Part II: Air Canada, too low on...
 

guess! :uhoh:

link


Controller alerts Air Canada flight after it descends too low too soon in New York landing

Published on Friday December 14, 2012
Share on twitter Share on facebook
MARK BLINCH/REUTERS The airline has launched an internal probe into the Nov. 27 incident involving Air Canada Flight 748.
Bruce Campion-Smith
Ottawa Bureau chief


OTTAWA—An air traffic controller had to alert the pilots of an Air Canada flight after they descended too low during a bad weather approach to New York’s La Guardia airport.

The airline has launched an internal probe into the Nov. 27 incident involving Air Canada Flight 748, which happened as the twin-engine Embraer 170 jet was arriving from Montreal.

The pilots were using electronic aids to guide the aircraft through the low clouds, rain and late-day darkness to a landing on Runway 4 and had been told by air traffic control not to descend below 520 metres until passing an approach fix.

But with autopilot engaged for the approach, the Embraer jet started down to the runway too soon, busting the altitude restriction issued by the controller, according to a preliminary report prepared by Transport Canada.

The jet — still enveloped in cloud — continued down and was just 300 metres above the borough of Queens when the controller sounded the alarm about the premature descent.

“The aircraft was one mile outside the (fix) when it reached 1,000 ft. and (air traffic control) issued an Altitude Alert,” the report said.

The pilots quickly aborted their faulty approach, circled around and made a successful landing at the busy airport on their second try.

The Transportation Safety Board of Canada conducted a preliminary investigation into the incident but has handed the file over to the airline to pursue internally, a spokesperson said Thursday.

Airline spokeswoman Isabelle Arthur said the controller told the crew that there “appeared to be a discrepancy with the aircraft altitude indication.”

“As per operating procedures the crew did a go around and landed without incident,” Arthur told the Star in an email. “These types of events are extremely isolated and we always conduct internal reviews to ensure we maintain the highest levels of safety standards and operations.”

Arthur did not say whether the discrepancy had been attributed to human error or a mechanical problem.

Not an infrequent occurence at KSAN - no big deal?
 

I occasionally hear the tower at KSAN issue altitude alerts to aircraft east of KSAN where there are many 3500' hills within 25 miles and on the published approaches to the sea level KSAN RWY 27.

No one seems very excited about the alerts when I hear them on the radio. Seems like no big deal when I hear them.

I heard a rumour jungle jets don't have proper VNAV. :oh:

told by air traffic control not to descend below 520 metres until passing an approach fix.

"The aircraft was one mile outside the (fix) when it reached 1,000 ft. and (air traffic control) issued an Altitude Alert,” the report said.


600' low one mile prior to the FAF and descending. That's a problem.

"600' low one mile prior to the FAF and descending. That's a problem."

Yessir, it certainly is, and it will be interesting to see how it occurred. I wonder if they run a flight data analysis program on the type?...

Landing at Dallas one day in an MD80 noticed the DME vs full down glide slope didn't make sense with FO flying so told him to level off. Soon after level off we were advised of a low altitude alert by controller. We were in the clouds around 1500 ft AGL. DFW ceiling was reported at 1,000 ft. A B747 was stopped next to the GS antenna causing the problem but tower said they didn't have to protect the sterile approach area until the ceiling got below 800 ft. Isn't that nice? The GS suddenly popped to full up so beware.

bubbers,

In any approach there is a "path" component and a height vs. distance component. We all know the formula that checks altitude against distance, and many who fly (I'm retired) have a glass map that shows the runway.

Why do crews (in general...we don't know yet what happened here), sometimes slavishly follow the glideslope indication and bust the FAF altitude and even minimums? Seen it in the data due to a stuck GS indication...near CFIT.

Avoiding the hindsight bias trap, what makes sense to a crew such that they continue a descent below the altitude-vs-distance checks, and why does it make sense at the time?

Why are ANY references to meters made, when discussing an approach not made in China or Eastern Europe? Common references would be appropriate to Canada and NY, I woulda thought. Sam

Because it is a Canadian newspaper's article?

What's a "jungle jet" ?

Jets made in Brazil.

Three things...No 1: This is a non-issue....No 2: "Automation seems to always come into the conversation when discussing these issues, and 3: A previous poster hit it on the head....A Canadian Newspaper from an area in Canada where (god forbid) French is widely spoken...

Simple Answer...if able, fly the A/C, don't let the Autopilot Do It!!! And if not able, go to Roadmaster and drive a truck...

