Data of the two fast telescopes are processed by two Fast DPUs (F-DPU). The data acquisition and handling of each exposure is similar to that of the N-DPU, except that the cadence is 2.5s instead of 25s; only imagettes will be processed; FGS algorithms are implemented to provide pointing error measurements with an accuracy better than 0.032 arcsec/√Hz and transmitted to the AOCS (Fine Guidance System: FGS).
In the nominal science observation mode, the main tasks of the F-DPUs are:
- To acquire and store CCD windows as a stream of pixels sent by the F-FEE from 4 CCDs with 451032255 pixels every 2.5 At the FEE level windowing is done to reduce the data volume transferred over SpaceWire. Full images are transmitted in calibration and test modes only. The raw data acquired from the ADC are just serialized and sent to the each N-DPU through 4 SpaceWire links.
- To reconstruct the star windows/imagettes (939 pixel windows / 2.5s).
- To send imagettes to the ICU (nominally for all stars in the P2 sample)
- To extract and compute offset, background and smearing data
- To send offset, background and smearing data to the ICU
- To calculate/estimate the attitude for the FGS/AOCS every 2.5s
- To collect HK data and send them to the ICU
Data products provided from Fast Cameras
|Data product (cadence)||Number|
|Imagettes (2.5s)||636 ([P2: 1 camera3586 + Guest observer: 1350] star windows)|
|Background values (50s)||100 (1 camera 3100 Background windows)|
|Offset values (2.5s)||8 (1 camera 3234 CCDs x 2255 pixels) offset windows|
|Smearing values (50s)||5724 (1 camera, 636 imagettes 39 pixels on 4 CCDs)|
|FGS data to S/C and ICU
|Quaternions, Covariance matrices, Status word(s)|
The FGS algorithms are based on an extended Kalman filter (EKF) used for recursive nonlinear optimisation. In full image mode with 4 links (one per CCD output) a peak data rate of 80 Mbps is transferred including the SpaceWire overhead. In order to cope with this bit rate with margin, the links are configured at 100 Mbps. In window mode a lower data rate for transferring imagettes windows only a much lower data rate from the F-FEE is required so that the links can be operated at 10Mbps.
The CPU load needed by the data acquisition, correction and reduction process is about 40% with a MDPA LEON2 FT processor running at 80 MHz. In total 512 Mbytes of SDRAM will be needed per F-DPU.
The FEU is an integrated electronics box, which consists of two data processing (F-DPU) boards (each within one module frame) and 2 power converters (PSU) integrated into a single frame.
The MEU/DPS team
The N-DPU SW is under responsibility of France. The N-DPU MEU is under responsibility of Spain. The FEU is under responsibility of Germany. The IAT DPS is under responsibility of Germany with contribution from Brazil.