Non-precision approaches and descent planning

[Roger, Roger]

There is an ongoing discussion around my Skool o' Flite about non-precision approaches and when and how to descend.

There are essentially 2 schools of thought on the issue.

The official policy is that we teach students to stay at the the last published or assigned altitude until a normal descent will cause them to cross the FAF at the minimum altitude, then continue on downward at the rate of descent published on the Jepp chart for their groundspeed. The idea here is to not level off until reaching MDA. Here is a quick example: minimum altitude outside the FAF is 1700. You are being vectored at 3000'. ATC says, "fly heading 340, maintain 2400 'til established, cleared VOR-DME 31". Rather than immediately descending to 2400', then descending to 1700 when established on the FAC, you would maintain 3000' until about 2.5 minutes from the FAF (1300 feet of descending at 500 fpm), aiming to cross the FAF exactly at 1700 and continue to MDA without leveling off. If it helps at all, here is the approach chart I'm thinking of (assume the vector takes you in between KAPPS and ARGEN).
http://204.108.4.16/d-tpp/0904/00807VDT31.PDF

A simpler procedure, and one that I've heard is used a lot out there outside of the 141 flight training bubble, would go more like this.

When you recieve your approach clearance, hustle on down to 2400' (~800 fpm in a 172). Then, when established inbound between ARGEN and KAPPS, hustle down to 1700'. Lather, rinse, repeat until at MDA.

Now, I'll admit that the first certainly gives a smoother feel to the approach. However, we have people who will argue up and down that it is much, much safer than the second method because (I quote) "it makes it more like an ILS approach". The obsession with this technique borders on cultish.

Now I've seen this technique cause people trouble in a few ways. First, if you don't keep a close eye on both your rate of descent and your distance/time remaining, it can be pretty easy to end up too high at the FAF, or farther in. Think getting to MDA at the MAP. Second, it gets thrown out the window when people get stressed (think one engine gets canned in the Twinstar) and they're falling back to technique 2, because it's easier...but they're not used to that technique so they blow through altitudes.

Of course both of those problems can be solved by giving the student more experience in instrument flying, but it seems to me like the second technique is a bit more idiot-resistant.

I don't really buy the ILS approach argument because you have no vertical guidance. Altitude awareness is actually more difficult in the first method because you have to monitor ROD and time/distance a lot more.

Any thoughts from other CFIs?

Any thoughts from anyone else with extensive real-world IFR experience (airline and freight pilots)?

 

[CFIscare]

I was taught, and thus now teach, to perform the second way that you mentioned. I'm not sure I buy the "much safer than the second method because it makes it more like an ILS approach" train of thought. As long as you don't descend below MDA, there is no danger flying at or above MDA for a bit.

Especially for a new instrument student/pilot, there are more things you can do wrong on the first method than can affect whether you can safely shoot the approach, think seeing the runway right at MAP and diving down to the runway.

 

[MikeD]

Any thoughts from anyone else with extensive real-world IFR experience (airline and freight pilots)?

Well since you asked.

As you describe, the two accepted techniques are the so-called "stair step" method, and the constant descent method.

The constant descent method is ok, and generally works, but as you say...it does take some mental math to make happen, and I wouldn't want to use it at all from FAF to MAP...that final segment of the approach being where I want to get down to the MDA and get stabilized so I can find the runway environment and only have to worry about VDP, so far as any more descents go.......or get the setup going for circling, etc. Point being, the plane is stable, and I can devote more attention percentage to the outside, and have my inside crosscheck just looking for something deviating from stable (VVI, speed, etc); rather than assuring I don't blow through a level off, or otherwise make finding the runway more difficult because I'm adding extra work to my plate inside the final segment.

Also, another thing I don't like about the constant descent method is just what you mentioned......crossing the FAF a little high. Normally no big deal, but what if the altitude at the FAF (or at any of the fixes) is a hard altitude, not an "at or above" altitude? This was common in the AF, where the airfield's overhead traffic patterns have an outside, as well as an inside downwind. The outside downwind pattern flows planes around the field and sets them up for turning 5 mile initial for the overhead. The point where they turn initial at the 5 miles, geographically sits above the FAF for the TACAN/VOR/LOC app to the same runway; so the IAP plate will show the FAF altitude as a hard altitude. Constant descent makes missing this a possibility, step down lets you get there and not have to worry about it.

Now since I personally like the stair-step technique, I will add that there's no requirement to have to descent at 5000 fpm (exaggeration) from step-down to step-down; you can go rapid if by yourself, but if you have pax, you can always make the step-downs smooth and still accomplish them in an efficient manner.

