1 Introduction

Total solar irradiance (TSI) has been identified as an important factor influencing the Earth’s climate system. Continuous measurements of the TSI have been conducted on various space platforms since November 1978 (Fox et al. 2011; Frohlich 2011), including the Earth Radiation Budget (ERB) experiment on the NIMBUS-7 spacecraft (Hoyt et al. 1992), the Active Cavity Radiometer Irradiance Monitor (ACRIM I) experiment on the Solar Maximum Mission (SMM) (Wilson 1997), the ACRIM II experiment on the Upper Atmosphere Research Satellite (UARS), the Variability of Solar Irradiance and Gravity Oscillations (VIRGO) experiment on the Solar and Heliospheric Observatory (SOHO) (Frohlich et al. 1997), and the Total Irradiance Monitor (TIM) experiment on the Solar Radiation and Climate Experiment (SORCE) (Kopp and Lawrence 2005; Kopp, Heuerman, and Lawrence 2005).

The FY-3 TSI database has been constructed from the data obtained since 2008 with the Total Solar Irradiance Monitor (TSIM) on the Feng Yun-3 series satellites (FY-3; Feng means wind in Chinese, Yun means cloud in Chinese; Wang and Wang 2012). The FY-3 series satellites are Chinese meteorological polar-orbit satellites (Yang et al. 2012). FY-3A, the first satellite of the series, was launched on 27 May 2008. FY-3B, the second satellite of the series, was launched on 5 November 2010. TSIM/FY-3A recorded the TSI daily from June 2008 to December 2011. The FY-3 TSI database is further extended by overlapping observations by TSIM/FY-3B from November 2010 to the present.

The TSIM/FY3-B instrument has almost the same design as TSIM/FY-3A, except for some software improvements for the onboard components. The TSIM is installed on the leading surface of the satellite, with its opening parallel to the flying direction of the satellite. Neither FY-3A nor FY-3B make the TSIM point accurately toward the Sun. In addition, the TSIM has no solar tracking device (Wang, Wang, and Fang 2011). Thus, the Sun is only observed when the satellite leaves the Earth’s shadow near the terrestrial North Pole and sunlight sweeps the TSIM field of view (Fang et al. 2002).

The TSIM was calibrated to World Radiometric Reference through ground-based comparison experiments (Wang, Li, and Fang 2014). Absolute radiometers SIAR-1a and SIAR-2c were used as transfer standard, which had been calibrated to World Radiometric Reference by World Standard Group TSI instruments in the tenth International Pyrheliometer Comparison (IPC-X), 2005.

2 Design

The instrument overview of the TSIM is shown in Figures 1 and 2. The electronics of the TSIM is shown in Figure 3. The instrument consists of the three absolute radiometers AR1, AR2, and AR3, as illustrated in Figure 2. The three radiometers and an amplifier for the temperature signal are integrated into the instrument, while other electrical systems for the TSIM are implemented in the electronics package. Some parameters of the TSIM are presented in Table 1.

Figure 1
figure 1

An overview of TSIM/FY-3B. It consists of the three absolute radiometers AR1, AR2, and AR3.

Full size image
Figure 2
figure 2

Schematic drawing of the instrument. The three absolute radiometers and an amplifier for the temperature signal are integrated into the instrument.

Full size image
Figure 3
figure 3

The electronics package of TSIM/FY-3B. It includes among others units of A/D conversion, communication with the satellite, and precision voltage reference.

Full size image
Table 1 Parameters for TSIM/FY-3B.
Full size table

The three absolute radiometers take independent measurements or simultaneous measurement of the TSI on command. The radiometers AR1 and AR2 are designed for regular observation of the TSI, while the radiometer AR3 is designed for occasional observation to investigate the degradation of AR1 and AR2. The three radiometers are aligned with different inclinations with respect to the mounting surface of the TSIM to increase their fields of view (Liebetraut et al. 2013; Memarian and Eleftheriades 2013; Witte et al. 2014). The angle between the optical axis of each radiometer and the satellite X-axis is given in Table 1.

2.1 Absolute Radiometer

Each absolute radiometer is constructed based on the principle of electrical substitution (Wilson 1979). The radiometer has two inverted cone cavities, as shown in Figure 4. One is the primary cavity, the other the reference cavity for compensation. The two cavities are installed together in a single aluminum cylinder. The cylinder serves as a heat sink to stabilize the thermal system of the radiometer. A precision aperture is located in front of the primary cavity. The precision aperture was fabricated carefully, and its area has been measured by repeated experiments. A shutter unit and an aperture for view limiting are located at the top of the heat sink to control sunlight input to the primary cavity (Booth 2014; Girshovitz and Shaked 2014). Several apertures are mounted between the precision aperture and the view-limiting aperture to prevent stray light from entering the primary cavity.

