When things don't add up

A month ago, I wrote a post explaining how that 747 might have landed at the wrong airport. Instead of calling them stupid or incompetent or joining in the chorus of indignation, I simply showed that it was an easy mistake for a human to make. I've been flying long enough to know better than to say "that could never happen to me", so I just gave a calm analysis and my best guess as to how it did. A month later, after some more information has come out, it turns out my guess was pretty close.

This week's AOPA/Air Safety Institute quiz was on the LOC BC-B approach into Rogue Valley International (KMFR) in Medford, Oregon. To keep my skills up, after I take the quiz I usually load up X-Plane or Microsoft Flight Simulator X and fly the approach. They're interesting challenges, since they're obviously not going to select a run-of-the-mill one for a quiz. This week was par for the course: a DME arc over high terrain to a localizer back course with a steep descent to a circle-to-landing. This is one of those approaches that leaves you feeling mentally drained after touchdown, even when you're just practicing from the comfort of home. That's why I enjoy flying them so much: they're excellent practice exercises, as Bob Hoover would agree.

Don't even think of using this for navigation.

Everything started off rather uneventfully. I tuned the Nav 1 radio to I-MFR and the Nav 2 radio to the OED VOR, which is the way I like to set up the radios when I'm flying an arc to an ILS or localizer.

I selected the OED 216 radial on the OBS, then loaded the approach into the GPS to use for a backup and a situational awarness cross-check. Since FSX isn't perfect, it didn't have WISEP intersection, but it did have one called D216X, which is the same thing.

It all looked good so far, so I took off and made a straight-out departure. I kept climbing until I intercepted the outbound radial, turned left, and kept on climbing to 9800, which is the minimum altitude for the procedure turn. Another 20 miles of seemingly straightforward flight followed on the way to the 24 miles from OED mark, which is where the DME arc was supposed to begin.

Both of those are tuned correctly. So why aren't things adding up?

However, as the distance increased and WISEP/D216X got closer, the needle wasn't doing quite what I thought it should be doing. It wasn't off by enough to be blatantly wrong; instead, it was just enough to make something in my intuition tell me something wasn't adding up, but not enough to ring any alarm bells.

I chalked it up to me perhaps being rusty, since this was only the second time this year I've shot a DME arc. After all, arcs are extremely rare nowadays, since you'll almost always get radar vectors instead. Of course, the thought that I hadn't even started the arc yet, so it couldn't just be that, hadn't occurred to me at the time. Hindsight is 20/20.

By the time I'm getting to about a mile away from the point where I should be starting the sharp left turn to begin the arc, things have really begun to not add up, and I'm really starting to get a bit confused. The GPS shows that I'm already past that point, yet the DME still says I have 0.7 miles to go!

The GPS claims I'm a mile past the beginning of the arc, but the DME says I'm only 23.3 miles out. Obviously, the GPS is wrong, because I could never make a mistake like that.

I dismissed that as having something to with FSX's database not having WISEP, so maybe the rest of the procedure was off by a little bit in the database. To make matters worse (or, more accurately, to make it easier for me to talk myself into believing that), there is an unlabeled intersection (the triangle on the GPS that the top of the red oval cuts through) almost exactly where I made up my mind that WISEP should be.

So just before I reached 24.0 DME, I began my left turn. I added in 10 degrees of right crab to compensate for the 20 knots of almost direct crosswind. This should have kept me from being blown inside the arc. It didn't, so I added five more degrees. Still getting blown inside. Added five more. Still getting blown inside. Added 10 more degrees of crab just to claw my way back to 24.0 miles. By this point, where I think my heading should be and what I have selected on the VOR disagree with each other by quite a bit.

After some effort, I've gone about halfway through the arc and have it bracketed well enough that I'm doing a good job being "on course". Unfortunately, at the halfway point, instead of fighting to keep from being blown inside the arc, I'm starting to drift further and further outside the arc, even though my heading "should" be correct. By the time I've figured out how to keep it from getting worse (which means I'm only halfway to figuring out how to fix it), my lead radial is only 10 degrees away.

A lead radial is something that tells you when you're getting close to the end of your arc and will be making your next sharp turn inbound. Ideally, if you do everything perfectly, you've sacrificed the right amount of chickens, and the stars are in perfect alignment, as you finish your turn, you'll roll out exactly on your inbound course.

Not this time. As the lead radial centered, I began a standard rate turn inbound. I rolled out on my inbound heading, but the Nav 1 display showed that I was way left of course. If anything, I should be right of course because the wind is now off my left side. No problem, easy to fix, and in just a few moments I'm back on course. For the first time, I'm actually on course and don't have to add scare quotes around it.

So I look at the chart to verify my stepdown points. I notice almost immediately that what I had read as 9800 feet is actually 9300 feet, so I'm already 500 feet high. Not a big deal, since I still have several miles to get down to 9300. But after that came the facepalm moment that explained the last half hour of confusion.

I double-checked to make sure my stepdown fixes were off the localizer's DME and not the VOR's. Then, like a good instrument pilot should, I verified that I had the correct DME source selected. Then it all made sense: after I tuned the Nav 2 radio to the VOR before takeoff, I never switched the DME source to Nav 2.

The entire time, I had been flying a ragged 24-DME arc off the localizer, not the VOR! Those two things are almost seven miles apart, as this shot from SkyVector shows:

Who's the idiot who decided to put the VOR at the top of the big hill instead of on the airport? (Oh, wait, that's where it's supposed to go. Signal propagation and all that.)

So I went back to where the GPS said I was past where I should have been but the DME said I hadn't reached it yet. I flipped the source switch to Nav 2 like it should have been the whole time, and lo and behold, the GPS wasn't lying to me—I was lying to myself. Quite convincingly, too, since now I was obviously 0.8 miles past where I was supposed to be.

It's only one little switch. How much difference could it make?

So you can see why I didn't get all smug when a professional aircrew landed at the wrong airport. Mistakes can and do happen to all of us. After all, I'm not a new instrument pilot; I'm a CFII with a decent amount of experience, and as you can see, I'm not invulnerable either.

In real life, this wouldn't have gone on as long as it did, because a real controller would eventually have asked why I was several miles away from where I should have been, and I wouldn't have so quickly dismissed the disagreement between the DME and the GPS. If this had happened to me in a real aircraft, I would have coupled the autopilot to the GPS and let it fly the arc while I figured out the problem. When you have a safe altitude, as is the case here, shedding some of the mental load to the autopilot lets you free up some of your mental resources to think things through. Just don't use the autopilot to do your thinking for you.

Humans make mistakes all the time; you just don't hear about them in the news every day because they didn't hurt anything except someone's pride. In fact, as Tom Turner over at Mastery Flight Training pointed out after the 747 at Jabara incident, an Antonov AN-124, which is a huge cargo aircraft that is almost as big as the 747, did exactly the same thing a decade ago. You just never heard about that one because they turned around and took right off again, since they needed a lot less runway.

The crew of National Airlines Flight 193 (for just one example of many) ignored the signs that something wasn't quite right and they ended up putting a perfectly-good 727 in the water, killing 3 people and injuring 11. So when things don't add up, don't just blithely continue on and pretend that reality is wrong because you just have to be right. That's just one more way that flight lessons make good life lessons, even if you're learning those lessons the hard way.