Air Canada A320 accident at Halifax

[peekay4]

@flyingchanges

There was a change to 7110.65 last year (version "V") following NTSB safety recommendations after the CO1404 accident. Gusts are now to be included when reporting surface winds (when applicable).

See paragraph 2-4-17(g). The new wording in bold:

g. Surface wind. The word “wind” followed by the
separate digits of the indicated wind direction to the
nearest 10−degree multiple, the word “at” and the
separate digits of the indicated velocity in knots, to
include any gusts.

[jaytee54]

It is always interesting that YHZ Airport Authority can find money for a shinny new terminal but no money for proper airport navaids, and they are not alone in Canada.

I suspect that runway 32 was not available as it had not been cleared of snow, the airport operations mgr. would make that decision.

A shiny new terminal is a fine place for high rent retail opportunities. Navaids on the other hand are an unnecessary expense, whose maintenance is a constant drain on resources.

I never flew buses, but the bottom line is that you cross the NDB at 2000 feet (QNH used I assume), then descend to 740 feet at which time somebody calls DECIDE and somebody else calls "Go-around", unless he/she has an adequate view of the runway and its lighting system to allow a safe landing.
So how come the aircraft touched the ground 40 feet BELOW the TDZ and 1100 feet before it? I await the report on that one. All else is of little relevance.

[peekay4]

A shiny new terminal is a fine place for high rent retail opportunities. Navaids on the other hand are an unnecessary expense, whose maintenance is a constant drain on resources.

And as I understand it:

New terminal = investment by the airport authority
New navaid = cost to NavCanada

Considering the transition to RNAV GNSS approaches, there's going to be very little appetite to invest in net-new ILS equipment, even though a large fraction of commercial aircraft in Canada cannot perform precision GPS approaches.

[fireflybob]

So how come the aircraft touched the ground 40 feet BELOW the TDZ and 1100 feet before it? I await the report on that one. All else is of little relevance.

Windshear, microburst?

[de facto]

The MDA was (is?) considered a hard floor. Descend though it on a Sim ride during the Go around (with no runway in sight) and it could be a Failure, or at least an RS..

With the old "Dunk and Drive"...MDA was a hard floor...times have changed and so have procedures...

[jaytee54]

So how come the aircraft touched the ground 40 feet BELOW the TDZ and 1100 feet before it? I await the report on that one. All else is of little relevance.

fireflybob:- Windshear, microburst?

Yes, or a technical problem, or whatever. At least they have the crew, aircraft and recorders to tell us what.

[TheInquisitor]

With the old "Dunk and Drive"...MDA was a hard floor...times have changed and so have procedures...

The methodology of getting to MDA may have 'evolved', but it is STILL a hard floor - unlike a DA, you do NOT go below it unless you have the required visual references. It is not the altitude at which you initiate the MAP.

[vilas]

This accident could be similar to the one that happened to a 744 at Frankfurt where the pilot was flying the approach manually and drifting fog started covering the runway from the far end causing an illusion of overshoot and he inadvertently kept adjusting by pushing the nose down and in the rapidly reducing visibility got fixated and failed to notice GPWS warnings resulting in an undershoot and hit the raised approach lights 700 mtrs. short but was successful in executing a go around.

[peekay4]

The methodology of getting to MDA may have 'evolved', but it is STILL a hard floor - unlike a DA, you do NOT go below it unless you have the required visual references. It is not the altitude at which you initiate the MAP.

No, it depends -- many operators in US / Canada have approved op specs to treat MDA as DA when performing SCDA approaches. Approval is granted to dip below the MDA floor during a missed approach.

[West Coast]

Perhaps in Canada. My company and others I surveyed at other carriers add 50 ft to MDA to ensure one doesn't dip below MDA on a missed. Can't say with any certainty there isn't a US carrier out there that allows you to dip below, but they're the exception.

[peekay4]

I know several US Part 121 and 135 ops authorized for ops spec C073, using MDA for DA. I would say it is not an uncommon authorization to have in the US especially for regionals & charters serving smaller airports having non-precision approaches.

By now the Transportation Safety Board must have a pretty good idea of why AC624 crashed...

[Groucho]

"By now the Transportation Safety Board must have a pretty good idea of why AC624 crashed"

Exactly. All this talk about MDA/DA/'hard floors'/'add 50ft'/'Dive and drive' is really irrelevant here.

The a/c hit the ground BELOW the threshold elevation and short of it. Regardless of 'limits', there was either a lack of visual reference and/or some sort of performance issue, either of which will already be known.

