensors

-1.5
-2.0
10 20 30 40 50 60 70 80 90 100
Reference (%RH)
Figure 2 Example for hysteresis measurement. A path
from dry to humid and one from humid to dry is measured
(full dots on graph). Dwell time at each log point is 40min.
-6 -4 -2 0 2 4 6 Open dots represent the mean values i.e. calibration
Number of standard deviations σ accuracy.
from the mean μ
Calibration accuracy and hysteresis values are
Figure 1 Distribution of sensor deviation around average.
The μ±2σ range is targeted to fit within the typical
determined by running the sensor in a full humidity
accuracy limits, while the μ±4σ range is targeted to fit loop 15% → 30% → 50% → 70% → 90% → 90%
within the maximal accuracy limits. → 70% → 50% → 30% → 15% with dwell times of
40 minutes at each log point. For the determination
Typical accuracy: The accuracy distribution of the of calibration accuracy at a certain humidity value,
ensemble up to μ±2σ is targeted to fit within the the mean value is calculated from the measured
typical accuracy limits. values of the ascending and the descending path.
Maximal accuracy: The accuracy distribution of the The difference between the measured values and the
ensemble up to μ±4σ is targeted to fit within the mean values determines the hysteresis.
maximal accuracy limits. This is equivalent to From a sample of sensors an average value and
targeting a process capability index Cpk ≥ 1.33. A standard deviations of the various calibration
one-hundred percent end test and a sample test accuracy values can be determined. With these
ensure that only sensors satisfying the maximal values compliance of typical limits can be checked
accuracy limits qualify for sales. (see Figure 3).
Accuracy limits are specified for the full range of
supply voltage unless stated otherwise. RH
accuracies for different temperatures are stated in
the respective sensor datasheets.

© Copyright Sensirion AG, Switzerland 2/8

 4 The sensor with largest drift during such an exposure
3
compliant not compliant divided by the calculated exposure at 25°C
determines the long-term drift value per year. The
Accuracy Deviation

2
specified value additionally contains some margin to
1
compensate for modelling assumptions and for
0
higher base temperatures.
-1
typical limit 1.5 Short-term stability
-2
maximal limit
-3 Short term stability may be characterized by
-4 repeatability – repeated measurements with the
Figure 3 Examples for accuracy distributions same sensor in identical conditions. The measure for
complying (left side) or not complying (right side) such term is the standard deviation for the sample of
with the specified accuracy limits. Open dots repeatability measurements.
represent average values. Error bars denote the 1.6 Extreme conditions
μ±2σ range and are targeted to remain within
typical limits. Capacitive humidity sensors undergo a reversible
drift at very humid and very dry conditions. In case of
Sensirion humidity sensors, and at relative humidity >
1.4 Accuracy - Long-term drift 80%RH – and the higher the stronger – the sensor
reading creeps to higher values. The reference value
The aging of the sensor element may lead to drift of is given for constant exposure to 90%RH for 60 hours.
the measured value compared to reference. Such a The same is true at very dry conditions – such as
long-term drift is about random – it may move to the reflow soldering – where the sensor reading creeps
upper or lower side or may change direction in the to lower values. A re-hydration procedure is then
course of time. The long-term drift value is a maximal required to take the sensor back to its normal reading
limit for such drift per year. state.
In the case of Sensirion, long term drift is determined 1.7 Non-linearity
by exposing a sample of sensors to High
Temperature Operating Lifetime (HTOL) with This term stands for systematic deviations outside
operation at 125°C for 408 hours. This exposure is the calibrated log points. Such deviations of the
equivalent to aging at 25°C for a much longer sensor output as well as temperature compensation
duration, which can be calculated with the following may be corrected by a linearization formula. Hence
formula: non-linearity may be made very small by using the
−1
appropriate formula. Non-linearity values are
E  1 1  included by the accuracy tolerance and shall not be
tT0 = tT1  exp a   −   considered as an additional term.
k  T1 T0  
1.8 Response time
T1 corresponds to the higher temperature (125°C in
the example) and T0 to the lower temperature (25°C). For response times, Sensirion specifies a so-called
Note that these temperatures must be expressed in τ63% (read “tau 63%”) or equivalently τ1/e time. For a
Kelvin in the equation above. t1 and t0 stand for the sensor exposed to an abruptly changing environment
durations (hours) spent at T1 and T0 respectively. k is (step function of measured physical value), the
the Boltzmann constant (8.61*10-5 eV/K) and Ea is the sensor reading approaches the final value typically
thermal activation energy. on an exponential function over time. The τ63% time
extends from the moment of the environmental
With Ea ~ 0.65eV for degradation within CMOS change at the sensor until the sensor reading
structures and Ea ~ 0.75eV for hydrolytic degradation reaches 63% of the step height (see Figure 4).
within the humidity sensor element, one gets that
storage at 125°C for 408 hours corresponds to 27 When testing response time, it is important that all
and 71 years spent at 25°C, respectively for CMOS other parameters, except the one to be tested,
structures and the sensing element. Note that aging remain constant. Also, one must ensure that there is
within humidity sensor elements is very complex and no dead time between the step function initiated by
difficult to model. Activation energies are chosen as the system and the sensor experiencing the step
the lowest among various processes and therefore function.
leads to the worst-case consideration.

