Pelican's Perch #24:
Sloppy, Sorry VNAV
Flying a non-precision approach has traditionally been a
I suspect Cap'n Jepp is rolling in his grave over the latest Briefing Bulletin — Vertical Navigation (VNAV), dated 12 NOV 99 — from Jeppesen in Denver. Not that it's Jeppesen's fault; they're just charting the latest industry foolishness. But Captain E. B. Jeppesen was an airman of the old school, and I believe he would have rejected VNAV descents during non-precision approaches in all but a very few cases.
This time, it's not the FAA, although I don't exactly see them doing any good in the area, and they're buying into some pretty bad procedures. As Pogo said, "We have met the enemy, and he is us!"
VNAV, in this case, refers to the deplorable tendency to make wimpy gradual descents during the final descent to MDA (Minimum Descent Altitude) during non-precision approaches, when an immediate step-down is better, followed by a period of level flight at MDA to the MAP (Missed Approach Point). Some pilots will even work out the math in advance, and preplan a rate of descent that will put them at the MDA at the MAP at about the same time! I've had people do this on check rides, and when I fussed at them for not getting down sooner, they look puzzled and say "But I got down at the MAP!" Yeah, but barely! I don't fail them for it, but it will draw a comment from me in the debrief. You're supposed to get down in good time to make a normal landing on the runway, and it doesn't matter where the MAP is!
Many pilots seem to think this glide-slope-like descent is a classier or easier way to do this maneuver. Maybe it is easier, but it often doesn't get the job done, it's not the way the procedure was designed, and above all, it can be dangerous.
With the advent of glass cockpits, inertial systems, RNAV, approach-approved GPS, and fancy computers, this technique is even being institutionalized and computerized, with "pseudo-glide slopes" being constructed and put right in the databases of these sophisticated systems. And now, alas, Jeppesen is formalizing it with the data right on the approach plates, presumably for everyone to use!
Yes, yes, I know, it says, "Use of this descent angle by certified VNAV-capable avionics equipment will ensure a stable, constant rate of descent that will clear all intervening altitude restrictions." I suspect that one-liner will get lost in the shuffle, and we'll see a lot more people trying to use the technique without adequate equipment. Further, even with adequate equipment, we'll see a lot more missed approaches. Meanwhile those who know what they're doing, and who do it in the classic manner, will quietly, legally, and safely make the approach and landing while the glass cockpits go merrily to the alternate.
Bad idea, I say ... BAD idea.
Back to Basics
Bear with me, please, this may be a little different way of looking at this subject than you've seen before. First we need a couple of definitions, and a mini-review of the basics.
There are two basic types of SIAPs (common shorthand for "Standard Instrument Approach Procedure"), the "Precision," and the "Non-Precision."
- The precision approach has vertical guidance for the descent.
- The non-precision approach does not.
Pretty simple, eh? (No, I'm not Canadian.)
While those are the basic definitions, "precision" also implies highly accurate horizontal and vertical guidance to, or very near, the landing area of the runway. Most common is the sixty-year-old ILS system, which provides a LOC (Localizer), and GS (Glide Slope), plus other components. There are a surprising number of other types, PAR, MLS, TLS, WAAS, LAAS, GLS, AND SCAT-1, but none are germane to this column.
"Nonprecision" approaches are "all those others," with no vertical guidance. Examples of non-precision approaches are Localizer-only, VOR, ADF, Loran, RNAV, current GPS, etc.
All the rules for constructing these approaches are contained in FAA Order 8260.3B "United States Standard For Terminal Instrument Procedures (TERPS)." It is very important stuff for those designing SIAPs, but bedtime reading material for the rest of us. It's heavy going, with lots of angles, formulas, and very rarely any explanation of just WHY the designers chose to do something a certain way. I have to move my lips when I'm trying to puzzle something out from it.
If you've read even a little of TERPS, it should have become crystal clear that the procedures shown on those approach plates should be followed, and followed very, very carefully, lest the ground rise up and smite thee. The Rockwell Number of granite is much higher than that of aluminum — or skin and bones!
