The immediate effect is that any command to the spacecraft can execute only after it has reached the spacecraft -- after the propagation delay. Consequently, the command can be verified as "okay" only once the relevant telemetry has come back, building in another propagation delay.
"All of this has to be modeled in our control software," says Accomazzo. "For example, when a command is sent it shall open the so-called 'verification window,' or the time during which it expects to receive a confirmation for a command at time-of-release of the command plus two times the propagation delay."
The telemetry that comes in is also time-stamped. When a return message is received, the reception time on-ground (known very precisely) must be stamped and the propagation delay subtracted in order to know the time when this telemetry was generated on the spacecraft.
"Since we need to know the generation time very precisely (less than 3 ms) then we need to model the propagation delay (i.e., know the distance between receiving antenna on ground and spacecraft) very well," Accomazzo stresses.
Because 3ms at 300,000 km/s (the speed of light) equates to a distance of 1,000km, the propagation delay has to be modeled for every single antenna used and the time of the day. Otherwise, significant errors could be introduced simply by not taking into account the position of the antenna in use on the surface of the Earth and its location.
As a case in point, if the assumption is made that Rosetta is fixed in space (which it is not) the rotation of the Earth alone causes an antenna to be 6,000 km closer or further from Rosetta in 12 hours.
Currently, at a twice-a-week interval, the team computes and groups a large set of instructions, send them to the on-board, time-tagged function, and then executes the command with the correct "y" time. But as the mission enters its critical phase, the frequency will increase.
"From an orbit control point of view the lander delivery in November is the most important phase of the mission, which means that we need greater accuracy in our predictions, so we will need to increase the frequency of the planning cycle with more frequent transmissions of commands to the spacecraft," says Accomazzo.
He says that in this phase engineers will have to take some "shortcuts" and acknowledges that they will, in essence, be running at the edge of risk, likening the situation to driving a car at 200 kph on a motorway.
"You can do it, but if you do it every single day you will eventually crash -- and that's exactly what we need to make sure doesn't happen. We know that we cannot run permanently at the edge of risk."
— Karen Field, Director of Content, EE Live and EE Times