[7478ti]

Many accidents with this profile, over 5+ decades

There have been many accidents following this erroneous "visual reference" profile over the past 5 decades. In fact, it is important to note that even FAR25.1309 [see the preamble] was largely instituted due to the two famous fatal accidents very similar to this one, that took place at KBFD, both CV580s, at the same airport, both winter nights, two weeks apart. The first was AL736 on Christmas eve, Dec 24, '68 and the second AL737 Jan 6, '69. These two accidents were key events in the direct chain that led to RNP's creation, assuring that every SIAP has a safe vertical and lateral path defined all the way to the TDZ, all the time,... as well as a safe path back out from the TDZ, in the event of some factor like "loss of visual reference".

As to the MDA being used as a DA, Theo VanDeV at KLM clearly showed during AWO HWG deliberations over two decades, ago why this policy to set MDA=DA is safe and appropriate, without needing any 50' additives, but some states and authorities have nonetheless been slow to take advantage of the increased safety of using VNAV paths with MDA=DA, in lieu of "dive and drive", by still assessing a completely unnecessary 50' additive penalty.

[Retired DC9 driver]

The Inquisitor wrote

The methodology of getting to MDA may have 'evolved', but it is STILL a hard floor - unlike a DA, you do NOT go below it unless you have the required visual references. It is not the altitude at which you initiate the MAP.

Actually, in the Airbus if you have the "magic" working properly, and have set the FPA properly at the FAF, and crossed the FAF at the proper altitude, when you reach MDA without the visual reference there should be no hesitation, it's a Go Around. You should have been in a perfect position to "see" the runway, stabilized on the correct approach path to touch down with correct Threshold Crossing Height.
So in that sense MDA is the altitude at which you initiate a MAP, in a constant descent, if you can't see the runway, ie a Decision Altitude.
Whereas on the DC-9 you got down to MDA, inside the FAF stabilized (but early!) so you could level off, and sit there, peer out, at MDA looking for the runway..

[L39 Guy]

GLS/RNP vs LPV

Aterpster,

I agree with you 100%. As an individual that is involved with procedure design and has design many and flown hundreds of LPV approaches, LPV is a vastly superior system than GLS/RNP. I state that for the following reasons:

Procedure Design
- the design criteria affords a very tight final, tighter than RNP (even 0.1 in close). There are a handful of places in the world where RNP 0.1 would yield lower approach limits than LPV
- the vertical path of RNP, being barometric altimetry driven, is subject to all of the barometric altimetry problems in both warm weather (a flight path angle that is steeper than the nominal path) and cold weather (flatter than the nominal path). As well, there are temperature restrictions at both ends of the temperature spectrum
- GLS may solve the barometric altimetry issue however it creates many others (discussed shortly)
- LPV uses satellite-derived height, thus eliminating all of the limitations of barometric altimetry
- how would it work at airports and not just the ones where airlines go exclusively? For each runway end there would be an GLS/RNP approach for the airliners and an RNAV (GPS and LPV) for the bizjets and GA aircraft?


Financial
- North America is covered with WAAS coverage; my understanding of the European Galileo satnav system is that it will have the qualities (accuracy, integrity, etc) to support Cat 1 ILS like approaches, ie LPV/APV
- with WAAS the navigation source is free; with Galileo it will come with a small licensing fee
- Galileo will have worldwide coverage
- GLS means, last time I heard, a $3 million per airport investment in the ground infrastructure; the investment in the aircraft to either install new multimode receivers or modify the existing ones is massive (last I heard about $250K per box)
- naturally, Honeywell, Boeing, Airbus and the others would love to have the airports and airlines pony up that kind of dough
- it is interesting to note that a brand new, certified WAAS receiver from Garmin is about $14K; upgrading an existing box to WAAS is $3K. Why Honeywell and the others can't do it for something similar (or even double) is beyond me.
- $3 million per airport times how many airports? Big, big dough.
- Given that there is no operational benefit there is no financial benefit from the cost to the industry of GLS boxes on the ground and expensive MMR's in the aircraft

With the knowledge the industry has with LPV/APV and with Galileo on the threshold of operational capability, the only logical and financially responsible solution is LPV/APV - airlines should be pushing the avionics providers to get with the program and upgrade the existing GPS's and FMS's to a WAAS/Galileo capability that would provide the best operational capability with the lowest (by a large margin) implementation cost to the industry.

[scud]

And as I understand it:

New terminal = investment by the airport authority
New navaid = cost to NavCanada

Considering the transition to RNAV GNSS approaches, there's going to be very little appetite to invest in net-new ILS equipment, even though a large fraction of commercial aircraft in Canada cannot perform precision GPS approaches.