Figure 4
figure 4

Schematic drawing of the absolute radiometer.

Full size image

The primary cavity is a silver cone cavity with thin walls. The thickness of the cone wall is 0.06 mm. The internal surface of the primary cavity is coated with a thin layer of absorptive black paint. The reflectivity of this paint is nearly uniform in the wavelength range of the instrument. The effective absorption coefficient of the cone cavity is 0.9997. The heating wires are embedded into the thin walls of the primary cavity using electroplating. Nearly all the power produced by the heating wires is absorbed by the cone cavity with little heat loss, thanks to this method.

A ring with a radius of 13 mm, surrounded by 180 pairs of thermocouples, is laid at the opening bottom of the primary cavity. The reference cavity also has a ring of thermocouples. The hot junctions of the thermocouples are connected to the primary cavity, while the cold junctions are connected to the heat sink. The ring of thermocouples is used as a passive temperature sensor to detect the temperature difference between the primary cavity and the heat sink. The reference cavity receives no sunlight and is not electrically heated either. The ring of thermocouples of the reference cavity is connected in series to the ring of thermocouples of the primary cavity. The purpose of the series connection is to compensate for any variation in the output from the ring of thermocouples that is caused by a temperature change in the heat sink.

2.2 Electronics

A block diagram of the TSIM electronics is given in Figure 5. The electronics package includes an amplifier for the temperature signal, an A/D converter for the temperature signal, three D/A converters for electrical substitution, three driving circuits for the shutters, three driving circuits for heating the primary cavity, a communication unit with the satellite, a precision voltage reference, a remote measurement unit, a micro-controller unit, and others. The three radiometers have the same circuits as the A/D conversion, communication with the satellite, precision voltage reference, and the micro-controller. Each radiometer has independent circuits for D/A conversion, shutter driver, and heating driver. A Zener-based integrated circuit LM199H was selected as the precision voltage reference for electrical substitution. The voltage applied to the primary cavity for heating is always calibrated by LM199H. This has excellent temperature stability over the operating temperature range of the instrument, namely −55 ∘C to 125 ∘C and for different operating current conditions; its typical long-term stability is \(8~\mbox{ppm}/\sqrt{\mathrm{kH}}\). Changes in voltage with temperature are almost eliminated by integrating the output from an active Zener diode on the monolithic substrate of LM199H.

Figure 5
figure 5

A block diagram of the TSIM electronics. Most of the TSIM electronics is implemented in this electronics package.

Full size image

When sunlight is fed to the radiometer, the temperature of the primary cavity is measured by the temperature sensor, that is, by the ring of thermocouples. The temperature of the primary cavity is further converted into a digital signal by a 16 bit A/D converter. The digital temperature signal is then sent to the micro-controller. The power of the incoming sunlight is estimated by the micro-controller, and then an equivalent electrical heating power is determined. The estimation method is described in detail in Section 4. Conversely, when the shutter is closed, the power for electrical substitution is calculated and a proper voltage is generated by the D/A converter. The heating voltage is applied to the heating wires through amplification in the driving circuits for the heating of the primary cavity. Specific experiment data are collected and stored by the micro-controller as remote sensing data. The remote sensing data are then sent to the data management system of the satellite through the communication unit.

2.3 Measurement Sequence

When no operation is performed, as described in Section 3, a constant electrical power P m is always applied to the primary cavity to keep it at an almost constant temperature T c in the orbit.

The value of the TSI is calibrated once in two subsequent phases: the observation phase and the reference phase (Wilson 1979). In the observation phase, the shutter is open and sunlight is allowed to fall onto the primary cavity. Less electrical power is needed as a result of the absorbed energy of the incoming sunlight. In the reference phase, the shutter is closed and more electrical power is applied to keep the primary cavity at an almost constant temperature. Because the applied electrical power can be measured precisely, an accurate measure of the TSI is available (Wilson 1979; Kopp and Lawrence 2005; Mekaoui et al. 2010).

In the observation phase, a voltage u eo is applied to the primary cavity. After the temperature of the primary cavity stabilizes at a constant value T o, the absorbed radiant power P s is given by

$$ P_{\mathrm{s}} = E \cdot A \cdot \alpha, $$
(1)

where E is the solar irradiance, A is the area of the precision aperture, and α is the effective absorption coefficient of the primary cavity.

