 RECALIBRATION AND ANALYSIS OF THE PHOENIX RELATIVE HUMIDITY SENSOR DATA.  E. Fischer1, G. M. Martínez1 and N. O. Rennó1, 1University of Michigan, Department of Climate and Space Sciences and Engineering, 2455 Hayward St., Ann Arbor, MI 48109, USA. erikfis@umich.edu   Introduction:  Here, we show initial results of our work in producing high-level relative humidity (RH) data from measurements made by the Thermal and Electrical Conductivity Probe (TECP). Due to resource limitations, the preflight calibration conditions of the TECP [1] only partially overlapped the environmental conditions found at the Phoenix landing site [2, 3, 4] and only the raw TECP RH experiment data records are available in the NASA Planetary Data System (PDS). We are working on recalibrating the flight unit using a spare engineering unit in analogous conditions simulated in our Michigan Mars Environmental Chamber (MMEC). This will allow us to obtain and archive high-level RH data from the Phoenix mission. Low-Level RH Data Available in the PDS:  The low-level TECP RH data stored in the PDS is stored in binary format, with each binary file consisting of a cluster of un-calibrated data records, typically in a time-series, grouped together by virtue of being acquired on the same sol with the same token, product type and record length. We have successfully converted the entire set of TECP RH binary files into one single readable file. In particular, we have converted into ASCII format values of the raw output of the TECP RH sensor (RHADC) and the temperature of the TECP board (Tb) returned from Mars, along with the sol number and the Local True Solar Time (LTST) at which these measurements (RHADC and Tb) were taken. These results are shown in Fig. 1. Laboratory Setup:  We use the MMEC to simulate Martian surface conditions for the sensor recalibration [5]. We have designed and built a structure inside the chamber to ensure good thermal contact between the chamber's thermal plate and the TECP, while maintaining insulation from the (warmer) chamber walls. An electrical feedthrough connects the TECP sensor inside the chamber to the TECP Interface unit outside of the chamber. Preliminary tests showed that the TECP sensor temperature converged to the thermal plate temperature within an error of < 0.5 K within a few minutes depending on the set temperature. The interface unit is controlled via Labview code and a custom software interface. Initial Experimental Results:  The original calibration of the TECP RH sensor flown to Mars was done at the University of Washington Mars Atmospheric Simulation Facility [6]. It was performed using a pair of frost point hygrometers (a Buck CR-1 chilledmirror hygrometer and an EdgeTech DewPrime I chilled-mirror hygrometer) as a reference. A calibration function Tf = f (RHADC, Tb) was produced, where Tf is the frost point temperature measured by the reference hygrometers [1]. Due to resource limitations, values of RHADC and Tb covered in the original calibration only partially overlap the environmental conditions found at the Phoenix landing site, and therefore the function Tf = f (RHADC, Tb) could not be extrapolated with enough certainty to such un-calibrated conditions.   Fig. 1. (Top) Complete set of measurements of the temperature of the TECP board (Tb) returned from Mars sorted out as a function of the local true solar time (LTST) and sol number (color bar). (Bottom) Same as in left figure, but for the raw output (counts) of the TECP RH sensor (RHADC).  We have conducted a set of laboratory experiments to: 1. Test the Engineering TECP probe for measurement stability and offset in the measured board temperature.  2. Check for consistency in the raw output (RHADC) between the TECP unit flown to Mars (Fig. 1, bottom) and the Engineering TECP. We have conducted the first experiments to check the TECP probe for a possible drift in the measured Tb values at constant environmental conditions. No such drift was observed over time frames of multiple hours. Furthermore, power cycling the probe showed no resulting shift in measurements either. Next, we analyzed the response of Tb to changes in the thermal plate temperature. We found that depending on the amplitude of the temperature change in the thermal plate, Tb in the TECP probe followed within minutes, up to about 30 minutes for tested sudden variations of 20 K. After Tb converged to the plate temperature, the stability in Tb measurements was within the amplitude of thermal plate temperature variations (< 0.1 K). Next, we checked for consistency in the raw output (RHADC) between the Engineering TECP unit and the TECP unit flown to Mars. Specifically, we have conducted initial experiments to simulate the same environmental conditions in terms of Tb and Tf as in the original calibration with the expectation of obtaining similar RHADC values from both TECP units. In both units, the RH sensor output voltage is measured with a 12-bit ADC. However, whereas the flown unit presents a range of 140 units in RHADC values between 2800 at RH ~ 0% and 2940 at RH ~ 50 % (highest calibrated value), our Engineering TECP unit shows a similar amplitude in RHADC but shifted to higher values. This is shown in Fig. 2. This offset is just within the maximum expected variability between different sensors according to the manufacturer. We are conducting preliminary experiments to obtain a relation between the RHADC values obtained by the flight unit and by the engineering unit under comparable conditions. We are simulating in the MMEC the complete set of environmental conditions in terms of Tb and Tf covered during preflight calibration with the goal of a functional relationship to convert RHADC values measured by the flown TECP unit to RHADC values measured by the Engineering TECP (Fig .3). Using this relationship between the flight and engineering unit outputs at the same conditions, we will methodically cover the entire range of conditions observed on the Martian surface (gray markers in Fig. 2) and use the RHADC output of the sensor to obtain a new calibration function.  References: [1] Zent, A. P. et al. (2009) J. Geophys. Res., 114, E00A27. [2] Rennó, N. O. et al. (2009) J. Geophys. Res., 114, E00E03. [3] Davy, R. et al. (2010) J. Geophys. Res., 115, E00E13. [4] Tamppari, L. K. et al. (2010) J. Geophys. Res., 115, E00E17. [5] Fischer, E. et al. (2014) Geophys. Res. Lett., 41, 4456-4462. [6] Cobos, D. et al. (2010) TECP Calibration Report. http://pds-geosciences.wustl.edu/phx/phx-m-meca-4-nir dr-v1/phxmec_1xxx/calibration/meca_tecp_calib_repo rt.pdf   Fig. 2. Low-level RH data returned from Mars (gray), preflight calibration (red) and MMEC re-calibration (blue). RHADC values measured by the Engineering TECP model in our MMEC (at Tb ~ 231 K and RH between 0 and 50%) show consistently higher values compared to those measured at the same Tb and RH conditions during preflight calibration.   Fig. 3. Values of TECP board temperature (Tb) and frost point (Tf) simulated during preflight calibration for the flown TECP unit (red) and in the MMEC (blue). The goal is to obtain a functional relationship to convert RHADC values measured by the flown TECP unit to RHADC values measured by the Engineering TECP by simulating in the MMEC the complete set of environmental conditions in terms of Tb and Tf covered during preflight calibration. 