DownIn3Green;

"Three things...No 1: This is a non-issue...."
You don't say what you mean by "non-issue."


The event itself is serious. The intent of an internal (airline, not Transport Canada) investigation under SMS is to find out why the event occurred and make changes where indicated. If the airplane performed incorrectly the reasons will be discovered and communicated to the appropriate parties so the problem can be addressed. If the crew performed incorrectly, that will be discovered and a review of procedures, of the SOPs carried out on the subject flight, and perhaps some training will take place if mistakes were made and then they'll be returned to the line. That's what I would call a "non-issue", and that's how it's done under SMS.

"No 2: "Automation seems to always come into the conversation when discussing these issues," and
"Automation" has not been discussed in this thread.

"3: A previous poster hit it on the head....A Canadian Newspaper from an area in Canada where (god forbid) French is widely spoken..."
French is widely spoken in Quebec and much less so throughout Canada although there are many French-immersion schools across the country. You should learn a bit about your neighbour to your north :)

bubbers,
I think something else was going on. The entire GP reflection area is pretty much off limits, i.e. no roads, taxiways etc (though I've seen, and immensely disliked, access roads on the runway shoulder). If you walk in front of the GP, you're enough of a disturbance to trip the near-field monitor (located a short distance towards the threshold) and turn lights red in the tower (not a good way to make friends), so something doesn't quite add upp in their 747 story.

I wonder what they do with those new taxiways going around the runway end. If they cross in front of the GP, an aircraft there should cause a disturbance. There are probably restrictions in place.

Edit: Oh, Dallas... yeah, plenty of places to put a 747 in front of the GP there. There should be restrictions in place. :)

According to this AvHerald story, they were flying the RNAV Y approach, so a false glideslope capture is not what we're looking at here.

Incident: Air Canada E170 at New York on Nov 27th 2012, descended below minimum height

Semaphore Sam,

It's because Canada is a metric country (although aviation uses mostly Imperial units) and all newspaper reports, not only in Quebec, use metric units, even when the original units are Imperial.

This occasionally leads to nonsensical precision, for example, someone in MN ;) may say "about a mile away" and the Canadian report will say "about 1.6 km away".

Based on the Avherald info:
http://i521.photobucket.com/albums/w...LGA04RNAVY.jpg
No Altitude/Distance scale...

Yes it does, 3000 at PONAE etc.... :ugh:

Living up to your name Clown, I see...

Example Distance/Altitude scale:

http://i521.photobucket.com/albums/w...loggs/Palm.jpg

Set minimums and not in VNAV? A/c descended under the Ejet version of FLCH or V/S?

Like we've never seen that before....:ooh:

See TAWS ‘Saves’, incidents 2,3 and 6, also incident 5 (same airport?).

Does this a/c have a G/slope display on an RNAV approach?

Yes. If you mean an approach ath indicator not the ILS GP.

Not sure what an 'ath indicator' is, but the737 can give an 'ILS' type display on a PRNAV approach.

Any images of what you get on this type?

Not sure what an 'ath indicator' is, but the 737 can give an 'ILS' type display on a PRNAV approach.

Any images of what you get on this type?

I missed the 'p'. Like the 737.

It also shows a horizontal view of the vertical profile on the MFD. Lots of situational awareness available.

Tried to post a pic but only get the blue cross of death.

E 190 VGP
 

Do not have a picture to hand, but on the 190/195 (I suppose the 170/175 are the same) you get a VGP indication (Virtual Glide path) indication which the AP can capture and follow.
The setup and checks are a little different but actually there is little difference in flying an ILS or an VNAV approach. You even arm the AP with the same button. The computer figures out depending on the setup (FMS/Radios etc) what type of Approach you are flying.
The most "difficult" approach you can get in an E-Jet is a loc only because you have to control the descent rate old school style:E (VS or FPA)
So it is going to be interesting for us E-Jet drivers to know how and why this incident happened.

If I do find a picture I will upload it...

Possibly a rather commonly seen error?
 

ATC instructs the flight to descend and maintain 1700 feet until DNNIS and clears them for the approach. Those who've flown into NYC area airports know that things often happen quickly and the likelihood of distraction and task saturation is increased. This is when "finger trouble" with making the proper inputs to the MCP and FMS usually happens. It's no coincidence that several levels of error catching measures are in force within the total system to stop such an occurrence from resulting in a CFIT:

1) Approach phase SOPs are designed (among other things) to catch input errors by cross checking altitude against distance to next fix. If you know the crossing altitude for each fix (approach plate) then you can easily see whether you're on profile. Since this can be misinterpreted or unintentionally neglected then;

2) ATC approach radar systems are designed to warn the controller if a flight is either below or predicted by the computer to be below the approach profile. A low altitude alert is generated and passed to the flight by the controller. This is what reportedly occurred in this instance. If the alert had not been issued then;

3) GPWS/TAWS/EGPWS (in order of sophistication and discrimination) provide the "last chance" warning of impending ground contact.