In the reference phase, voltage u er is applied to the primary cavity to stabilize the temperature of the primary cavity again at T o. When the stable temperature of the primary cavity in the reference phase is the same as the stable temperature in the observation phase, the relationship

$$ P_{\mathrm{s}} + \frac{u_{\mathrm{eo}}^{2}}{R} = \frac{u_{\mathrm{er}}^{2}}{R} $$
(2)

holds, where R is the resistance of the heating wires for the primary cavity.

When the TSIM works in the operation modes that are introduced in Section 3, the temperature of the primary cavity is sampled every 5 s by the 16 bit A/D converter.

3 Operation Modes

The TSIM works continuously in the orbit, and the value of the TSI is observed once every orbit by the radiometers. The radiometers have three operation modes: a test mode, a solar observing mode, and a background mode.

The temperature of the primary cavity presented in this section is given as a binary number from the A/D converter in units of the least significant bit (LSB) of its output. The temperature of the primary cavity is not transformed into kelvin for simplicity.

3.1 Test Mode

When the TSIM needs to be checked in the orbit, such as in the power-on sequence of the instrument, the TSIM enters the test mode. The electrical power as illustrated in Figure 6 is applied to the primary cavity. The temperature record of the primary cavity in the test mode is shown in Figure 7. The sequence of procedures followed during the test mode is listed below:

Figure 6
figure 6

Electrical heating power profile of the test mode.

Full size image
Figure 7
figure 7

Temperature response of the primary cavity in the test mode. The data were obtained by TSIM/FY-3B in space at 1:02 UT, 11 November 2010. This is the first remote-sensing data set obtained by TSIM/FY-3B. The temperature of the primary cavity is given in units of the LSB of the A/D converter output.

Full size image

1) Close the shutter of the radiometer, apply a low power P L to the primary cavity for 6 min, and the temperature of the primary cavity stabilizes at a temperature T L.

2) Keep the shutter of the radiometer closed, apply a high power P H to the primary cavity for 5 min, and the temperature of the primary cavity stabilizes at a temperature T H. The value of P H is equal to the maintaining electrical power P m.

3) Compute the sensitivity of the primary cavity of the radiometer s pt by

$$ s_{\mathrm{pt}} = \frac{P_{\mathrm{H}} - P_{\mathrm{L}}}{T_{\mathrm{H}} - T_{\mathrm{L}}}. $$
(3)

The sensitivity indicates how much electrical power is required to increase the temperature of the primary cavity by a specified amount. The sensitivity s pt is shown in Figure 8 in units of W/LSB.

Figure 8
figure 8

Sensitivity (s pt) of the radiometer AR2 in TSIM/FY-3B from 11 November 2010 to 10 November 2013. Some data gaps occured (e.g. in July 2011) because the instrument occasionally stopped working as a result of electronics problems.

Full size image

4) Produce a remote-sensing data set of the radiometer in the test mode.

3.2 Solar Mode

When sunlight is visible within the TSM field of view, the instrument enters the solar mode. The electrical power as illustrated in Figure 9 is applied to the primary cavity. The temperature response of the primary cavity in the solar mode is shown in Figure 10. The sequence of procedures of the solar observing mode is listed below:

Figure 9
figure 9

Electrical heating power profile of the solar mode in the observation and the reference phases.

Full size image
Figure 10
figure 10

Temperature response of the primary cavity in the solar mode. The data were obtained by radiometer AR1 of TSIM/FY-3B in space at 1:51 UT, 11 November 2010. This was the first sunlight for TSIM/FY-3B. The temperature of the primary cavity is given in units of the LSB of the A/D converter output.

Full size image

1) Open the shutter of the radiometer and apply a power P soe to the primary cavity for 6 min. How to determine the power P soe is explained in detail in Section 4. The temperature of the primary cavity stabilizes at temperature T so.

2) Close the shutter of the radiometer for 5 min and apply a power P sre to the primary cavity to stabilize the temperature again at T so. When the temperature of the primary cavity stabilizes at the end of the reference phase, the value of E (TSI) is obtained from Equations (1) and (2) as

$$ E = \frac{P_{\mathrm{sre}} - P_{\mathrm{soe}}}{A \cdot \alpha}. $$
(4)

3) Produce a remote-sensing data set of the radiometer in the solar mode.