So the BL is this, as I see it:

Constant rate descent: Will it work? Yes. Is it efficient? Not really, due to multiple changes happening at once, esp inside FAF, and exp if you have a VDP; since the descent rate will need to have to calc'd for going from FAF to VDP; not FAF to MAP.... circling to land being exempted from this.

Stair-step: Will it work? Yes. Is it efficient? Yes. Less going on at one particular time, easier crosscheck.

This is what i've found from a high number of actual hours in a variety of aircraft types and operations.

Your mileage may vary.

 

[Douglas]

I favor the stair step approach.
I see that your example requires DME, but I was taught in aircraft with out DME (for the most of the flights).
I teach in that environment now, too. If DME is not required they don't get to use the GPS for distance or time info.

Knowing when you are 2.5 minutes-out and timing a descent is really not practical for my situation. Though if I did use all the bells and whistles, a nice constant stable approach would feel better for pax.
I don't think there is a wrong answer, just preference.

 

[ppragman]

I favor the stair step approach.
I see that your example requires DME, but I was taught in aircraft with out DME (for the most of the flights).
I teach in that environment now, too. If DME is not required they don't get to use the GPS for distance or time info.

Knowing when you are 2.5 minutes-out and timing a descent is really not practical for my situation. Though if I did use all the bells and whistles, a nice constant stable approach would feel better for pax.
I don't think there is a wrong answer, just preference.

For me, it depends, but mostly when I was at ace, I didn't like the stairstep. In JNU, our special approach had about 7 stepdown fixes from ASORT. Frankly, it was a lot easier in the ice and turbulence to plan a descent that would put you at the minimum altitude, at BARLO rather than going down for a minute stabilizing, then starting a new descent. Just my $.02

EDIT: That being said, with your example, it might be easier to stair step it, as your first drop down is only 300' so you could easily go down to that and stabilize in the 6NM you have, i dunno, there's no concrete answer to these things. Each approach is different, and each day is different. What works on saturday may not work on sunday, do what feels most comfortable for you because this is a technique issue, not a reg issue.

 

[nosehair]

They both have their advantages and disadvantages. On a familiar approach in normal wind, a constant descent is much easier and smoother.

In an unfamiliar situation, I'd rather arrive at altitude a little early than late.

Your question raises the point that we should be teaching both methods as standard, and the PIC chooses which way depending on the situation.

 

[MikeD]

They both have their advantages and disadvantages. On a familiar approach in normal wind, a constant descent is much easier and smoother.

In an unfamiliar situation, I'd rather arrive at altitude a little early than late.

Your question raises the point that we should be teaching both methods as standard, and the PIC chooses which way depending on the situation.

Yup...both techniques with advantages/disadvantages, as well as time & place, depending.

 

[Orange Anchor]

The studies for airlines flying NPAs showed that the Constant Rate of Descent (CROD) or Constant Angle Non-Precision Approaches (CANPA) had a higher success rate for completion (ie landing) than the stair-step or what some refer to as 'dive and drive'. The subject has been hotly debated in the gen-av community as to which is better.

But the argument over which is better misses the point that NPAs are inherently higher workload and with step downs the potential for descending to MDA or a step down altitude at the wrong time is there.

At my 'house' a couple of requirements were added which caused some huff-huff in the pilot community. First was to round the MDA up to the next 100ft. And MDA of 520 became 600. Sounds like it would be a factor in increased missed approaches bue findings were the incident of misses was so small as to be negligible. Second, ALL NPAs had to be flown using the autopilot. This reduced workload and improved tracking.

Like other things (V1 for example) some airline procedures are adopted without finding out if they really are applicable to gen-av use for small recips. Sometimes the answer is yes. Sometimes no. I think the main thing is to understand what you are trying to do with the CROD or the D&D, respect the limitations of each, round UP MDAs and use the autopilot.

 

[SpiceWeasel]

Our airline does the dive-and-drive. VS1.4down. Ohhh yeah. That's how I learned it until I got to ERAU.... but I definitely prefer being able to spend more time looking for the airport environment stabilized at MDA.

 

[BrewMaster]

Considering a lot of people can't even plan a descent out of a cruise altitude, I don't think they could handle doing the math in the middle of an approach.

I was taught the stair step method and probably wouldn't consider the other way unless I was required to do so. I'm not close minded to it, but for now that's what I'm gonna go with. I would much rather get to my altitude and stabilize as opposed to just pointing her down and hoping the math is right or that I'm not going to bust altitude on a segment. As far as passenger comfort goes, the descent would have to be just about perfect (with no vertical guidance), and smooth to not have to adjust pitch at all.