Around 1998, Nav Canada conducted a cost-benefit analysis concerning the installation of an ILS for runway 05 in Halifax. The request came from an operator who's 767 suffered a major tail strike a couple of years before, that cost 10 million dollars to fix. That was due to crossing the FAF at the minimum crossing altitude, giving a flight path angle of about 1.2 degrees to MDA, and coming in a little slow near the VRef.

The result was that Nav Canada's cost for an ILS would have been about 2 million dollars, plus another 2.5 million to the Halifax Airport Authority for upgraded approach lighting and landscaping, because the localizer antenna would have to otherwise be placed in a ravine, and there was a service road where the glide-slope antenna would be situated. Both agreed that the cost-benefit was positive: 4.5 million dollar investment compared to another potential (or worse) 10 million dollar loss to an air carrier. The airline in question balked at being responsible for such an outlay, despite the fact that they understood that the system would pay and not the airline directly. Nav Canada put the project on the back burner due to lack of airline interest.

Typical of airlines in Canada, at least at the time: our pilots are heroes, except whey they become when they crash an airplane.

[scud]

As to the MDA being used as a DA, Theo VanDeV at KLM clearly showed during AWO HWG deliberations over two decades, ago why this policy to set MDA=DA is safe and appropriate, without needing any 50' additives, but some states and authorities have nonetheless been slow to take advantage of the increased safety of using VNAV paths with MDA=DA, in lieu of "dive and drive", by still assessing a completely unnecessary 50' additive penalty.

"dive and drive" has nothing to do with adding 50 feet. It is irrelevant in this context because "dive and drive" is seldom used anymore. Use of a DA or MDA in this case both involve approaches along a predefined flight path angle using the aircraft vertical nav capability.

The issue of whether to add 50 feet is related to the accuracy of the approach, navigation equipment and the regulatory limitations of the operator.

A DH assumes that a decision to land or go-around will be made no later than at that height, and it is permissible for the aircraft to dip down below that height in case of a go-around because of its inertia. It is predicated on the aircraft being exactly at the MAP location, as guaranteed by the precision approach system (ILS) for example, so that adequate clearance above obstacles is maintained.

An MDA assumes a lower degree of precision, in which case the MDA is the lowest altitude the aircraft can descend, taking into consideration the highest obstacle within a certain radius of the MAP (plus the path from the FAF); the exact location of that obstacle in not considered in the approach. The typical height above this highest obstacle is 250 feet. Because of this uncertainty, the aircraft must never be allowed to descend below MDA unless visual, because of the potential but uncertain proximity to this highest obstacle. This is why 50 feet is added: to allow for the dip in altitude due to inertia and reaction time when deciding to do a go-around, assuring remaining above MDA.

RNAV approaches can either use DA or MDA, depending on the operator certification and the accuracy of the nav equipment. If you look at a GNSS approach plate, you may see 2 sets of minimums: one for LNAV/VNAV and the other LNAV only. Under the former you will find a DA and under the latter a MDA. 50 feet must be added to the MDA, unless doing a 'dive and drive' descent, which is still legal but not SOP. Some major operators in Canada are not approved for LNAV/VNAV, despite having such limits on the approach plate.

[peekay4]

Incidentally the blanket exemption in Canada allowing the use of MDA as DA (without adding 50ft) will expire at the end of this month -- April 30, 2015.

I believe large jet operators may still request exemption, but it will be interesting to see if the blanket exemption gets extended or not.

[YRP]

Originally Posted by L39 Guy

- it is interesting to note that a brand new, certified WAAS receiver from Garmin is about $14K; upgrading an existing box to WAAS is $3K. Why Honeywell and the others can't do it for something similar (or even double) is beyond me.

Maybe off topic, but the cost is largely NRE (development cost). From working in the electronics industry on even smaller product developments, without the same reliability / robustness targets as aviation, the development costs can be pretty big.

For a general aviation Garmin box, the manufacturing cost is sub $1K. The rest is engineering work.

The market is much smaller for commercial airline avionics. They are going to sell/upgrade between 1000 and 10,000 units (at a guess) for any given product. Those have pretty high standards for reliability/certification and a lot of integration with other a/c systems.

If you add up the engineering work (design, verification, certification) and divide it across the possible market, and add probably not a small amount of liability insurance, those cost figures are probably not unreasonable.

[AJW709]

On the investment case (or not) for ILS upgrades at YHZ it's interesting to compare to YYT (St John's, Newfoundland), another place with a surfeit of interesting weather conditions which will install Cat III ILS next year (becoming only the 4th or 5th airport in the country with Cat III ILS & the first to use LED approach lighting. (Source:http://stjohnsairport.com/wp-content...ative-FAQs.pdf)

It strikes this SLF as odd that Halifax, as an arguably busier airport than St John's, hasn't put in the same investment in addition to the shiny new terminal/car parks/hotel.