How do you fly your non-precision approaches?

[Right Way Up]

Seems quite simple! Those that fly a level segment set bugs to MDA, those that fly a constant descent set a minimum DECISION height of MDA +50-100 so as not to infringe MDA. The reason some companies in the UK went to constant descent was that during the level segment to the MAP some pilots were getting visual contact with the runway(way too late) and performing a dirty dive to get in. IHMO I don't particularly like flying level at 300 feet fully configured.

[411A]

Right Way Up,

Dive & Drive works quite nicely if done the proper way, ie; gear down and approach flap selected, landing flap only when the runway is in sight, and landing is assured, without excessive sink rates.

[4dogs]

for the benefit of miserlou,

Yes, I understood exactly what the discussion was about.

I apologise for not predicting that you would fail to read my words with the same precision that I wrote them.

Because the circumstances of the approach change, I merely stated the last place from which to go around, not the first. We fly a mixture of runway aligned and circling approaches and there are times when a runway approach may need to be converted to a circling approach. As I mentioned, the process is sufficiently robust to cater for each option - my only requirement is for the plan to be briefed prior to commencing the approach.

My reference to planning a lower profile never reverts to "dive and drive" in that we are talking of 50-100' only - while mildly destabilising, it greatly improves the chances of success.

Stay Alive,

[FlapsOne]

If you Dive n Drive, with approach flap and a minimum of, say 400', then you are not stable by 500' (no land flap) so you must go around.

You could be up there all day!

Flying a level segment in a med/lge transport aircrafdt is barking mad. Even more so on one engine!

Notional glidepath and add 50' or so to the MDA and use it as a DA.

If you can't get in from that, go somewhere else!

[PPRuNe Towers]

I've enjoyed watching this discussion develop and held off from commenting for now.

For those relatively new to the industry its a topic that's rattled around the electronic bulletin boards for 15 years and has traditionally been far more divisive and angry than you're witnessing now.

Discussion tended to split geographically with our American colleagues resisting the basis of Canpa far longer than elswhere. Major european airlines such as KLM formalised and introduced the concepts many years ago and it has spread at varying speed around the world.

At differing times companies have spotted and acted, usually independently of the legislators, to cover the MDA point discussed in this thread. In a handful of cases individual airlines have gained permission to use the MDA effectively as a decision height - some of our Canadian readers might have experience of this.

Here in the UK we have an a very unusual situation with CANPA. As you'd expect, progress along the way with the early adopters such as BA and the mature charter companies percolating through the rest of the industry. They've mainly covered the introduction of the 'pad' above MDA at their own pace and in their own time - the requirement sometimes spotted by pilots far lower in the food chain than those in charge.

11 months ago, when approached by the main UK pilots' union with a written question regarding formalising the requirement for UK public transport to observe the precise definition of MDA we were aghast at the response.

The UK CAA stated there would be no requirement to do so in order to keep UK companies competitive in Europe - presumably with Pittsle's operation

Ah, thank god for a safety minded regulator I really do hope they're well on the way to a change of heart since last December.

Regards
Rob

[LEM]

DIVE & DRIVE IS SAFER!

Yes Dive&Drive is safer.

And also much easier.

Those who find the level segment destabilizing simply have lost all their handling skills and try to hide behind the constant angle magical solution (which doesn't work, and puts them in an even worst and dangerous situation).

They try to turn a NON precision approach into a PRECISION one, simply because they don't know how to fly it and are a bit scared of it.

These are the pilots who crash in the non precision approach, the dangerous **** of Guam and many other places!







Ok, let's cool down and try to explain.

Basically, there are two kinds of NPA: the easy ones and the difficult ones.
That means: the ones with various altitude vs distance check points reported on the plate, and those without anything.
These can be flown in good weather or bad weather.
When I say bad weather I mean ceiling just above the MDA. Poor visibility, of course.
Avionics can be advanced or old.
AP ON or raw data manual flying.