There are two further approach variants beyond "precision" and "non-precision." One is the "Straight-In," and the other is the "Circling" approach. The latter is not well understood by many general aviation pilots and for that matter, some airline pilots who fly only into airports served by an ILS. In fact, most airlines now flatly prohibit circling approaches entirely, unless they're done in basic VFR conditions. If they can't land straight in with lower weather conditions (preferably using ILS), they go somewhere else. I suspect ole Cap'n Jepp would shake his head at that, too, for the death of circling approaches represents the loss of a useful tool, and it's another loss in overall airmanship.
In fairness, a low circle with a jet transport is not a trivial matter, and unless the airline is prepared to spend a lot of money on periodic training on this specific maneuver, the risk is probably not worth the reward. Until recent years, the visual displays in simulators were simply not capable of showing the view during a low circle with any accuracy, and mostly as a result of this, they are generally banned for jets today. Ironically, the very latest simulators do a credible job at showing the low circle in real time, but I guess no one has the heart to bring them back. Too bad, I love them — at least in the simulator. A low circle in minimum visibility with one engine out, and a 50-knot direct crosswind will get the old adrenaline a'pumpin', for sure!
VORs, NDBs, and Other Strange Critters
But for the purposes of this column, let's narrow the subject down to straight-in non-precision approaches, primarily VOR and NDB, for these are the ones the glass-cockpit folks are a'monkeyin' with.
There are two separate and distinct requirements for a straight-in, non-precision approach to be established:
- The final approach course must lie within 30 degrees of the extended runway centerline.
- Whatever final course is chosen, it must comply with the obstacle clearance criteria in TERPS.
To give you an idea of obstacle clearance requirements, here is a generic picture extracted from TERPS, cleaned up, and some color added. There are many variations on this, but this will give you a general idea.
To give you some idea of how much variation there is, here's the specific diagram for the VOR approach:
There are a couple of interesting things about this diagram. Note the faint inner dotted line. While it is shown in all three segments, it really has significance only for the final segment, and the extension of those dotted lines form the boundaries of that final segment. The area within that dotted line can be up to 30 nm long! Its only purpose is to make the final segment wider if the VOR is further from the runway.
Many pilots seem stuck on the idea that a non-precision approach is to the runway, and some of the TERPS requirements seem to give that impression. It may even be technically correct. But in reality, a non-precision approach in a non-glass cockpit is pretty loosely attached to a runway, and in most cases, you'll be much better off thinking of it as simply an approach to the general area of the runway, or even to the airport itself. Here is perhaps the simplest possible alignment of the final approach course, from a VOR located away from the airport:
Note this diagram has nothing to do with obstacle clearance! This simply shows the limits for the angles of the straight-in, off-airport VOR. The perfect final approach course must fall within the blue area, and that area can lie anywhere between the 30-degree lines from the runway. In rare cases, it might even be a nice straight shot!
Now let's take a look at the on-airport VOR case:
At Last, a Point!
Yes, I'm finally coming to it.
Now, look at the blue areas depicted on the last two graphics. Those tell the designer of the VOR approach where the precise CENTERLINE of the approach path can be. The MAP must also be within the blue areas.
But that's perfection, theory only!
TERPS also allows for errors in the ground facility, errors in the airborne radios, and some considerable sloppiness of the pilots, too (well, for me, at least). This is where the obstacle clearances enter the picture, and there are a whole bunch of graphics and text describing those. All you need to understand here is that those final approach paths are ideals, and the reality can be very different.
By way of example, at the legal maximum, a VOR located 30 nm away, the protected primary area at the runway will be FIVE miles wide, or two and one-half miles each side of the exact centerline.
Maximum permissible error of your airborne VOR is four degrees, give or take, right? You are supposed to know what the actual error in your equipment is, and allow for it, but all things considered, four degrees of total error (ground radios, too) is not an unreasonable thing to think about. Four degrees of error at 30 nm is about two miles off-course, barely within the TERPS criteria for protected airspace! Remember that lurking granite, just beyond the obstacle clearance boundary.
For a Localizer approach, the glass cockpit folks can punch in their data, hit the FAF or the "pseudo-glide slope," and slide on down just like it was a real glide slope. The VNAV path is computed to the landing threshold, and the navigation is very accurate, being continuously updated by double-DME fixes, or GPS, or the like. I agree, that's nice.