© Copyright Sensirion AG, Switzerland 3/8

For temperature measurements the thermal mass the average power consumption through the
and thermal conductivity of the substrate play an constant supplied voltage VDD.
important role. Therefore, temperature response time
Note that additional power is consumed during start-
values are specified very vaguely.
up phase or during operation of other sensor
100% functionalities.
90%
80%
Relative Humidity

70%
2 Testing
60% τ63%
50% 2.1 Preparation
40%
63% For calibration and end testing of RHT sensors,
30%
20% Humidity Setting
Sensirion is using expert calibration equipment and
10% Response Data
procedures guaranteeing highest precision and
0% reliability. To ensure consistent testing of Sensirion
-10 0 10 20 30 40 SHTxx relative humidity & temperature sensors the
Time (s) following preparation items should be carefully
Figure 4 Measurement profile of response time testing considered:
for relative humidity. a) Test Objects: The test series should consist of at
least 5-10 sensors, which should be taken out of
1.9 Supply voltage the original packaging.
The supply voltage (VDD) range is defined with an b) Conditioning: Make sure that sensors have not
upper and a lower limit, plus a typical value. Any been contaminated prior to the testing. Original
supply voltage in that range may be used for packaging should not have been fixed with
continuous operation. Absolute maximum voltages scotch tape, and packaging should not have
that may be applied for a limited time only are been stored in plastic bags. If you are unsure
specified in the datasheet of the respective sensors. whether sensors might have been contaminated
The typical value defines the supply voltage at which or not, follow the re-conditioning procedure (80-
the sensors are calibrated and at which quality 90°C [176-194°F] at < 5%RH for 24h (baking)
control is performed. followed by 20-30°C [70-90°F] at > 74 %RH for
48h (re-hydration)).
1.10 Current and energy consumption
c) Re-hydration after soldering: In case sensors
During operation the sensor pulls a certain supply have been soldered to a PCB board please
current (IDD). This current is different in idle state make sure that the re-hydration procedure (20-
than while measuring/communicating. Furthermore, 30°C [70-90°F] at >74%RH for 48 hours) has
in an ensemble of sensors there is a certain variation been applied after soldering.
in current consumption. The average of the d) Test set-up: Make sure that test and reference
ensemble is specified as typical value, while sensor experience identical humidity and
minimum and maximum values define the lower and temperature conditions. If possible, make use of
upper limits. a professional humidity chamber. If such a
The power consumption (P) is calculated as P = IDD chamber is not available to you, put the
* VDD with IDD the supply current and VDD a specified reference and test sensor into a closed box and
constant supply voltage. give the set-up time to homogenize. Please
make also sure that there are no humidity
To determine the average value of power absorbing materials (silicone sealing, rubber,
consumption over time, we consider a specific etc.), nor contaminating materials present near
measurement resolution and measurement the sensor. More information on contamination
frequency (indicated in the datasheet of the can be found in the info sheet “Handling
respective sensor). The average power consumption Instructions”.
is then the power consumption averaged between
e) Reference sensor: The reference sensor should
the time spent measuring and the time spent in idle
be of high reliability. If possible, use a dew point
state over a measurement interval. The average mirror or recently calibrated RH probes.
supply current over time for the specified
measurement resolution and frequency is related to