To each his own, this is just what I prefer.

 

[Blackhawk]

Here is a quick example: minimum altitude outside the FAF is 1700. You are being vectored at 3000'. ATC says, "fly heading 340, maintain 2400 'til established, cleared VOR-DME 31". Rather than immediately descending to 2400', then descending to 1700 when established on the FAC, you would maintain 3000' until about 2.5 minutes from the FAF (1300 feet of descending at 500 fpm), aiming to cross the FAF exactly at 1700 and continue to MDA without leveling off.

I'm not sure if this is what you meant, but I do have a problem with not beginning a decent immediately after being given one by ATC. I have flown approaches, particularly to non-towered airports in non-radar environments, where airplanes are stacked up waiting for the approach. Even in a RADAR environment you could cause a traffic conflict and even a talk with ATC if you do not begin your descent to 2400' immediately- or whatever altitude assigned by ATC. ATC is expecting a descent, you begin a descent. How you execute the descent...

 

[minitour]

The official policy is that we teach students to stay at the the last published or assigned altitude until a normal descent will cause them to cross the FAF at the minimum altitude, then continue on downward at the rate of descent published on the Jepp chart for their groundspeed. The idea here is to not level off until reaching MDA. Here is a quick example: minimum altitude outside the FAF is 1700. You are being vectored at 3000'. ATC says, "fly heading 340, maintain 2400 'til established, cleared VOR-DME 31". Rather than immediately descending to 2400', then descending to 1700 when established on the FAC, you would maintain 3000' until about 2.5 minutes from the FAF (1300 feet of descending at 500 fpm), aiming to cross the FAF exactly at 1700 and continue to MDA without leveling off.

So you were assigned an altitude of 2,400' and you're just going to decide to stay at 3,000' until some arbitrary distance from the FAF based on your descent rate for the approach (which is only computed from FAF-MAP)?

What's that reg about complying with ATC instructions...?

A simpler procedure, and one that I've heard is used a lot out there outside of the 141 flight training bubble, would go more like this.

When you recieve your approach clearance, hustle on down to 2400' (~800 fpm in a 172). Then, when established inbound between ARGEN and KAPPS, hustle down to 1700'. Lather, rinse, repeat until at MDA.

Dive and drive? Eh...more later.

Now, I'll admit that the first certainly gives a smoother feel to the approach. However, we have people who will argue up and down that it is much, much safer than the second method because (I quote) "it makes it more like an ILS approach". The obsession with this technique borders on cultish.

It's nothing at all like an ILS approach unless you have vertical guidance and a Decision Altitude (height).

Now I've seen this technique cause people trouble in a few ways. First, if you don't keep a close eye on both your rate of descent and your distance/time remaining, it can be pretty easy to end up too high at the FAF, or farther in. Think getting to MDA at the MAP. Second, it gets thrown out the window when people get stressed (think one engine gets canned in the Twinstar) and they're falling back to technique 2, because it's easier...but they're not used to that technique so they blow through altitudes.

They need to be taught to think farther ahead. Part of the brief needs to be "we will descend at x feet per minute to arrivate at MDA y miles from the MAP...the right crosswind will put the runway off to the left side of the airplane, so we'll look there until the MAP." Something like that would be fine. Then they've already done the math and know what they're doing before they do it.

You wouldn't have them look at the missed approach instructions for the first time as they reach the MAP, would you?

Of course both of those problems can be solved by giving the student more experience in instrument flying, but it seems to me like the second technique is a bit more idiot-resistant.

IMHO (and this is just an opinion), the "dive and drive" technique is just a crutch for those that can't or won't get ahead of things when it's critical to be ahead of things.

I don't really buy the ILS approach argument because you have no vertical guidance. Altitude awareness is actually more difficult in the first method because you have to monitor ROD and time/distance a lot more.

I agree on the vertical guidance, but not sure why you wouldn't be monitoring ROD and time/distance using the dive and drive method as well. That's all stuff that should be part of the scan...especially on an approach. Next time you get the chance, throw them in the sim and when they're on their way to MDA, toss a Pitot/Static problem at them. See what happens.

[quopte]Any thoughts from other CFIs?

Any thoughts from anyone else with extensive real-world IFR experience (airline and freight pilots)?[/QUOTE]
Again, the following is only my opinion.

When you're given an altitude assignment, unless you have a valid reason...you go to that altitude. So when you're given "maintain 2,400' until established..." you descend to 2,400. This isn't "dive and drive" yet.

