@lmsl1967 I think part of the reason is historical inertia, and part is training for all the privileges the licences provides. i.e. there are still a small community of AOC, or aerial work, operators using piston twins with vacuum gyros in IFR. In fact there are still corporate turbines that are mainly non-glass, or just might have an EHSI. You then have the small private operations IR pilots, although their fleet might have a higher proportion of updated avionics than a well run outfit using a couple of Islanders.
I doubt any of it is still relevant to those who fly on PPL/IR?
It’s better to leave these legacy partial pannel, hardcore emergencies, mental calculus…to wanna be CPL/MEIR holders, they are the ones who are likely to shoot NDB approach on turn coordinator with 300ft ceiling on one engine before going around (they can’t “legally” use ForeFlight iPad Synt-Vision to “dead-recon” their way out to VMC in an emergency, it has to be the old way or die hard trying 
)
In gliders that may get wet and frozen sometimes on the windshield, which is an accurate indication that your non-heated pitot tube and airspeed is about to give up
You need serious IFR version of “dual heated yaw string” !
Ibra wrote:
they are the ones who are likely to shoot NDB approach on turn coordinator with 300ft ceiling on one engine before going around
@lbra you may be amused not totally in jest, in the not so distant past some of the FAA check rides (Part 135 eg) might be somewhat Soviet, with testing to destruction of the candidate. The check ride, which might take some hours, after a two or three hour oral exam, and going to every airport in the vicinity to test every type of approach available from the avionics, the check ride would conclude with a non precision, limited panel, asymmetric approach. The FAA has a life preserving allergy to asymmetric instrument go arounds in a wheezy piston Navajo, so the limited panel asymmetric approach is to a landing.
I am sure in the 1960’s the Canadians and the Alaska DPEs would have required an asymmetric, bat-and-ball four range LF night time approach into some mountain airport, requiring you to level off (this time in a state of the art wheezy Apache) to count mountain obscuration lights before continuing. In the Apache no CDFA as single engine performance would be de facto dive, short drive, and dive. So the level off to tick off the mountain obscuration lights would need to be done in about ten seconds before the inevitable continued descent. To give the four range system it’s due, it was less prone to all the various equipment errors (cumulative in my experience, they never cancel themselves out) that the ADF is prone to.
In fact I hope one day, if GPS signals become chronically unreliable, some savvy civil servant might re introduce this very cheap, and reliable system of radio navigation.
Indeed, I heard that from someone who flew piston twin commuters on round dials in FAA land they also love to kill the engine on takeoff (for real in real minima) but again, I am not sure that kind of business is still around there or around here? although, to be fair with the required skill for CPL/MEIR, things has to be sharp !
For PPL/SEIR operating single engine GPS+ILS (and nothing else) with the typical 2km visibility & 600ft ceiling (unless one can afford to base it in LeBourget or Southend), flying for fun makes zero incentive to push a trip when avionics fail, I would expect the required level of precision and amount backup is rather “very low”, any backup would work 
I think the whole partial panel stuff needs a complete revision, but
All equipment malfunctions on a DPE check ride are simulated. The FAA would pull their ticket if they failed an engine low level for real. Part of the check-ride may include a scenario of engine fire which would require a shutdown, and emergency descent, with a subsequent re start. However simulated engine failure would occur in low IFR.
It is a reasonable constant that a proportion of relatively low currency pilots will attempt to feather the wrong engine, assuming their motor memory kept the aircraft under control, so no DPE would fail an engine for real (fuel cut off) low level. In a shutdown scenario the candidate carries out the shutdown.
My Bonanza has a mix of capabilities. I have a G500TXi and a G5 for backup. I have a GTN750xi and a KNS80 both connected to the G500TXi. The KNS80 also has a dedicated GI106A indicator. The autopilot is an Stec 60-2 with a ST901 GPSS and the G500TXi can provide GPSS as well. If the G500TXi fails, I can still use the autopilot and guide it via the GTN750xi and the ST901 GPSS. That provides capability fly routes and the lateral of a GPS approach. The KNS80 has VOR, DME, RNAV (Rho Theta) and ILS and is independent of the G500TXi when using the GI106A CDI. The G5 has a standby battery with up to 4 hours of standby power. The vertical Magnetic Compass can provide heading information. I have a standby B&C alternator that provides 20 amps of power that will drive all the avionics and instrument lights including the GTN750, the G500TXi, G5, GTX330, GDL88, JPI, Shadin Miniflo, and Stec 60-2 All the CB have been replaced with pull type CB, so I can adjust load as necessary. I also have both a Flight Stream 210 and a 510, although they are not intended as backup as I kept the 210 mainly for testing. I also carry two iPads and an iPhone that have the latest version of ForeFlight on them with both Jeppesen charts and FAA charts. I keep a user generated map overlay on Foreflight that identifies all the airports that are MON airports graphically. A MON airport is one that has at least one approach that is totally independent of GPS, DME, ADF, Radar and may be flown solely with a VOR or ILS capability. I keep a few flashlights and spare batteries handy as well.
The thing is that you understand all those systems. Most people would not.
I too have a backup alternator now but the “avionics bus” is still a single thing – a single point of failure.