© Copyright Sensirion AG, Switzerland 4/8

2.2 Testing RH Accuracy temperature accuracy of SHTxx sensors, additional
improvements to the measurement set-up need to be
It is recommended to run the following test profile:
applied. Consider especially point e) of the
Set temperature to 25°C (77°F) and run a low –
preparation guidelines: the best practice is to use a
medium – high RH profile (e.g. 15% – 30% – 50% –
reference which is much more accurate than the
70% – 90%) and check hysteresis with a high –
sensors (ideally by a factor 10). To make sure that
medium – low RH profile (e.g. 90% – 70% – 50% –
the sensors measure the same temperature as the
30% – 15%). See Figure 5. Allow at least 40 minutes
reference, and therefore to make the measurements
settling time at each relative humidity set-point,
valid, consider the following best practices:
before starting a measurement. For the first set point,
a longer waiting time might be required to allow the a) Thermal coupling: Attach the sensors and
temperature to stabilize (~3 hours for temperature the reference to a highly temperature
changes <30°). conducting material. This can be achieved
easily by attaching them to a piece of metal
100% such as copper. More thermal mass is
90%
advantageous.
Measured RH

80%
70%
60% b) Homogeneity: Place the sensors and the
50% reference close to each other to limit the
40%
30% influence of air flows or heating/cooling
20% effects inside the test chamber. Additionally,
10% use a measurement chamber which is as
0%
small as possible. This can be achieved by
15%

30%

50%

70%

90%

90%

70%

50%

30%

15%

putting a small measurement box in a

RH set point profile bigger climate chamber.

Figure 5 Measurement profile for accuracy testing (full c) Settling time: Similar to the testing of RH
squares: reference, open diamonds: test sensor). accuracy, a long settling time is required
before measurement to let the system reach
0.2% equilibrium (~3 hours for temperature
0.0% changes below 30°). When choosing the
RH Deviation

-0.2%
-0.4%
settling time, due consideration should be
-0.6% given to the thermal masses involved in the
-0.8% measurement setup.
-1.0%
-1.2% 2.4 Testing RH response time
-1.4%
-1.6% For properly measuring RH response times a test
set-up should guarantee a step function from dry air
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0%

to humid air at the very same temperature, or vice

RH set point profile versa. Therefore, the test set-up in Figure 7 is
Figure 6 RH deviation profile, full line: ascending order, recommended. The sensor is placed into a climate
hatched line: descending order. chamber whose volume must be very small to allow
for fast environmental change dynamics. A rotary
It is also recommended to operate the SHTxx valve controls the air intake of the chamber, enabling
sensors with the Sensirion SEK Evaluation Kit to to switch between dry air and humid air. It is a crucial
avoid communication or conversion errors. element for creating a step function in environmental
IMPORTANT: Relative humidity is highly condition. Dry air is taken from a compressed, oil-free
temperature dependent. During the measurement, air tank – while humid air is taken from a humidity
the temperature of reference and test sensor must chamber. As the compressed air cools down with
show the same value to make RH values comparable. expansion, the temperature shall be brought to room
For your test records it is highly recommended to temperature by a heater. The heater alternatively can
note RH and T for each set-point and sensor. be placed between air bottle and rotary valve or
between rotary valve and test chamber. The tubing
2.3 Testing T accuracy between the valve and the test chamber shall be kept
For testing temperature accuracy, the same general as short as possible. This allows for an abrupt
procedures apply as for RH. Due to the very high change from dry to humid air. A flow rate of the dry

© Copyright Sensirion AG, Switzerland 5/8

air and humid air of at least 1 m/s is recommended to temperature remains constant. To avoid any
avoid that the setup is limiting the response time. temperature changes it is recommended to run the
measurement at ambient temperature.

Test Compressed air

The data should behave roughly inverse exponential
Humidity (see Figure 4). The SHTxx sensors readings should
Chamber
Generator reach 63% of the full RH step within the τ63% of the
sensor. For example, if the RH step is between
RH = 90% 10%RH and 90%RH, then the RH value shall
read >60%RH within τ63% seconds. Note that the last

RH = 5%
T = 20°C T = 20°C percentage points of the response curve to complete
the full step will develop with a lower time constant.
Rotary
Valve There is an alternative option to measure response
time in a very simple and effective way – which,
Exhaust however, may not fulfill scientific standards: put the
sensor into a humidity chamber with high humidity
Figure 7 Response time test set up. Dry air from (90%RH for example) or if no humidity chamber is
compressed air is blown into test chamber with sensor –
available, put the sensor into a closed box with an
a heater makes sure the air is at room temperature. With
a rotary valve the relative humidity is taken from a open glass of water. Make sure that the temperature
humidity chamber and therefore abruptly changed from inside the chamber or box is equal to the outside
low to high. temperature. Start measuring (for example with the
Sensirion SEK Evaluation Kit) and when the values
To start the measurement, blow dry air into the are high and stable remove the sensor from the box
sensor box and give the sensor time to display a or chamber and expose it to the outside conditions.
constant value. It should be in the range of 5 - 10% Give the sensor enough time to stabilize the RH
RH. Control the temperature carefully as well. Then value and in this way define the full RH step. The
change the air source abruptly and make sure the response time can then be derived from the data.

