12 ms·
Interesting that they use rate 1/6 Turbo Codes for communication with the probe, i.e. for any given data block, 1/6 is the data and 5/6 is error-correction info
by udev 11y ago
Interesting that they use rate 1/6 Turbo Codes for communication with the probe, i.e. for any given data block, 1/6 is the data and 5/6 is error-correction information.
This might seem inefficient, but it allows for reliable communication even in cases when the SNR is low, i.e. when the signal is faint compared to the noise, and multiple bits in the packet are inverted.
- tzs 11y agoThe signal is indeed faint [1]. With their 70m dish in Australia, they are getting a received power of 400 zW [2], or in dBm units -154 dBm. Note that this the power they are getting after the gain from a freaking 70m dish. [1] https://www.reddit.com/r/amateurradio/comments/3d93en/the_radio_signal_from_the_new_horizons_spacecraft/ https://www.reddit.com/r/amateurradio/comments/3d93en/the_ra... [2] z is for zepto, the SI prefix for 10^-21.
- soneil 11y agoI think it's not just the SNR involved, but the time. With a round-trip of nearly 9 hours, a TCP-style "I'm sorry, I missed that?" response is hugely impractical. The higher the penalty for re-transmission, the more effort goes into getting it right the first time.