Historical Review

The Envisat RA-2 has the capability to provide limited bursts of individual, unaveraged echo sample data at the full PRF rate for research purposes. In this concept, the full rate data are stored, for a short burst, into an internal buffer memory, in parallel to the normal averaging and other functions of the instrument. The buffered data are subsequently read out at a much lower rate and appended to the normal science data.

These individual echoes data (only in phase and in quadrature components) are not processed onboard and therefore cannot be readily used for any geophysical application.
In this section we give you all the information required in order to understand the ad-hoc needed processing of the individual echoes, from their collection on-board and transmission to ground, to the way the processing on-board the altimeter is effectively emulated on ground for these echoes.

Here follows some basic information on the way the Individual Echoes are stored and sent by RA-2.

During nominal operation, RA-2 always transmits pulses 20 microseconds long at Ku-Band (13.575 GHz) with a Pulse Repetition Frequency (PRF) of 1795.332 Hz (at S-Band, 3.2 GHz, it becomes 448.833 Hz). Depending on the measurement mode the pulses can be un-modulated (Detection) or linear-frequency modulated (Acquisition and Tracking).

Under nominal operation (during tracking) RA-2 collects therefore 1795 echoes per second, but it transmits to ground only averages over groups of 100 echoes. The averages are transmitted to ground in Source Packets (SP) which contains 20 averaged echoes (Data Blocks) relevant to the transmission of 2000 pulses.

In order to allow the analysis on single (not averaged by 100) echoes (e.g. collected during tracking) it is possible to collect on-board the individual echoes and transmit them to ground (see Figure 1). This is possible in addition to and without interfering with the normal operations (e.g. Tracking, Preset Tracking). These echoes are only collected in Ku band, not in S band.

As can be seen in Figure 1, the individual echoes are buffered on-board before the Fourier Transform, and so they are still in the Time domain, and not in the Frequency domain as the normal RA-2 averaged waveforms.

Due to the high data rate of the full rate echo data (>4Mbps) the individual echoes are sampled in short bursts into a buffer memory which is read out at a lower rate into following source packets (see Figure 2).

This process has however some limitations. First of all the maximum number of individual echoes (128 samples each) that can be stored on-board (selectable in multiples of 100) is limited to 2000 (equivalent to 1 Source Packet) by the capacity of the on-board memory. Then, the time after which a new set of individual echoes can be stored is constrained by the time it takes to deplete the on-board memory buffer into the source packets. The latter have available for this operation 1600 words (16 bit) only. Furthermore, by macro command we can choose to store individual echoes in groups of 100 up to 2000 and transmit them to ground. The operation can only be repeated once the whole set of previously collected individual echoes is transmitted to ground.

This operation takes place in a cycle. There is one empty SP between cycles, regardless the length of the cycle (or number of on-board IE´s collected).

Table 1 shows all the possible lengths of the stored burst data. As an example; to collect 400 individual echoes it will take 0.2228 sec. They will be stored in 51200 words of the on-board memory. The transmission of these echoes to ground will need 32 source packets for a duration of 35.648 sec. This also implies that the collection of further 400 individual echoes cannot be repeated before 36.762 sec (35.648 + 1.114 seconds).