Fairly obviously there must be a point where the GS signal becomes useless, simply because the GS transmitter is not in the middle of the runway, but to one side of it, and it radiates diagonally across.
Strangely enough a quick google doesn’t turn up any obvious diagrams.
JasonC wrote:
LPV slopes work below the minima of course – they are GPS derived.
They “work” yes, but the integrity is a bit scary. It promises to tell you within 6.2 seconds (TTA) if it’s lying to you by more than 50 m (165 ft) (VAL) about your vertical position. That’s not much of a promise when you’re below 200 ft. :)
bookworm wrote:
They “work” yes, but the integrity is a bit scary. It promises to tell you within 6.2 seconds (TTA) if it’s lying to you by more than 50 m (165 ft) (VAL) about your vertical position. That’s not much of a promise when you’re below 200 ft. :)
Indeed. I was by no means advocating their use below minima. I was merely describing the technical difference between that and a ground sited system.
Unless I have no choice but to land and have declared an emergency (ie very low fuel and fog) I would not rely on any of these below minima. You shouldnt have to. In that emergency scenario, I would prob rely even more on synthetic vision to stay straight and flare. And hope it is accurate.
Peter, the issue isn’t a lateral problem. It is the relationship between the lower and upper beams (150 & 90hz). This image isn’t the best but perhaps it will help explain.
There are two predominant elements at play, Signal Strength and Difference in Depth of Modulation (DDM).
Firstly, signal strength – The closer you get to the runway the nearer you get to the edge of both the 150 and 90hz signals. Despite the fact you are closer to the radiating source, the strength of the signal does not noticeably increase within about 6nm as you are progressively easing towards the edge of the signal. Furthermore, any deviation at the edge of the signal creates a significant drop-off. The baseline requirement is 190 µA and most signals get perilously close to this value. A lower value will (should) present a flag on the HSI.
More interestingly, DDM has a major part to play. This is the bit where the aircraft instrument ‘understands/decodes’ the guidance. In simple terms, DDM is a ratio between the 150 & 90hz signals defined as a percentage. If the ratio is 0% then you have a precisely balanced 150 & 90hz signal – you are on the glidepath. If the ratio is changes, you get a fly up/down (centrelines work on the same principle). As you get closer to the runway the physical distance between the top of the 150hz and the bottom of the 90hz reduces. Consequently, any deviation brings a bigger variance in DDM – it is this that creates the increase sensitivity as range reduces which is very closely aligned to the angular displacement issues we see on the HSI. Immediately prior to the antenna there are a number of issues that create interference between the two signals. Consequently, they tend to not be straight until some way up the nominal glidepth. I wish I had a graph here to explain but basically the GP can either take a dive or a fly-up. Engineers can adjust the phasing of the 150 + 90hz to adjust this bend and most choose to create a fly-up scenario rather than a fly down.
Hopefully you will note from the diagram that the closer you get to the runway the closer you get to the edge of the 90 & 150hz beams (risk of signal drop-out – red flag) and the tendency for the DDM to vary, wildly. CAT I systems are susceptible to DDM issues beyond system minima and locations that have terrain/obstacle issues also tend to have signal strength & alignment issues. As with anything that involves RF, if you tweak a signal for its size (structure) you pay a price with it’s strength, and vice-versa. Consequently, it is a fine balancing act when setting the system for best results. More modern systems deal with some of the issues in a number of ways by introducing things such as pseudo ground planes, additional antennas (working slightly out of phase) etc but the bottom line is that the GP propagation is a fundamental weakness on ILS. CAT II/III systems have been confirmed to have acceptable DDM/Signal Strength characteristics down to 100ft (some even achieve 50ft) providing additional criteria such as more remote holding points, LVP, topography, grass length, ground wetness/density etc are met. Each system is unique, many require additional measures if used below CAT I minima, and a pilot cannot assume that because ‘it worked’ at Little Snoring last week it will work at Little Snoring this week, never mind Big Wheeze just down the road.
Lastly, I’m struggling to think of any fixed wing operation whereby an approach below 200ft DH is authorised without an appropriate A/P (flying the approach) and two crew monitoring; there’s a good set of reasons for that and I would suggest it isn’t just red-tape.
OK; I see that; thanks.
FWIW I used to know a guy who had a twin TP Commander who got himself CAT2 certified; down to 150ft I believe. He was the famous Ole Henriksen, from here.
Bookworm’s post is scary – that sounds really bad. But how would the GS flag on an ILS compare with that?
Oh far better. One of the biggest stumbling blocks with CAT III GPS approaches is refresh rate. Our DGPS guidance has a 10hz rate and that still has trouble keeping up. We may know where we are to within 20mm at almost real time but we will not fly a calibration in less than 500ft cloud base and 2000m visibility.
I don’t have the technical knowledge to wade into this debate with any clout, but I have been employed by two separate companies flying two different aircraft variants that were certified CATIIIA. Both these companies had/have SOPs to say that during a CATIII approach the RHS pilot’s (that’s me) job is to monitor GS and LOC deviations right until touchdown and call out any deviations, no looking outside allowed. Now I appreciate how the aircraft flies this and how the autoland is completed, using averaged descent rates, tracks and rad alt rate of change – but in the years I have been doing this, watching those needles down to 0 feet (or even more realistically until 30’ or so whenever other stuff needs monitored) – I can’t say I have ever seen any issues with the raw data glide slope info in front of me – and this also includes practice autolands whenever there are no LVPs in place at the airfield. I can say this confidentially as it is in that instance literally my own professional job to announce it and if great enough to call a go around.
On another note, Dave and Timothy – You both individually and collectively share a wealth of knowledge, information and opinions that are of much interest to myself and I am sure many otherss. I would personally prefer to see the one-upmanship that is so common in other places left aside.
I don’t think that there is any one-upmanship. Dave and I are at different points on a spectrum when it comes to envelope pushing and we therefore disagree from time to time on the subject. But I don’t think there is any personal rancour, just the very statement of opinions that you say you wish to read.
We are unlikely ever to to agree on this type of issue. I take greater risks at the boundaries and accept both the upside and downsides of that. Dave thinks that I shouldn’t and that is absolutely his right to express.
I have a rather low opinion of some regulators and some of the regulations they produce, and have some sometimes rather energised meetings with them as I try to explain the negative consequences of their focus on Type 1 risk (I have given three examples in this thread, but there are plenty more.) Those meetings are ongoing and gradually progress is made. Indeed the UK CAA is very good, at a policy level, at understanding the nature of proportionate regulation and pilot responsibility, and they do seem to be successful in bringing EASA along with them.
Thus, focus on obscure and rare theoretical risk is gradually being teased out of the system in favour of pragmatic risk-based analysis, and thank goodness for that.
But, at the same time, there is always room for the technical guy to say “have you thought of this?” and I value that. We can then debate and decide. I hope that this thread has achieved that balance.
Cat III and CAT I are somewhat apples and pears – the ILS system on the ground for CAT III and CAT II is more sophisticated.
Peter wrote:
Bookworm’s post is scary – that sounds really bad. But how would the GS flag on an ILS compare with that?
Let me redress the balance a bit. The VPL in the southern UK is typically more like 15 m, and the confidence level is 2*10^-7. The typical 95% (2 sigma) position error is 4 m vertically, or 10-15 ft.
I don’t see a problem with using the needles below DH on either sort of approach, as long as it’s not total reliance. Once you add the visual picture to the integrity budget (and I think it would remain in peripheral vision), there’s an extra check that the needles are not providing hazardously misleading information.
