AI-generated illustration for Galileo Whispers. A generic focal-plane concept inspired by time-delay integration; no exact Gaia hardware, moving luminous charge packets or measured timing is depicted.
Gaia collected starlight from drifting images by moving charge across its CCD detectors at a matching rate. In time-delay integration, a charge packet follows the image and keeps accumulating signal as both progress across the detector. The spacecraft can scan while a star contributes to one coordinated measurement rather than an ordinary long streak.
Moving charge is the key mechanism. It differs from pointing a telescope steadily at a star and from aligning separate photographs afterward. Gaia provides a historical engineering example of matching detector timing to a deliberately moving image.
Follow the image and its charge packet
A CCD converts absorbed light into collected electrical charge. During a scanning observation, a star’s image progresses through successive detector rows. If charge stayed fixed while the image moved, the signal would be spread over those rows during the exposure.
In time-delay integration, or TDI, the detector transfers charge in the same direction at an appropriate line rate. When the image advances by one row, the relevant charge advances too. Photons received at successive locations can then add to the packet following that image.
Crowley and colleagues’ Gaia detector paper, submitted in 2016, describes the synchronized line-transfer and spacecraft rotation rates. The design had 4,500 along-scan pixel rows and a nominal TDI line interval of 0.9828 milliseconds, giving about 4.42 seconds for a full CCD transit.
Multiplying 4,500 × 0.9828 milliseconds gives 4,422.6 milliseconds, or 4.4226 seconds. The rounded 4.42-second value is therefore a detector-transit timescale, not the time Gaia took to scan the entire sky.
Synchronization preserves concentration of the signal
The image velocity at the focal plane and the charge-transfer rate must agree closely enough for the intended image quality. Optical geometry connects the spacecraft’s angular scan to motion measured in detector pixels. Detector clocks then implement the corresponding transfers.
The image is not stopped in physical space. Its signal is gathered along a matched path. Finite transfer steps, the point-spread function, optical distortions, and imperfect transfer remain parts of the instrument’s response and calibration.
A simple invented mismatch shows the problem. Suppose an image moves at 100 pixels per second while charge is shifted at 99 pixels per second. After five seconds, their relative displacement is five pixels: (100 − 99) × 5. This is a constant-rate illustration, not a Gaia performance measurement.
Matching only the average rate is insufficient if substantial jitter or other motion occurs during the transit. TDI addresses predictable scan motion; it is not a universal correction for every pointing error.
Bright stars and moving asteroids need different treatment
ESA’s Gaia DR2 solar-system-object documentation explains that the nominal TDI rate was calibrated on stars. Shorter integrations could be selected for bright sources using gates, limiting accumulated charge rather than allowing every source a full-length exposure.
An asteroid can move relative to the stellar background. Its image therefore does not exactly follow the stellar charge-tracking rate. During a transit it can spread and drift away from the center of the small detector window assigned to it, affecting both the recovered light and position.
That is relative motion during the observation. It is different from a star’s slow catalog proper motion considered over years. The relevant question here is how far the source moves relative to the detector’s assumed tracking path during one measurement.
TDI is not postprocessing stacking
Image stacking combines separately recorded exposures after acquisition. TDI combines accumulating charge during the detector transit, before the final readout of that signal. Both exploit alignment, but at different stages and with different noise and calibration consequences.
Likewise, spacecraft scanning architecture determines how lines of sight move. TDI is a detector-level response to that motion; it does not replace the spacecraft’s attitude knowledge or control.
The broader lesson is that an exposure need not mean holding an image on one stationary pixel. A well-characterized moving image can be measured by moving its accumulating electrical signal with it—provided the geometry, timing, and limitations remain part of the measurement.