Apparently they don’t do it exactly. The CJ4 (Proline) gets close.
It’s an academic question but the implementation is interesting.
By “exactly” I mean flying an exact circle which joins the two tracks. In zero wind the roll angle would be constant all the way round. In some wind, the roll angle would vary.
Put it another way: does your jet, Jason, fly geometrically perfect holds, or does it fly the outbound turn at Rate 1? I am told by airline pilots that their systems don’t fly the “perfect” hold if there is wind.
Peter wrote:
Put it another way: does your jet, Jason, fly geometrically perfect holds, or does it fly the outbound turn at Rate 1? I am told by airline pilots that their systems don’t fly the “perfect” hold if there is wind.
Holds are irrelevant as these aircraft are often limited on bank angle so can’t fly a perfect oval, but it flies a near perfect intercept of final in a strong crosswind.
Peter wrote:
With GPSS (also called roll steering) the GPS transmits a digital data stream (ARINC429) which contains a roll angle, which the GPS adjusts according to where in the turn you are.Roll steering is not necessarily digital. E.g. in a G1000 + KAP140 setup, the G1000 gives analog roll steering commands to the KAP140. (And, no, the G1000 doesn’t fake roll steering using the heading deviation input of the A/P like a roll steering converter does. The KAP140 has separate roll steering inputs and roll steering is used in NAV mode, not HDG mode.)
The ARINC signal is bank angle and the GPS can only use whatever inputs it has available to it. However, the GPS has no knowledge of the autopilot, so it has no way of knowing if the autopilot is rate limited or bank angle limited or the type of roll steering converter being employed. So I suspect the limits are in the autopilot and the roll steering increases to some internal limit not known by the GPS which has track, GS, desired track, maybe IAS, Temp, heading, pressure altitude available to it.
As far as track accuracy is concerned, if the aircraft is flown at a speed above which the autopilot can command an adequate rate of turn or bank angle, there may be some overshoot, but no where near the error introduced by pilots. With GPSS, the track is being continuously updated at a 5 HZ rate on a WAAS unit, and once the course is captured, it will track center line and be correcting long before there is even a fraction of a needle width deviation.
The holding pattern shown on a GNS/GTN with heading and air data input will not be an oval and the shape is adjusted such that the inbound leg should be right. So unlike CDI navigation, the curved paths and the three DR legs of a holding pattern are flown as directed by the GPS.
The downside of the accuracy with which the course is flown is that aircraft using the same course guidance are at greater risk of coming in contact with each other, so the big sky gets much smaller.