In an argument like this we must assume the worst scenario and examine it.
No doubt once we master the most difficult scenario we can master the easy one.

So the scenario is: NDB over the field, and nothing else!
Outbound leg of, say, 3 minutes, turn back inbound and descend to the MDA.
MAP over the beacon.

Those who plan for a constant descent will calculate a certain VS versus a certain speed, trying to correct for the wind.
They think they'll be able to reach VDP (Visual Descent Point, the point at which you can leave MDA on a 3° slope to the runway) using the calculated parameters.

Very funny!
Actually, I'll tell you what will happen:
First of all, when you overfly the NDB outbound your timing will be inaccurate, because of the dead cone and the instrument tolerances, and I've noticed this error can reach one mile (!);
Second, all your meticolous calculations will be falsified by the real wind, which is different from what you had expected at different altitudes throughout the approach; also, because of the difficulty involved in maintaining exactly the speed and VS you had planned.

The result will be: you'll certainly NOT reach MDA at the VDP, and what is worst you won't know if you are high or low on your imaginary profile.

In both cases, a missed approach is very likely, since you will end up too far out, or too close and high.
All these unnecessary goarounds reduce safety (more time spent in the air thus more chances of failures, more stress and pressure with low fuel and so on...).


But now, Pittsle, you are with me in the cockpit:
we'll fly this approach using the Dive&Drive method.
With this technique, it's not even necessary to make any meticolous calculations, as we'll know for sure on which side of the imaginary glide we are, so we'll know exactly what to expect and what to do.

Leaving each altitude step you will initially set about 1500ft/min, then approaching the selected altitude you'll reduce to about 1000ft/min.
I say about: it does't matter if your VS is not exactly a certain value.
This will certainly NOT trigger any GPWS warning.
You'll realize soon how easier it is to concentrate on one thing at a time: descending to your next altitude and levelling off, till the next one, instead of worrying all the time if your VS is still correct, if you haven't screwed it up, and if you'll reach the next step high or low...

Just one figure in your mind: the altitude you are descending to.

We'll have the Landing cklist completed down to flaps 15 on the 737 (flaps 10 if single eng.) before turning on final.

Once established on final you'll descend to the MDA the same way: there again, you'll notice how easy and safe it is to descend worrying only about one thing: levelling off at the MDA.

You have certainly noticed how difficult it is for your friends who use the other philosophy to control the path of the airplane and make a decision at MDA: they have to fly a descending airplane, monitor the altitude not to bust the minima, look outside for the runway which can be offset by a large amount in a non prec. approach, find it, assess their height in relation to the PAPI to decide about the feasibility of the landing, ALL THAT SIMULTANEOUSLY.

Not so in our cockpit: after a nice and smooth leveloff at MDA, the airplane will be stabilised, we'll know for sure on which side of the profile we are, and we'll have plenty of time to set the Missed approach altitude in the MCP, look for the runway, find it, and make our decision.

Another important point to consider in bad weather: if the ceiling is roughly coincident with your MDA, if you descend with a constant rate and plan to decide at MDA, one little limb of the cloud is sufficient to keep you in IMC and force you to goaround! while with a longer level segment at MDA you have many more chances to exit the cloud and become VMC.

Ok, Pittsle, now we are approaching the VDP and have the runway in sight: as soon as the PAPI becomes three red one white, you call flaps 30, set Vat, and leave the minima on the PAPI and land.

If the minima is lower than 500ft, you have two options:
you can plan a flaps15 landing, or you can arbitraryly set MDA=500ft.
This way you ARE configured by 500ft.


Summing up: you know exactly what is happening instead of hoping for a miracle to end up exactly at the VDP at MDA, you know what to do, much easier job thus safer job, and what is a sort of nightmare to more than one pilot becomes very easy and effective.