© Copyright Sensirion AG, Switzerland 6/8

3 Revision History

Date Revision Changes

18. March 2010 1.0 Initial release
07. June 2018 1.1 Update for new sensors
19. April 2021 2 Merged testing guide into document; visual and text update

© Copyright Sensirion AG, Switzerland 7/8

Important Notices
Warning, Personal Injury

Do not use this product as safety or emergency stop devices or in any other application where failure of the product could result
in personal injury. Do not use this product for applications other than its intended and authorized use. Before installing,
handling, using or servicing this product, please consult the data sheet and application notes. Failure to comply with these
instructions could result in death or serious injury.

If the Buyer shall purchase or use SENSIRION products for any unintended or unauthorized application, Buyer shall defend, indemnify and
hold harmless SENSIRION and its officers, employees, subsidiaries, affiliates and distributors against all claims, costs, damages and
expenses, and reasonable attorney fees arising out of, directly or indirectly, any claim of personal injury or death associated with such
unintended or unauthorized use, even if SENSIRION shall be allegedly negligent with respect to the design or the manufacture of the
product.
ESD Precautions
The inherent design of this component causes it to be sensitive to electrostatic discharge (ESD). To prevent ESD-induced damage and/or
degradation, take customary and statutory ESD precautions when handling this product.
See application note “ESD, Latchup and EMC” for more information.
Warranty
SENSIRION warrants solely to the original purchaser of this product for a period of 12 months (one year) from the date of delivery that this
product shall be of the quality, material and workmanship defined in SENSIRION’s published specifications of the product. Within such
period, if proven to be defective, SENSIRION shall repair and/or replace this product, in SENSIRION’s discretion, free of charge to the
Buyer, provided that:
• notice in writing describing the defects shall be given to SENSIRION within fourteen (14) days after their appearance;
• such defects shall be found, to SENSIRION’s reasonable satisfaction, to have arisen from SENSIRION’s faulty design, material, or
workmanship;
• the defective product shall be returned to SENSIRION’s factory at the Buyer’s expense; and
• the warranty period for any repaired or replaced product shall be limited to the unexpired portion of the original period.
This warranty does not apply to any equipment which has not been installed and used within the specifications recommended by
SENSIRION for the intended and proper use of the equipment. EXCEPT FOR THE WARRANTIES EXPRESSLY SET FORTH HEREIN,
SENSIRION MAKES NO WARRANTIES, EITHER EXPRESS OR IMPLIED, WITH RESPECT TO THE PRODUCT. ANY AND ALL
WARRANTIES, INCLUDING WITHOUT LIMITATION, WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR
PURPOSE, ARE EXPRESSLY EXCLUDED AND DECLINED.
SENSIRION is only liable for defects of this product arising under the conditions of operation provided for in the data sheet and proper use
of the goods. SENSIRION explicitly disclaims all warranties, express or implied, for any period during which the goods are operated or
stored not in accordance with the technical specifications.
SENSIRION does not assume any liability arising out of any application or use of any product or circuit and specifically disclaims any and
all liability, including without limitation consequential or incidental damages. All operating parameters, including without limitation
recommended parameters, must be validated for each customer’s applications by customer’s technical experts. Recommended
parameters can and do vary in different applications.
SENSIRION reserves the right, without further notice, (i) to change the product specifications and/or the information in this document and
(ii) to improve reliability, functions and design of this product.
Copyright © 2021, by SENSIRION. CMOSens® is a trademark of Sensirion. All rights reserved
Headquarters and Subsidiaries
Sensirion AG Sensirion Inc., USA Sensirion Korea Co. Ltd.
Laubisruetistr. 50 phone: +1 312 690 5858 phone: +82 31 337 7700~3
CH-8712 Staefa ZH [email protected] [email protected]
Switzerland www.sensirion.com www.sensirion.com/kr

phone: +41 44 306 40 00 Sensirion Japan Co. Ltd. Sensirion China Co. Ltd.
fax: +41 44 306 40 30 phone: +81 3 3444 4940 phone: +86 755 8252 1501
[email protected] [email protected] [email protected]
www.sensirion.com www.sensirion.com/jp www.sensirion.com/cn

Sensirion Taiwan Co. Ltd

phone: +886 3 5506701
[email protected]
www.sensirion.com
To find your local representative, please visit www.sensirion.com/distributors

© Copyright Sensirion AG, Switzerland 8/8