Where that CANPA (Continuous Angle Non-Precision Approach, a new term) intersects the MDA, a VDP (Visual Descent Point) is established.
The other non-precision approaches are done the same way with glass cockpits, but unlike the Localizer, there is the added complexity of the final approach probably not being aligned with the runway centerline.
If that VDP on the VOR or NDB approach just happens to align pretty closely with the runway, then it too becomes a good deal with the glass cockpit magic, because you'll break out in a good position to just keep on descending to the runway without changing a thing. Once again, the extreme accuracy inherent in the glass cockpit assures that you will be right on the approach path plotted by the designer, at a perfect altitude to just continue with a nice, stabilized descent, but possibly with up to a 30-degree turn to the runway (new criteria may limit this to 15 degrees for these approaches). I'm still not fully happy with this, because there are a lot of times you can do it the old way, break out in a clear spot, then maneuver just a little bit to stay clear of clouds (sometimes called "Dive and Drive"). Doing it with the magic will cause more missed approaches, but perhaps it is worth it, in terms of requiring less airmanship and skill, which makes it safer.
Why is the glass cockpit so magical, and so much more accurate than the older "steam-driven" cockpits as found in 727s, older 737s, and all 747s except the -400?
The key is that the glass cockpit is NOT using the VOR for the VOR approach at all! Yes, you read that right. The folks on the ground could shut the VOR station down right in the middle of the approach, and it would not only not affect the approach, but the folks in that glass cockpit probably wouldn't even know it!
How does this happen? These cockpits are usually using inertial reference systems, with automatic updating from any available source. It's a pretty smart system, it will usually pick the most accurate possible method of updating itself, and it's doing it all the time, as long as any radio facility is available. If GPS is hooked into the system, that'll probably be the primary source for updating the inertial units. Lacking GPS, it will probably auto-tune two DME facilities as close as possible to 90 degrees apart from the airplane, do the math, and slew (update) the inertial systems as needed to match. The result is that the nav computers in the airplane know exactly where the airplane is, down to a few feet with GPS updating (or GPS as the primary), a few hundred feet with double DME. If there are no outside radios available, the inertial systems are capable of great accuracy all by themselves. We still go across the Pacific Ocean on the inertial units alone, updating only when "coasting in." (Most airlines have not adopted GPS, yet.)
Now, if you tell this "magic" to "go directly to XYZ VOR," it does NOT tune in XYZ, and go there. It mumbles to itself, takes a quick peek into the database and looks up the latitude, longitude, and elevation of the XYZ VOR, does a quick calculation, and says to itself, "If I'm at this lat/lon, and XYZ is at that lat/lon, the course from 'here' to 'there' is 123 degrees."
It gets better. It "remembers" where "here" was, and where "XYZ" is, and does whatever is necessary to get the airplane on the Great Circle between those two points, keep in on that course with great accuracy, and headed in the right direction. Whether XYZ VOR is working or not simply doesn't matter, it's probably never even used, or tuned by the system! When you program the magic for the approach, it hits XYZ dead on, available or not. After XYZ, it not only does all the real-time number-crunching to keep the airplane on the Great Circle to the MAP, but it also computes the proper descent angle to clear all crossing restrictions to arrive at the runway end. All this time, it's also driving a nice big color display in the cockpit with a moving map, and a glide slope needle, keeping those humans informed.
("What's it doing now?" "I dunno, but it's doing it with fantastic accuracy!")
You may wonder "Well, why not just go ahead and tune that VOR to be safer?" Go ahead, try it. You'll be appalled at the wild needle swings, and the common errors in the VOR system! You'll sometimes see that VOR needle swing full-scale, while the magic calmly flies a dead-straight course, never wavering, never seeming to correct. If I'm destroying your faith in VOR, GOOD! VOR navigation is slap-dash at best, and always has been (though still better than ADF). Take THAT a step further, and recall the last time you shot a VOR approach, with high terrain all around!
Now perhaps you can see why the magic will do such a great non-precision approach. It will reliably place the airplane exactly on the precise centerline the approach designer selected.