Once again, if you master this scenario, you can apply the same rules to an easy one (as described at the beginning) or whatever else you like, even the constant descent, of course if you are in CAVOK even my dog can fly a nice touristic approach



One last pearl of wisdom : busting the minima by 30 or 50 ft, although very very ugly, has never killed anybody.
Being flown by the airplane, instead of flying it, like the Guam ****, has.

LEM

[Flight Safety]

LEM, outstanding post. It seems clear enough to me.

[411A]

LEM...indeed seems to have learned in the old school of ops, hence he certainly has the right idea.

It is a shame that many newer guys today just seem to never have been taught or, as sometimes is the case, a group in a particular airline 'thinks' that they have a better way...often it is not.

Flying level at MDA is required for the issuance of the FAA ATPL.
No performance...no license, simple as that.

And, it is not all that difficult in a heavy jet transport, even with an engine unserviceable.

This is just why many aircarriers demand flying ability from their flight crews.
Automatics off...fly the aeroplane.

[Wing Root]

Pelican's Perch #24: Sloppy, Sorry VNAV
I think this is a good article written by John Deakin. It explains the issues involved when CANPA can work and when it can't. There is a follow up article too.
Pelican's Perch #25: How I Learned to Love CANPA
-----------------
And from an Australian perspective...
DME Arrivals - Especially for those who use Airservices DAPs since Jeppesen present them slightly differently as I understand.
The DME Arrival is a procedure unique to Australia and involves descending in steps at certain DME distances on a particular inbound track to an aerodrome. These charts actually have the phrase "AT DME or GPS DISTANCE...NM DECEND TO....FT" Seeming to discourage a constant descent.
Here is an example

BUT, in the latest amendment, they have basically made them look like the profile view of an ILS approach and now present an altitude/distance scale to give a constant descent to the MDA. I.e. arrive at the MAPt at MDA. Another confirmation that constant angle approaches are becoming the norm. Is this good or bad?
Here is an example
and here
Looking at the first arrival from TBD I can't see how arriving on top of the airport at MDA is any help to you when you could be happily (and safely) flying along level at the appropriate MDA for your aircraft category before the MAPt ready to manoeuvre within the circling area to line up with a runway. But in the second example in Sector A they do depict a level segment. I guess this is to keep a standard 3 degree approach angle from 6 miles out. The actual chart used to fly the approach without the explanations is here.
Sorry for my ramblings I just found all this rather interesting. I am keen to know how pilots that perform DME arrivals actually do them.
Any comments?

[Pittsle]

Good discussion!

That´s what I like about pprune: those good discussions about a subject!

I really enjoyed the "pelicans perch" link.

Any more of those around ??

[Wing Root]

Pittsle,
There are plenty more Pelican's Perch columns about many topics here. He is probably most "famous" for his material about operating piston engines. I find his writing very interesting. It’s funny to note that he is now flying an aircraft that is totally computer and glass cockpit oriented – see here.
There are many more good columnists at avweb too.

Enjoy!

[alf5071h]

One more head bang against the cultural brick wall; or an alternative view for the undecided, or just a personal view?

Approach procedure design, to a certain extent, considers the inaccuracies and variables of a range of pilot skills when flying NPAs; very little in our industry is precise and many operations are based on probabilities. A skilled crew should be able to calculate the required vertical rate and timing with sufficient accuracy to arrive at, or close enough to the visual descent point (VDP) when using a constant descent angle stabilized approach. The risks due to any position error at MDA, whether continuing to land or go around, should be less than the risks involved with attempting a late descent from level flight or ducking under before VDP: - and for those who think otherwise, there is no requirement to fly level until MAP. The majority of these risks are due to human inadequacies or failings, notably the failure to follow procedures by attempting a steep approach after VDP, or accepting a high landing airspeed or unstable approach due to a late configuration change.

For those crews who think dive and drive (D&D) is easier; is easier an alternative description for lazy or lacking in those airmanship skills required for NPAs? If you are not very good at something (time / vertical speed) then practice should improve it.
Just because a national authority (or their representative) reportedly requires level flight does not mean that it is the best option. Our industry is still evolving; we still have to incorporate many of the hard learnt lessons from other’s accidents.