When you're established and it's an NPA, you go down to the altitude for that sector of the approach. I think (without looking back up) you said it was 1,700. We'll just use that as a number for now.

After the FAF, you use your computed rate of descent for your ground speed to arrive at the MDA at least a mile before the MAP (this is how I teach it). In slower planes like 172s, Duchii, Seminoles, etc. 1 mile is fine. It still gives them almost a minute to look around and, to be honest, if I'm still in the crap...a minute is all I want to spend down low, still in the crap, down by all of the radio antennas, rocks, buildings, guys taking off "VFR/IFR" out of uncontrolled fields, etc. That's plenty. Get to MAP and execute your missed approach procedure.

The other way, which I teach once or twice, but hate performing... is "dive and drive". It's the way my former employer teaches NPAs. From the FAF, you drop the gear and drop out of the sky at 1,000 fpm or more until MDA. Then you stabilize on your speed and look around. I don't like hanging out down low, in the soup, looking for something I probably won't see from 5 miles out anyway (if the vis is down to 1.5 miles...why such a hurry?).

I'd much rather see a constant rate of descent or, as was mentioned "CANPA" approach....from the FAF inbound. That's what I'd prefer to see, not what is 'right' or 'wrong'. Both techniques can produce results equally well.

-mini

 

[Roger, Roger]

First, I want to thank everyone for the input. It's been interesting and educational to read some of the different opinions and techniques.

I want to specifically address some of the things that Mini said.

So you were assigned an altitude of 2,400' and you're just going to decide to stay at 3,000' until some arbitrary distance from the FAF based on your descent rate for the approach (which is only computed from FAF-MAP)?

What's that reg about complying with ATC instructions...?

That has bothered me for a while. We're essentially teaching that altitude as pilot's discretion every single time. It is VERY standard on our flights at our airport, and ATC never bothers us about it. Typically if they want us at a lower altitude, they'll say "cross PODCI at and maintain 2500, cleared VOR-DME 31".

They need to be taught to think farther ahead. Part of the brief needs to be "we will descend at x feet per minute to arrivate at MDA y miles from the MAP...the right crosswind will put the runway off to the left side of the airplane, so we'll look there until the MAP." Something like that would be fine. Then they've already done the math and know what they're doing before they do it.

You wouldn't have them look at the missed approach instructions for the first time as they reach the MAP, would you?

I agree. That descent planning IS included in our standard approach brief. I believe that the real problem at the heart of these issues is simply a lack of the ability to think ahead and visualize what is going on. I hope it's not me, I give my students as many tools as I can think of for planning these descents (rules of thumb, quick tricks with mental math, etc) but I think at the most basic level, these people just don't have enough experience to keep themselves 3 or 4 steps ahead of the airplane, in IMC, when an engine quits. We can talk about the benefits of reduced-time commercial/instrument training later, but for now suffice to say I'm not a huge fan.

I agree on the vertical guidance, but not sure why you wouldn't be monitoring ROD and time/distance using the dive and drive method as well. That's all stuff that should be part of the scan...especially on an approach. Next time you get the chance, throw them in the sim and when they're on their way to MDA, toss a Pitot/Static problem at them. See what happens.

Interesting idea, that. Also, I agree it should be included in the scan. I just see it as consistently a weak area (not just with my students, either).

Again, the following is only my opinion.

When you're given an altitude assignment, unless you have a valid reason...you go to that altitude. So when you're given "maintain 2,400' until established..." you descend to 2,400. This isn't "dive and drive" yet.

When you're established and it's an NPA, you go down to the altitude for that sector of the approach. I think (without looking back up) you said it was 1,700. We'll just use that as a number for now.

After the FAF, you use your computed rate of descent for your ground speed to arrive at the MDA at least a mile before the MAP (this is how I teach it). In slower planes like 172s, Duchii, Seminoles, etc. 1 mile is fine. It still gives them almost a minute to look around and, to be honest, if I'm still in the crap...a minute is all I want to spend down low, still in the crap, down by all of the radio antennas, rocks, buildings, guys taking off "VFR/IFR" out of uncontrolled fields, etc. That's plenty. Get to MAP and execute your missed approach procedure.

The other way, which I teach once or twice, but hate performing... is "dive and drive". It's the way my former employer teaches NPAs. From the FAF, you drop the gear and drop out of the sky at 1,000 fpm or more until MDA. Then you stabilize on your speed and look around. I don't like hanging out down low, in the soup, looking for something I probably won't see from 5 miles out anyway (if the vis is down to 1.5 miles...why such a hurry?).

