Here is a spreadsheet for computing link budgets in a hard-limiting bent-pipe transponder. Many of the input numbers currently shown are just for testing, so take the results with a grain of salt.
You are encouraged to play with the input numbers. The most interesting results are down near the bottom of the sheet: the data rate supported by the link and the loss relative to regeneration.
The loss relative to regeneration is the loss in received S/No due to the use of a bent-pipe transponder instead of one that demodulates and decodes the uplink data and then re-encodes and remodulates it on the downlink. This loss comes from two sources: the increased noise floor at the downlink receiver due to repeated uplink noise, and the transponder output power that is wasted on this noise.
Although regeneration is more efficient than bent-pipe repeating, it is considerably more complex, especially when (as here) highly power-efficient modulation and coding schemes are used.
One possibility is to use a less power efficient modulation method on the uplink that is easier for the satellite to decode. It could then encode the uplink data in the more power-efficient mode for the downlink. Most power-efficient modes are relatively easy to generate even if they are complex to demodulate.
This would be indicated when the loss of the bent pipe relative to regeneration is greater than the Eb/No loss due to the use of the less efficient uplink scheme. For example, QPSK with concatenated RS/Viterbi FEC at a rate of about 0.4 can operate at an Eb/No of about 3 dB, while simple noncoherently demodulated FSK requires an Eb/No of about 13 dB. So for those two modes, the regenerating scheme would be better whenever the bent-pipe loss relative to regeneration exceeds about 10 dB.
Last updated: 23 July 2001