I prefer to have the lowest workload when flying a NPA. With D&D or ‘stepping down’ there is additional workload with altitude selection, power adjustment, and aircraft trim (trim changes due to changes in power, configuration, and airspeed) and, although use of autos can reduce this workload, the monitoring tasks in a dynamic situation are more complex. Alternatively CANPA gives a stable and consistent basis for monitoring during all NPAs. The aircraft should be stable in airspeed, vertical speed, pitch, and configuration. You can check that the flight path is at or is above each step-down altitude during the descent. There is reduction in the number of altitude captures; for every altitude change there is the risk of an alt bust, for every selection of the MCP there is opportunity for error, hence avoid the step downs and level off at MDA.

Alternative landing configurations, procedures, or approach speeds lead to inconsistencies, each with opportunity for error or requiring higher mental workload. The process of choosing which landing option to use, which configuration, speed, etc requires knowledge and experience. Not all crews have the opportunity to gain these attributes, and even less time to practice them. Thus the industry needs to look for simplicity through consistency and familiarity.

Approach timing is important, but timing or the available time is important in assessing the visual situation leading to a decision. Flight with an intermediate configuration often requires a higher approach speed which decreases assessment time. A level flight path gives a higher nose attitude than when descending, this reduces the available visual segment that could be seen on which a decision is to be made. Although the decision to land may ultimately be judged on what is seen at the far point of the visual segment (lights / runway), the ability to decide also depends on how long something has been seen. In low visibility the first point of contact is at the ‘near point’, a function of the cockpit cutoff angle and pitch attitude, the earlier the first point of contact the longer time there is to assess the aircraft position and flight path with respect to terrain etc. Thus use of the optimum landing configuration and a continuous descent improves the likelihood of seeing sufficient visual reference and maximizes the time for a quality decision.

There may not be a clear cut solution to the problems represented by the differences in culture seen throughout this thread, but here are some thoughts:
Avoid NPAs: National authorities / Airports take action. Operators use FMS if capable.
Use charts with range altitude checks; if the procedure has none then operators supply a crib sheet with details or at least tables for calculating timing / vertical speed. (The more astute may decline to operate to a runway in those circumstances).
Minimize the opportunity for error; publish and use consistent procedures, avoid configuration and speed changes. Report all ambiguities, both to your organisation and to the authority.
Strengthen error detection procedures; publish the required crew calls (and the limits of deviation) for non stabilized approaches.
Practice the error recovery techniques: go around.
Do not rush; maximize the time available for monitoring, checking, and deciding.
Report the violators; CFIT accidents are not Captain’s accidents, they are crew accidents, you are all in the same aircraft.

[LEM]

alf, I understand from your post you don't even select the various intermediate mandatory altitude steps in the MCP, but directly the MDA, and you merely monitor to check if you actually clear those steps.

My God, do you really mean that?

And do you really think not calculating the required VSI for a CANPA is a matter of laziness or difficulty?

[Right Way Up]

411a, when I say dirty dive, I mean some pilots were getting visual just before MAPt at the threshold at 300 feet and trying to get in from it. I personally prefer a constant descent approach. but there is always more than one way to skin a cat. Whichever way you fly the approach the MDA should reflect the type of approach you fly so as not to bust the minimum DESCENT altitude.

[LEM]

Right Way Up,

diving after the MDA has nothing to do with "diving" before the MDA.

It must be absolutely clear in every pilot's mind that the MDA can be left only on the PAPI, unless you fly a Piper or an Islander, of course!

Actually, what happens more often, is that with CANPA they end up high and have to dive - yes dive - to make the landing, while with the other method there's plenty of time to correctly capture the PAPI.

You are right when you say flying level fully configured is bad, but that's not the case: we have repeated various times that you fly level with the approach flaps setting, not the final one.