I'd much rather see a constant rate of descent or, as was mentioned "CANPA" approach....from the FAF inbound. That's what I'd prefer to see, not what is 'right' or 'wrong'. Both techniques can produce results equally well.

-mini

Thanks again, Mini.

 

[Orange Anchor]

Our airline does the dive-and-drive. VS1.4down. Ohhh yeah. That's how I learned it until I got to ERAU.... but I definitely prefer being able to spend more time looking for the airport environment stabilized at MDA.

How do you fly the final leg? Flaps and gear configuration?

My point is to maintain MDA with approach flaps and gear, you have a high power setting. BAM! there's the runway so you saw off a wad of power, drop the nose, retrim and go landing flaps.

What is the definition of an unstable approach (if applying airline definitions)? See previous paragraph.

 

[SpiceWeasel]

How do you fly the final leg? Flaps and gear configuration?

My point is to maintain MDA with approach flaps and gear, you have a high power setting. BAM! there's the runway so you saw off a wad of power, drop the nose, retrim and go landing flaps.

What is the definition of an unstable approach (if applying airline definitions)? See previous paragraph.

We have Flaps 45 and REF set between 1-2 MI prior to the FAF. Yeah you need a bit higher power setting when you get to the MDA, though the time you spend there depends on the approach, because doing 137 KIAS makes the 1.4 down seem almost slow. Once you pick up the visual glide slope, you're golden though.

I've done way more visual approaches than non-precision approaches where I'm employed.... 1 NPA so far in 1.5 years.

 

[Blackhawk]

One other point would be the use of a VDP, or if one is not listed, being able to compute a PDP(planned descent point). While a PDP may not be a factor flying a non-precision approach with an MDA of 500' AGL in a 172 to a 12,000' runway, it can be a factor in airplanes with higher approach speeds (or even in a 172 on some approaches).

 

[Fly_Unity]

I favor the stair step approach mostly cause in the mountainous areas the approaches are pretty steep and I like to be as low as possible because its hard for me to slow down and keep engine temps warm doing the whole descent at once. Also in my experiance Its better to get low and out of the ice as fast as possible, I always found that clouds tend to have more moisture at the tops. Depends on the day, and the area. Im not used to see FAF altitudes lower then 7 thousand either. I guess I never had a problem having to keep leveling off at every fix, Im usually struggeling to get down by the fix without busting everyone eardrums

 

[Orange Anchor]

We have Flaps 45 and REF set between 1-2 MI prior to the FAF. Yeah you need a bit higher power setting when you get to the MDA, though the time you spend there depends on the approach, because doing 137 KIAS makes the 1.4 down seem almost slow. Once you pick up the visual glide slope, you're golden though.

I've done way more visual approaches than non-precision approaches where I'm employed.... 1 NPA so far in 1.5 years.

And if the MAP is at the end of the runway, take about :30- :45 seconds off the timing since one goes from a normal glide path to a 90deg path using a 700fpm descent.

The other point is yes, the studies showed that the average airline pilot made 1-3 NPAs (as I remember) per year. For that reason many airlines first did away with NDBs. Many have deleted VORs.

But your procedures are not uncommon (app flaps, gear down) and at 400-500ft changing config, power and trim. And no one seems to notice the disconnect over stable approaches. And why? Because statistically it can be supported... until the first big accident.

 

[Roger, Roger]

So you were assigned an altitude of 2,400' and you're just going to decide to stay at 3,000' until some arbitrary distance from the FAF based on your descent rate for the approach (which is only computed from FAF-MAP)?

What's that reg about complying with ATC instructions...?

I just talked to one of our TRACON guys. He said that that last altitude, the minimum altitude until course intercept, is basically a suggestion/notification of the minimum vectoring altitude, so in this case it is essentially pilot's discretion whether or not to descent to that altitude immediately. Basically if they want you lower, they'll notify you.

 

[ClearedToThe]

If hand flying, the stair step is easier IMHO.

If you are flying the approach on autopilot, set in the published rate of descent from the plate and select vertical speed mode on the autopilot when crossing the FAF and have the altitude selector set to the MDA rounded up to the nearest 100ft. Of if your autopilot can handle altitudes in 10' incriments, set a DDA (Derived decision altitude). DDA = MDA + 50ft. This is a CANPA, as previously mentioned, and works extremely well if executed properly (even in a 172) and IMHO is much safer than the dive-n-drive. Not to mention passenger comfort.