No EGPWS (what this airplane is equipped with) warning was reported in this instance. Had the airplane gotten much lower that far from the runway a terrain warning would presumably have been generated.

As to exactly what specific input error(s) could match with the scenario, I'll leave that to an E-Jet driver to explain. In each jet type I'm experienced in, the possibilities are different according to the avionics and vertical mode control design. It should go without saying that the airline in question will investigate and do what they must to reduce the likelihood of future similar occurrences.

Apologies to the vast majority of pilots who are already aware of the preceding. :uhoh:

No EGPWS terrain warning, but did the aircraft have the obstacle database option enabled?
Obstacles should be considered a major hazard in this area.

safetypee;

TAWS Saves is an excellent document.

"...but did the aircraft have the obstacle database option enabled?"

It's installed and I don't believe it can be disabled (turned off).

westhawk;

" Had the airplane gotten much lower that far from the runway a terrain warning would presumably have been generated."

The Terrain Clearance Floor (TCF) is, I believe, "standard" - 700', then from 15nm to 12nm the warning area descends to 400' and then at 4nm it descends to the airport level at between 1/2 and 2nm. The warnings would be for both terrain and obstacles.

ft, I have noticed a lot of airports have a protected area that doesn't allow you to procede past that point if weather is low. As I recall it is before the glide slope antenna about 1,000 ft from the end.

PJ2:

Thanks for the info, it inspired me to go look at an old Honeywell EGPWS pilot's guide applicable to the Hawker 800XP I used to fly. Indeed it appears that between 4 and 12 miles to the runway center a terrain warning would be issued only if the airplane had reached 400' AGL or if any terrain/obstacle penetrates the caution ribbon. (predictive terrain clearance)

So in this case it appears that the ATC radar low altitude alerting feature provides a considerably earlier warning than the EGPWS would have. I've seen a few nuisance low altitude alerts in my time but this one appears to have kept a premature descent from becoming a much closer thing.

For the types I'm experienced in I have pretty good idea how this scenario can occur because I've either seen or committed some of the errors possible. Fortunately effective error trapping made these occurrences non events, but eventually even the most effective SOPs break down for various reasons. I look forward to reading what a pilot with experience in this particular type might have to say about what equipment specific input errors might be likely to lead to such an occurrence.

- it might well help others if you shared this?

Easy:

Under ATC vertical control: level off here, descend there... so not in VNAV for the "approach" (PONAE) to the approach.

When established on final, set the MDA/DA (SOP-forgetting ATC said 1700ft...they did then say "cleared for the approach".) but still in VS: down you go all the way to the MDA, possibly before the FAF=bust 1700ft at DNNIS.

PJ2 (#30), thanks.
“If the obstacle database is enabled”
I assume that this depends on the equipment type / manufacturer / version.
For Honeywell EGPWS, the obstacle database appears to be loaded / able to be uploaded in all systems, but “Obstacle alerting is activated by a configuration item in the configuration module”; “This feature is an option, enabled by system configuration selection during installation.” (Pilot’s guide).
“… the obstacle database being accessed by the EGPWC application only if enabled via configuration module option.” (Installation manual).

I have yet to check what equipment this particular aircraft uses and the configuration / options in that system.
However, from discussions with several operators (some time ago), it was apparent that they (operators) and crews were unaware of the status of the obstacle mode configuration in their aircraft (various types) and the validity of the obstacle database.

bubbers44:

wx 800 and 2 in the U.S.

Thanks safetypee, if you're able to check on this aircraft specifically, I'll await results.

For the thread, there have been some instances at TEB recently where a relatively new taxiway crosses the landing runway in front of the GS antenna array. Aircraft inbound, coupled to the GS, have reported a several hundred foot altitude gain as the autopilot rather quickly attempts to fly up to a deflected GS signal. Even when the wx is above 800/2, stuff can happen.

Good job Capn Bloggs!