When a remote-sensing data set is produced in the solar mode, the shutter of the radiometer is opened to detect the sunlight. If the temperature of the primary cavity exceeds a given value, it means that sunlight enters the field of view of the radiometer again, and the radiometer will start the next TSI measurement.

3.3 Background Mode

When the shutter of the radiometer is open in the solar mode, the radiometer emits radiative power to space. The radiative power exchange between the cosmic background and the radiometer is given by

$$ \varphi_{\mathrm{rb}} = \int_{S_{\mathrm{r}}} \int_{S_{\mathrm{b}}} \frac{ ( L_{\mathrm{b}} - L_{\mathrm{r}} )\cos \theta_{\mathrm{b}}\cos \theta_{\mathrm{r}}}{r^{2}} \,\mathrm{d}S_{\mathrm{b}}\,\mathrm{d}S_{\mathrm{r}}, $$
(5)

where φ rb is the radiative power from the cosmic background to the primary cavity of the radiometer, L b is the radiance of the cosmic background, L r is the radiance of the radiometer, S r is the surface of the precision aperture of the radiometer, S b is the surface of cosmic background (the two surfaces are joined by a light ray of length r), θ b is the angle between the light ray and the normal to the surface S b, and θ r is the angle between the light ray and the normal to the surface S r. From Plank’s law and Equation (5), we obtain

$$ \varphi_{\mathrm{rb}} = \frac{\sigma ( T_{\mathrm{b}}^{4} - T_{\mathrm{m}}^{4} )}{\pi} \int_{S_{r}} \cos \theta_{\mathrm{r}}\,\mathrm{d}S_{\mathrm{r}}\int_{0}^{\omega} 2\pi \cos \theta_{\mathrm{b}}\sin \theta_{\mathrm{b}}\,\mathrm{d} \theta_{\mathrm{b}}, $$
(6)

where σ is the Stefan–Boltzmann constant (σ=5.670×10−8 W m−2 K−4), T b is the temperature of the cosmic background (T b=2.72548 K), T m is the temperature of the primary cavity, and 2ω is the field of view of the radiometer.

For a light ray leaving the primary cavity, we can set

$$ \theta_{\mathrm{r}} = 0. $$
(7)

Then, Equation (6) becomes

$$ \varphi_{\mathrm{rb}} = \sigma A \bigl( T_{\mathrm{b}}^{4} - T_{\mathrm{m}}^{4} \bigr)\sin^{2}\omega, $$
(8)

from which follows

$$ I_{\mathrm{rb}} = \sigma \bigl( T_{\mathrm{b}}^{4} - T_{\mathrm{m}}^{4} \bigr)\sin^{2}\omega, $$
(9)

where I rb is the irradiance from the cosmic background to the radiometer.

For the TSIM, T m=300 K, ω=13.3∘, and we obtain

$$ I_{\mathrm{rb}} = \sigma \bigl( T_{\mathrm{b}}^{4} - T_{\mathrm{m}}^{4} \bigr)\sin^{2}\omega = - 24.3059~\mbox{W}\,\mbox{m}^{ - 2}. $$
(10)

Here the irradiance I rb is measured in the background mode for the correction of the TSI data.

The heating power profile applied to the primary cavity in the background mode is shown in Figure 11, and the temperature response of the primary cavity in the background mode is shown in Figure 12. The sequence of procedures followed during the background mode is as follows:

Figure 11
figure 11

Electrical heating power profile of the background mode in the observation and the reference phases.

Full size image
Figure 12
figure 12

Temperature response of the primary cavity in the background mode. The data were obtained by radiometer AR1 of TSIM/FY-3B in space at 2:08 UT, 11 November 2010. This was the first measurement of cosmic background by TSIM/FY-3B. The temperature of the primary cavity is given in units of the LSB of the A/D converter output.

Full size image

1) Ten minutes after the TSIM receives the information that the satellite is in the Earth shadow, the shutter of the radiometer is opened. A power P coe is applied to the primary cavity for 6 min. The temperature of the primary cavity gradually decreases as a result of the radiative power from the primary cavity to the cosmic background, and finally stabilizes at T co.

2) Close the shutter of the radiometer for 5 min and apply a power P cre to the primary cavity to stabilize its temperature again at T co.

3) Compute the radiative power from the cosmic background to the radiometer and produce a remote-sensing data set of the radiometer for the background mode.

