If there is terrain around but not close to the runway (i.e. you are still in IMC when passing it) then a straight-in approach, with lateral guidance (CDI/HSI or autopilot) is vastly safer than flying an ILS to the opposite runway, going missed, and screwing up the missed approach and hitting that terrain.
If there is terrain near the runway, then an ILS to the opposite runway and a circle after that, is probably the safer option, which is why most King Airs and such like going into EGKA 20 would fly the 02 and circle to land. But this is a less common scenario, and is much less relevant if you fly the LOC-only approach on autopilot, with VS hold, and flying an accurate GS. Then you can hit every SDF within 20-30ft.
The objective in all this is to reach the runway and land on it.
I think BC approaches in Europe were simply a victim of European anti-American (“we do it better than the cowboys over there”) elitism. Europe can hardly make the case that there are adequate other options. Too few airports to start with, even fewer when you tick off the ripoff jobs, too few IAPs, silly opening times, some GPS IAPs nowadays but if there is a GPS issue then you have nothing.
BTW, is a BC implemented with the normal LOC antenna (which AFAIK always radiates a lobe on the opposite bearing) or is it another LOC antenna?
1 SM visibility
One can only just see the start of the runway then, with a DH of 800ft (EGKA 20) 
In US, we need to keep an identifiable portion of the airport (not the runway) in sight.
Same here; the circling is a fully visual procedure (not necessarily IAW VFR minima, of course). Some have done it in IMC… N403HP was one.
gallois wrote:
have a perfectly well designed final approach down to visual conditions?
Ah, I misunderstood your comment as VFR circuit with overhead & downwind under low clouds is better & safer than going straight-in IFR all the way
A final approach all the way down to the ground is even better than IFR approach & circling, I will take this every time: you don’t even need to do anything bellow DH, the aircraft is in love with the trim…
The choice is normally or used to be, before RNP /RNAV GNSS approaches, (Ignoring NDB and VOR approaches for this debate.) ILS to one runway even with 2 or even 3 runways on an airfield. Now of course the ILS precision approach down to a DH less than 250ft followed by a straight in landing if you can see the runway, is the best way to go. But what if that is not the runway in use? Well, you could go missed and go off to your chosen alternate. But if the visibility is above circling height and you can maintain that visibility you can go round to the runway in use and land on it.
For a category A aircraft a circle centred on the ARP with a radius of 2.4nm will be a protected area. If you are circling to a cross runway there will be another circle centred on the ARP of that runway, also with a radius of 2.4nm also a protected area. All obstacles within those areas must be marked on the plates. As long as you are aware of them and keep within the circles, providing you can fly straight and level in VMC, you should be perfectly safe.
My personal preference is to fly close enough to the runway to follow the centre line and as is usual in a twin fly a curve U base to final, ie not a rectangle. But I don’t think there is any one right way, it depends on what you get used to. Keeping the tip of the wing approx on the centre line at lower altitudes eg 500ft just means you are nearer the runway than if you were at 1000ft with the same visual. Eg 1km from runway instead of 2km depending on the aircraft you fly.
But this is no different from VFR circuit bashing or gliding for that matter.
gallois wrote:
For a category A aircraft a circle centered on the ARP with a radius of 2.4nm will be a protected area.
The area is defined first by circles with as radius (for cat A) of 1.68 NM centered on the runway thresholds and then extended by connecting the circles with tangents. The ARP is not involved. In other words, when circling with a cat A aircraft you can at all times be at least 1.68 NM away from any runway.
See this picture from PANS-OPS volume II. Of course for a single runway airport the boundaries of the circling area will look like a racetrack.
For aircraft with higher approach categories, only runways useable by those aircraft are considered. I would also guess that departure-only runways are not considered. (Unusual, but happens.)
Circling with prescribed tracks works differently, of course.
A_A Please look at the DGAC guide for Instrument procedure design and construction that I posted in another thread for Ibra.
It was the bible we were expected to use when constructing the GNSS approach here at LFFK.
Note the 2.4 nm I referred.
You reference to thresholds is more akin to the UK CAA guidelines for constructing an ATZ.
The runway protected area for “instrument circle to land” is an ellipsoide or racetrack that joins instrument RWY TCH to any non-instrument RYW TCH (as long as it’s not excluded), it’s the same for DGAC as in PANS OPS, I think US TERPS has different (tighter) definition, @NCYankee knows more on this?
The aerodrome obstacle chart survey area (mandatory for public transport in light or low performance B twins likely as they have to “visual circling after takeoff failure”) is defined as circle on AD ARP
UK ATZ is defined as circle around the center of the longest runway (but there MATZ, CMATZ)
Ibra wrote:
I think US TERPS has different (tighter) definition,
Yes. They used to be even tighter – the TERPS radii were increased some 10 years ago, but they’re still smaller than the ICAO radii.
boscomantico wrote:
In fact, Germany never had any.
Not recently but never? I recall that when I started learning about ILS’s in the 1970ties there was a mention of back course approaches in a German publication (Motorbuch Verlag) with an example from one of the larger airports there (I believe Hamburg but I may be mistaken).
When I did my initial IR in the late 1980ties I recall that back course approaches did exist in Europe, I saw some when doing the page revisions of my then Jeppesen 5 volume manual. Can’t remember where, but they were there. I also recall them being a subject of instruction at the time.
Peter wrote:
I wonder why Europe has not gone for them. They would enhance safety at the many airports which have an ILS on just one runway end and nothing in the other.
Because the reason for having ILS on just one side of the runway and not the other are typically not the cost for installing an additional set of ILS/LOC equipment but rather obstacle clearance, missed approach procedures, etc.
BC approaches are not a way of increasing options but a way of reducing cost.
With GPS-based procedures, BC-approaches have largely become obsolete in any case – but you can see evidence of the point above: How many airports in Europe you know which ILS/LOC-procedures only in one direction but now have GPS-procedures in both?
Malibuflyer wrote:
Because the reason for having ILS on just one side of the runway and not the other are typically not the cost for installing an additional set of ILS/LOC equipment but rather obstacle clearance, missed approach procedures, etc.
Many small airports have always had ILS just on one side of the runway, but now they also have LPV to the other side — and you’re saying the reason they didn’t have double ILSes in the first place was not cost?