Here's another from the old Westwind 1: (with old fashioned Collins F/D mode controller) AP is engaged and GPS NAV source selected with course captured. this is planned as a LNAV only approach with manual step downs. ATC tells you to "descend and maintain 1700 to DNNIS cleared approach". PNF dials up 1700 on the preselect and presses ALT SEL on the mode panel. You pitch wheel the nose down to establish a nice rate of descent then select VS to hold it at that ROD. Neither of you notice that the "ALT SEL" caption is not illuminated. (SimuFlite mantra: latching indicator on the button is a rumor and the lit mode caption is confirmation!) Since ALT SEL is not armed you'd better be keeping your scan up if you're gonna catch this error before it turns into an altitude bust. You either bumped it out of ALT SEL when pushing VS or it was never engaged when it was selected. Just one of those gotchas you learn to look for in this airplane, fancy GPS/FMS or not.

From the Lear 60 with Proline 4 and UNS1-B or C dual FMSs:

Same approach and instructions as before. 1700 is preselected and ALTS mode auto engaged. Select VS mode and spin the VS knob to desired descent rate. This will be a LNAV/VNAV approach and doing the proper things in the proper order at the proper time is essential in order for the vertical modes to behave as desired. Upon capturing 1700 the ALT (hold) mode will auto engage At this time the approach mode may be armed and new altitude (MDA or MAA as appropriate) may be set in the preselector. If the ALT preselect is changed prior to ALT capture at 1700 the AP will be bumped out of ALTS mode and will continue descending in PITCH mode until a vertical mode is selected. Another gotcha that you hopefully learn in a way that doesn't have any real consequences.

But as I indicated before, I've made these errors myself and I'm rarely surprised when I see these or other similar oopsies in the cockpit whether flying with very experienced PICs or brandy-new nugget co-pilots. To date neither has led to anything more than a need to make a timely correction. I have known other pilots who didn't catch it in time to avoid consequences and I wouldn't want to trade places with them. In the best of circumstances one's superior airmanship will make it unnecessary to display those superior skills or something like that...

In my experience these "rookie" mistakes usually happen when under time pressure of either the externally or internally generated variety. Tired or mentally drained pilots make plenty of errors, especially when they allow themselves to become task saturated. Being "go mission" oriented folks in general, pilots are particularly prone to continuing an approach even when behind the task curve. NTSB reports repeat the same theme over and again. Prior Planning Prevents P!ss Poor Performance (the 6 'P"s) is a good way to remind oneself of the importance of effective cockpit work flow management.

The magnitude of consequences resulting from whatever errors that ARE committed is determined in large part by how soon they are detected and rectified. A habitual pattern of crosschecking and rechecking coupled with a cynical refusal to believe everything will happen as intended goes a long towards having the right attitude. Admit that being 99.9% infallible isn't nearly good enough and be prepared to detect and correct those mistakes before anyone else notices. :eek: That way even your flying partners might begin to think you know what you're doing some of the time!:) And hey the flights seem to go pretty smoothly too...

Man I gotta get back in the air. This groundpounder teaching stuff I've been doing since I lost my flying job sucks.

Good Afternoon All:

I have copied this from the Canada Transport Safety Board Daily Notifications.

It would "appear" to a data problem in the Data Base.

Stay Tuned..

----------------------------------------Occurrence 2----------------------------------------
Occurrence No. : A12F0157 Occurrence Type: INCIDENT VOLUNTARY
Class : CLASS 5 Reportable Type:
Date : 09-12-2012 Time : 19:45 UTC
Region of Responsibility : ONTARIO
Location : LA GUARDIA, UNITED STATES

Aircraft Information:
Registration : C-FEKJ Operator : AIR CANADA
Manufacturer : EMBRAER Operator Type: COMMERCIAL
Model : 170-200 CARs Info: 705 - AIRLINER
Injuries: Fatal : 0 Serious : 0 Minor : 0 None : 72 Unknown : 0
Occurrence Summary :
A12F0157: The Air Canada Embraer 175 (AC730, C-FEKJ) was cleared for the VOR approach to LGA runway 4. Prior to crossing NORAY at 1,700 ft the aircraft was observed on radar descending below the minimum altitude to 1,400 ft. ATC issued a low altitude alert and the aircraft leveled at 1,500 ft. The crew reported that LNAV and VNAV were engaged and the vertical profile on the VGP (Vertical Glide Path) appeared normal. The operator suspects there is a data problem in the navigation database for the non-precision approaches for LGA rwy 04. The event remains under investigation by Air Canada Fleet Management and Corporate Safety in co-ordination with Embraer and Honeywell.