4 Estimation of Required Electrical Power

A stable temperature of the primary cavity is helpful for quick measurements of TSI values. Our experiments have proved that when the temperature of the primary cavity varies substantially, more time is generally necessary to complete a single measurement. In addition, the measurement accuracy is influenced by the temperature variations. The primary cavity of the TSIM radiometer is always expected to be maintained at an almost constant temperature. However, the precise electrical power required in the observation phase is not known in advance. It needs to be determined consistently and carefully, especially in the solar mode. The method for estimating the required electrical power is described in this section.

When a radiometer operates in the observation phase of the solar mode, the electrical power is reduced to zero as soon as the shutter is open. The temperature of the primary cavity decreases from a constant temperature T c. If a sufficiently long time t has elapsed in the observation phase, the temperature of the primary cavity will stabilize at T so, as described in Section 3. As a function of t, the temperature of the primary cavity will vary as

$$ T(t) = (T_{\mathrm{c}} - T_{\mathrm{so}})\mathrm{e}^{ - \frac{t}{\tau}} + T_{\mathrm{so}}. $$
(11)

For given times t 1 and t 2 (t 2>t 1), we obtain

$$\begin{gathered} T(t_{1}) - T_{\mathrm{so}} = (T_{\mathrm{c}} - T_{\mathrm{so}})\mathrm{e}^{ - \frac{t_{1}}{\tau}}, \end{gathered}$$
(12)
$$\begin{gathered} T(t_{2}) - T_{\mathrm{so}} = (T_{\mathrm{c}} - T_{\mathrm{so}})\mathrm{e}^{ - \frac{t_{2}}{\tau}}. \end{gathered}$$
(13)

Equation (12) divided by Equation (13) gives

$$ T_{\mathrm{so}} = \frac{T(t_{1}) - T(t_{2})\mathrm{e}^{\frac{t_{2} - t_{1}}{\tau}}}{1 - \mathrm{e}^{\frac{t_{2} - t_{1}}{\tau}}}. $$
(14)

The power of the incoming sunlight is approximated as

$$ P_{\mathrm{sa}} = \frac{T(t_{1}) - T(t_{2})k}{1 - k}s_{\mathrm{pt}}, $$
(15)

where k is given as

$$ k = \mathrm{e}^{\frac{t_{2} - t_{1}}{\tau}}. $$
(16)

The electrical power P soe required in the observation phase of the solar mode is found as

$$ P_{\mathrm{soe}} = P_{\mathrm{m}} - P_{\mathrm{sa}}. $$
(17)

Likewise, the electrical power P sre required in the reference phase of the solar mode is given by

$$ P_{\mathrm{sre}} = T_{\mathrm{so}}s_{\mathrm{pt}}. $$
(18)

The electrical power required in the background mode can be derived in a similar manner as discussed above.

As proved by the remote-sensing data set shown in Figures 10 and 12, the temperature of the primary cavity is maintained at a nearly constant value using the method presented in this section.

5 Comparison Experiment

The traceability of TSIM/FY-3B to the World Radiometric Reference (WRR) was achieved through ground-based comparison experiments in the air, at ambient temperature and ambient pressure (Wang, Li, and Fang 2014). No vacuum devices were used in the instrument calibration. Solar Irradiance Absolute Radiometer (SIAR) type absolute radiometers SIAR-1a and SIAR-2c were used as transfer instruments in the comparison experiments. (A similar process was also carried out for TSIM/FY-3C.) The duration of a single TSI measurement in the comparison experiment is 10 min; 5 min each for the observation phase and the reference phase (Wang, Li, and Fang 2014). To measure the Sun together, the three instruments TSIM/FY-3B, SIAR-1a, and SIAR-2c were pointed toward the Sun simultaneously using solar tracking devices.

SIAR-1a and SIAR-2c had previously been calibrated to WRR by World Standard Group (WSG) TSI instruments at the tenth International Pyrheliometer Comparison (IPC-X) in 2005. According to this calibration, the calibration factors to WRR are 1.001928 for SIAR-1a and 1.000016 for SIAR-2c. The calibration factors of the TSIM in the comparison experiments are presented in Table 2.

