Analog And Digiltal Data
JD-EE and others
Permit an old software designer to make some comments. For your knowledge
I am no pilot, the closest I have come is by designing and debugging software for high altitude rockets at Esrange. Those programs were only running once and for a short time - any problem and the rocket would self-destruct.
Air data (pressute, temperature, etc) in the A330 are measured as close to the ports as possible and immediately digitized (some types of very accurate probes even present their value initially as a number). From the schematics available for the A330, this primary air-data handling is done by the ADM-units - one for each probe. The ADM-units are connected to the ADIRU-units by signal cabling. On these cables data are requested and/or transmitted in digital form using a standard protocol. The use of a common bus reduces the amount of wiring needed.
The tranfer of data can usually operate in three ways:
Clocked mode: A central system clock triggers data transfer at given times between the various units via the common data bus. Each unit then has its own time window for communication
Polled mode: A unit higher in the system polls a unit lower in the system. The lower ranked unit then responds with either a "no data" answer or sends the requested data. Thus the higher unit controls the use of the common bus.
Interrupt mode: A lower unit can send a "data available" signal on a dedicated interrupt line to request attention. Also in this case the higer unit controls the use of the bus. However, I don't think this mode is used on large aircraft because of the extra wiring needed to carry unique interrupt information to and from all units and the rather slow variation in measured data probably doesn't require interrupt mode.
In none of these cases is any higher ranked computer "scanning" analog signals. However, such scanning may occur in some simple on/off cases, e.g. the postion of various switches. This will need additional wiring,
As far as I have understood, the A330 systems don't operate in a clocked mode. This makes debugging of software much more difficult than in a clocked system. An old proverb states that "given the right data all computer systems will give the wrong result". Regardless of all testing and verification, some strange bugs may still hide in the code. There is at least one published example of this kind of bug found hidden in several versions of code used in an ADIRU unit during a long time.
Regards