If you use the magic to do the non-precision approach to minimums using this "Constant Angle Non Precision Approach (CANPA)" with a ceiling at minimums, you will have only a few seconds to look for the airport, because now the MDA altitude acts just like a DH/DA, it becomes a DECISION POINT. You might well pop out of a minimum ceiling, not see the airport, and have to immediately execute the missed approach, because by definition, you've popped out at the VDP, and if you go past that, you won't be able to land safely even if you do spot the runway.
Without the Magic, Back to the Basics
If you use that shaky old VOR to get there while trying to make a gradual descent of some sort, and even if you time it right to hit MDA right at the VDP, your position when you pop out of the ceiling could be anywhere at all, well outside those blue areas! You'll still be safe from the terrain, because the design of the approach (probably) gives you sufficient room for error, but "Where's the airport, Charlie?" It may be left, it may be right, it may be underneath where you can't see it. You probably won't have the time to find it visually, and even if you do see it, it may take a near-acrobatic maneuver to get to the end of the runway, while completing the checklist, getting final flaps out, etc. Unless you just happen to get lucky, all you can do is miss the approach, try again, or go away. Yes, it's real easy to sit in a comfortable chair behind a desk, as chief pilots do, and say "Well, you shoulda done the miss!" The reality is that at the end of a long day, maybe the final landing at home with a hot date waiting, you just may make that split-second decision to try it. "Yeah, it looks a little worse than I'd like, but let's give it a try, we can always do a go-around, right?" Try this in the rain, at night, with a black hole and a bit of turbulence, and you've got a recipe for disaster.
Without the magic, the old way is better. That means you descend early to MDA, level off there beneath the ceiling (or even still in the clouds), and start looking for familiar landmarks, or signs of that elusive runway. How early do you start down? Well, remember that once you cross the last fix before the MDA leg, you could, if you had a helicopter, descend vertically just past that fix to the MDA. Few of us want to carry it to that extreme, but for very good reasons, it's more than appropriate to do it the way Captain Jeppesen did in the early days of instrument approaches, descend as rapidly as possible to MDA, and level out there. For practical purposes, 1,000 feet per minute after the fix is a good target to use, although some higher speed jets will need to use a bit more to even get down before the VDP or MAP, on some approaches.
Once at MDA, maintain at or above it (for check rides, it's +50, -0 feet) until the airport is in sight and it is time to descend, or until the MAP, then miss. Normally, descent to the MDA will put you beneath the ceiling, giving you a good deal of time to look for the runway long before the glass cockpit folks would have even popped out of the clouds. You will have an early opportunity to make any needed turns for lining up with the runway, and you will probably need that maneuver room, considering the inherent inaccuracies in the non-glass, non-precision approach. Your only job will be finding the runway, then descending on a reasonable final approach to touchdown.
All that said, I don't care how you get to that final fix before letting down to the MDA. You can step it down, or you can fake some sort of easy descent path, but I do suggest you hit that fix AT the minimum altitude, and fully configured for a fairly rapid descent. As soon as that fix is passed, DUMP IT, and hustle right on down to the MDA just as quickly as you can get there safely ("Dive and Drive.") You may need to put out gear and full flaps, and reduce power to a very low setting, but whatever it takes, short of Herculean efforts, GET DOWN to that MDA! As you approach the MDA, you'll want to ease the power up, and perhaps pull the flaps back up to some reasonable approach setting. On the prop airplanes, it's best to get back to a configuration you'd use with an engine failure, so you don't have the cleanup problems if one does fail. By doing the approach in this manner, you should catch sight of the runway very early, and will have plenty of time to get the flaps (and possibly the gear) back down again. Of course, if the airplane will fly well with an engine failure and the gear down, you probably ought to leave the gear down when you hit MDA, so there's one less thing to do.
Sometimes the old ways are still the best!
NOTE: I am indebted to Wally Roberts, the world-class expert on TERPS, who has taught me so much but not enough. He kindly skimmed this column for gross errors. He gave me the new buzzword "CANPA," and suggests it's a good thing for jet transports. I agree with that for the glass cockpit magic, but not for the "steam-driven" jets with the old electronics, many of which are still flying. For the average GA airplane, the old way is by far the best.
Be careful up there!