what_next wrote:
The problem with most (all) RNP capable navigation systems is that they will only switch to RNP 1 or better once established on the final track of an approach which is stored in their supplied database. Otherwise – and that includes DIY procedures – they can/will remain within RNP 5 tolerances. I can observe that myself often enough with our onboard equipment whenever an RNP approach is mishandled by the pilot (e.g. by setting a direct-to the final approach fix or by inserting waypoints between the stored points). In those cases, using the exactly same autopilot mode, the aircraft will deviate from the centerline no more than 20m in “APP” mode but several hundred meters in “TERM” or “ENRT” mode.(I guess by RNP 5 and RNP 1 you mean RNAV 5 and RNAV 1?)
If the equipment is RNAV (or RNP) 1 capable (as such a standard GPS device as GNS 430 is, even without SBAS), then it will be within RNAV/RNP 1 tolerances in TERM mode. I.e. it will switch to RNAV/RNP1 tolerances not when on final approach, but already when 30 NM from the destination. The switch on final approach to APP mode gives RNP APCH tolerances.
This for me is one of the problems many pilots do not understand about RNP approaches (official and selfmade) and a constant risk factor.
Indeed!
FWIW, I have been flying with my KLN94 set to 1nm full-scale since 2002. It is a preflight config setting. The default 5nm produces really sloppy tracking so why bother with it. Then, flying a GPS approach in the database, it narrows gradually to 0.3nm FS after the FAF. But one could fly a DIY IAP with the 1nm setting – subject to some reasonable minima like 600ft. The 0.3nm setting takes you to the runway centreline every time.
Why that helicopter crashed seems to be unclear. They were not using the GPS, apparently, but were tracking something else. The problem with the old GPSs is that they had no moving map; you just got a course deviation – like a CDI bar.
I think the overarching point is that there are a lot of (potential) gotchas associated with homemade approaches.
Sure, but the key is knowledge. One can screw up an ex-database GPS approach too if one doesn’t know which knobs do what. There are different levels of diligence. For example what you can do in your own plane which nobody else flies is very different to a syndicate or rental plane where people can reprogram user waypoints 
There are many DIY approaches around the place. I have a collection of DIY plates, from over the years. They are used by the locals and they get a lot of use, so I don’t believe there is any extra danger if in the right hands. They don’t get posted on forums. Some of them go back to some AOC operation which had a “private plate” but which is long gone.
I once designed one for Welshpool, with a missed approach, and cross-checks to some nearby VOR/DME, IIRC. I test-flew it in VMC but for unrelated reasons didn’t fly there again in IMC.
Ultimately, for your own plane, nobody else using it, proper design and a VMC test, the difference between a database IAP and a DIY IAP is the automatic FS tightening to 0.3nm past the FAF… but you can do it manually.
So I don’t think one can make blanket statements about danger. It all depends on the detail.
This for me is one of the problems many pilots do not understand about RNP approaches (official and selfmade) and a constant risk factor.
I couldn’t agree less. There’s no particular problem or risk. All we have to do is Read The Fff-fine Manual, set the CDI scale manually for a DIY cloud break, AND check that the auto-switch (if enabled) has operated correctly on a published approach. It’s about the same order of complexity as selecting whether to put kero or gasoline in the fuel tanks according to which engine the good old boys at Moultrie bolted on the front of your bushplane.
(I guess by RNP 5 and RNP 1 you mean RNAV 5 and RNAV 1?)
There is unfortunately some confusion with these designators. In some airspaces (USA, North Atlantic, …) the precision is designated by a suffix like RNP 1 or RNP 5. Others (Europe!) use a letter as prefix like P-RNAV or B-RNAV for the same thing.
BTW: the navigation system of the aircraft I fly will remain in ENRT (or B-RNAV or RNP5) mode with selfmade or modified stored approaches. The more precise modes TERM and APR will only engage with unmodified procedures from the database.
No manual setting for the full scale value? That’s interesting, but strange. Is that Collins Proline?
The GNS boxes had that setting too but AFAICS most users never discovered it, and fly with 5nm FS. On the KLN94 it is very obvious.
It may be that your equipment is unsuited to this use, but all the normal Navigators used in light aircraft (GNS, GTN, G1000, IFD) can be set manually to RNAV0.3, which is the same as APR on a GNS430.
As Jacko says, it is part of the approach check to set it to 0.3. I have this printed on my briefs on unpublished approaches.
At least one training school (not far from Gloucester) teaches to set the GNS/GTN permanently to 0.3 nm full scale, rather than the usual “auto” setting. That’s arguably OTT, but it certainly sharpens up one’s en-route flying.
At least one training school (not far from Gloucester) teaches to set the GNS/GTN permanently to 0.3 nm full scale
That is likely to produce instability when intercepting anything – unless you have GPSS (a.k.a. roll steering) whereby the GPS computes the intercept and feeds ARINC429 roll steering commands continuously to the autopilot. In that implementation, the roll angle is gradually reduced as the desired heading is approached, in a linear manner which while crude doesn’t generate oscillation. A “classic 1980s/90s” system, where the autopilot gets its heading from the HSI course pointer and gets a fine adjustment from the HSI deviation bar voltage, is going to overshoot massively and then oscillate.
That is why IFR GPSs tighten up from 5nm FS to 1nm FS 30nm from the airport (and that doesn’t cause instability) and then gradually tighten from 1nm to 0.3nm after the FAF. Doing it gradually avoids the control loop going unstable.
So you need an autopilot which can accept ARINC429 roll steering, or if it is an old one which doesn’t, you need the “fake heading bug” roll steering converter which converts the ARINC429 data stream from the GPS into a fake heading bug and then the autopilot is left in the HDG mode.
For examples of 0.3nm behaviour in the classic system, see here and search for “track fragment”.