One last question to those who say a level segment is unacceptable in a large aircraft:
How do you fly a circling approach at the minima?

Oops, sorry, I forgot in your world a good alternate with a good ILS into the wind is always available!

[Right Way Up]

LEM, see my earlier post. I was explaining why many UK companies originally changed to constant descent approaches. This is because of an increasing incidence of pilots diving for the runway when they were well past a reasonable descent point.

[Tinstaafl]

I really like the new format DME Arrivals. What a great change from the old format. Wish the UK & USA had DME/GPS Arrivals available.


As for how I do them (erm.....used to do them, when I was in Oz): Pick whichever sector or approach track is convenient then I aim to clip each step down altitude limit at the DME distance that marks the start of the next, lower altitude limit. I assess each descent segment at the half distance mark ie I should be halfway through the required altitude loss halfway between each DME step distance.

When the final step down to MDA occurs I descend as rapidly as is reasonable to get to the MDA ASAP to give me more time to get visual & orient myself.

Bear in mind this is a CIRCLING approach, not a RUNWAY/LANDING approach.

[LEM]

Mike,
1) as I said in my first post, the scenario I'm talking about is the "difficult" one, not the easy one.
No doubt if you have all the magic available (glass, FMS, Vpath) you can simply push a button...

What I've often seen in cockpits not that advanced, is that when they try to fly a continuous descent they end up high...

2) flying a circling app. on autopilot or manually is not the point.
The point is somebody said that flying level on a large transport aircraft is unacceptable.
Thus my question about the circling...


3) for every selection of the MCP there is opportunity for error

As for not selecting the various intermediate mandatory altitude steps in the MCP, as far as most approaches go in Europe, as alf says, you can check that the flight path is at or is above each step-down altitude during the descent by very slightly modifying the ROD.

I don't believe my eyes!
Ok, gentlemen, from now on we'll throw the MCP ALT window out of the window, because every selection is dangerous, and when given descent clearances, or when shooting a STAR or an approach, we'll simply MONITOR to make sure we clear all obstacles!

[Crossunder]

Totally agree with Mike there!

SAS & KLM use CANPA (don't remember the other ones but there are a lot of Companies doing it. Braathens, Malmø Aviation, Austrian and Lufthansa if I remember correctly). But the EAG (that make the charts for these companies) always have a DME vs. ALT table on their charts, and Jeppesen hasn't yet gotten around to doing this as far as I can see; only partial tables not covering the entire approach. And this is sort of a requirement for flying a CANPA IAL; the PNF must call each DME and state "On profile or slightly high - next D13 @ 2360ft" etc. This way both pilots will stay in the loop and are forced to monitor the AA every NM. Works like a charm, and you won't get any mode confusion (I'm sure we've all tried changing mode just as the AP goes into capture - and you end up with either PITCH/IAS or remove the ALT SEL or something...". You'll end up with a stable ROD, only small power changes are needed and you'll arrive at minimums with a small pitch-down, thus giving much better forwad visibility! Just like on an ILS.
I first thought that this procedure would give increased workload, but once you get used to it - and that won't take long - it's much easier and more comfortable than the old D&D style approaches.
But on to the original question; how do I fly my NPAs? Well, as it happens to be, my present employer hasn't yet introduced CANPA, and I sure miss it. So much easier than the seemingly endless fiddlin' with the MCP that we're forced to do with our current SOPs. But instead we're waiting for the CAA approval for FMS/baro-VNAV approaches, which will be esentially a CANPA. I've tried it during the initial testing phase, and let me tell you that D&D is a thing of the past...

[Pittsle]

Please allow me to mention, that my in company CANPA´s are flown, but with the use of MDA as a decision altitude without any add-ons for altitude loss during missed approach.

And that the missed approach altitude is set in the MCP when the descent to MDA is started.
No setting of stepdown fix altitudes or the MDA itself.

I take it from the discussion that this is really not standard procedure in other airlines.