Table 2 Results of the comparison experiment for TSIM/FY-3B.
Full size table

6 Evaluation of Measurement Uncertainties

The uncertainty in TSI measured with each absolute radiometer in the TSIM is modeled from Equation (4) as follows:

$$ E = f(u_{\mathrm{er}},u_{\mathrm{eo}},R,A,\alpha ) = \frac{u_{\mathrm{er}}^{2} - u_{\mathrm{eo}}^{2}}{R \cdot A \cdot \alpha}, $$
(19)

where u eo and u er are the heating voltages in the observation and the reference phases (see Equation (2)). The combined standard uncertainty in E is

$$\begin{aligned} u_{\mathrm{c}}^{2}(E) =& \biggl( \frac{\partial f}{\partial u_{\mathrm{er}}} \biggr)^{2}u^{2}(u_{\mathrm{er}}) + \biggl( \frac{\partial f}{\partial u_{\mathrm{eo}}} \biggr)^{2}u^{2}(u_{\mathrm{eo}}) \\ &{} + \biggl( \frac{\partial f}{\partial R} \biggr)^{2}u^{2}(R) + \biggl( \frac{\partial f}{\partial A} \biggr)^{2}u^{2}(A) + \biggl( \frac{\partial f}{\partial \alpha} \biggr)^{2}u^{2}(\alpha ) \\ =& 4 \biggl( \frac{u_{\mathrm{er}}}{R \cdot A \cdot \alpha} \biggr)^{2}u^{2}(u_{\mathrm{er}}) + 4 \biggl( \frac{u_{\mathrm{eo}}}{R \cdot A \cdot \alpha} \biggr)^{2}u^{2}(u_{\mathrm{eo}}) + \biggl( \frac{u_{\mathrm{eo}}^{2} - u_{\mathrm{er}}^{2}}{A \cdot \alpha \cdot R^{2}} \biggr)^{2}u^{2}(R) \\ &{}+ \biggl( \frac{u_{\mathrm{eo}}^{2} - u_{\mathrm{er}}^{2}}{R \cdot \alpha \cdot A^{2}} \biggr)^{2}u^{2}(A) + \biggl( \frac{u_{\mathrm{eo}}^{2} - u_{\mathrm{er}}^{2}}{R \cdot A \cdot \alpha^{2}} \biggr)^{2}u^{2}(\alpha). \end{aligned}$$
(20)

The uncertainties in the parameter values of the absolute radiometers and the combined standard uncertainty in E are summarized in Tables 3, 4, and 5. Most of the uncertainties for individual parameters were obtained from independent repeated observations, while some were obtained from the manufacturer’s specification, such as the datasheets of the A/D converter and the precision voltage reference.

Table 3 Uncertainties in the absolute radiometers AR1 and TSIM/FY-3B.
Full size table
Table 4 Uncertainties in the absolute radiometers AR2 and TSIM/FY-3B.
Full size table
Table 5 Uncertainties in the absolute radiometers AR3 and TSIM/FY-3B.
Full size table

The ground experiments for uncertainty evaluation were performed in the test mode of the TSIM as described in Section 3.1. It was found that the light resources used in the ground tests of TSIM/FY-3B, such as the laser (Rubenchik, Fedoruk, and Turitsyn 2014), were always changing. As a result, no light sources were used in the experiments for uncertainty evaluation to investigate the real characteristics of the TSIM. In the test mode of the instrument, the heating voltages u er and u eo were set by the micro-controller unit of the TSIM electronic system and were repeated as often as necessary. In the solar mode, it was quite hard to repeat the heating voltages for the primary cavity because of changes in the light sources. Therefore, the radiometer performance of light-to-heat conversion could not be evaluated.

Uncertain correction terms for TSI science results, such as corrections for traceability, Sun–Earth distance, satellite velocity relative to the Sun, and solar pointing errors were neglected in this uncertainty evaluation.

7 Summary

The FY-3 TSI database has been constructed from data from the TSIM/FY-3A and TSIM/FY-3B instruments obtained from June 2008 to the present. Daily observations of the TSI with traceability to SI have been performed by the TSIM, using three absolute radiometers. The performance of the TSIM has been improved in the following points: 1) The heating wires are embedded into the walls of the primary cavity and are not wrapped on its surface, to reduce loss of heating power. 2) Radiative power exchange between the cosmic background and the absolute radiometer is measured in the background mode every orbit, and they are identified as compensation terms in the correction of TSI data. 3) To quickly perform TSI measurements, the required electrical power for the heater is estimated by calculation. This helps to maintain the primary cavity at an almost constant temperature, especially in the observation phase of the solar mode.

For the remaining tasks, the in-flight data from TSIM/FY-3A are known to suffer from solar pointing errors due to some data problems of the satellite. This is expected to be solved in the future. A preliminary analysis of the in-flight TSI data obtained with TSIM/FY-3B will be presented in an accompanying paper (Wang et al. 2014). A complete reduction of the TSIM/FY-3B data incorporating the orbital position of the satellite requires a substantial development of analysis software and may take time.