Mike's Musings #8: Jet Aircraft IFR Flight Testing

By Mike McCarthy (20 March 2007)

With the kind permission of webmaster Nels Anderson, at some point in the future portions of this article will be republished in modified form on my pending writing web site, thewritingblog.com. There you will find articles and photography from, I hope, several different authors, on a broad range of subjects.

"The time has come", the Walrus said, "to talk of many things:
Of shoes and ships and sealing wax, of cabbages and kings,
And why the sea is boiling hot - - and whether pigs have wings."

Lewis J. Carroll, "The Walrus and the Carpenter"
From "Through the Looking Glass, and What Alice Found There"

The FSX 737-800 exiting the Great Plains. FSX Default Scenery(!)

What you have begun reading is really three articles in one. It discusses many things, the three articles being ...

A discussion of FSX and its default B737-800.
An IFR crash course (forgive) including links to additional tutorial information.
A detailed description of the promised jet aircraft flight test procedure.

The more I work with FSX the more I like it, especially the graphics and especially the default 737-800. While this aircraft lacks an FMC, it is satisfying in most other respects because once again Microsoft did their 737 homework regarding what the typical simmer is likely to want. This is the aircraft we will fly today.

As you know, FSX does have performance problems out of the box, but surprisingly good performance can be had by following the procedures I gave in my FSX performance article series. (Read the series backwards beginning with "FSX Nirvana Demystified" located here.)

My flight test procedure is designed to test every aspect of jet aircraft performance. Because the test sequence is the same every time, and because I have flown it so many times, I am soon alerted to any unusual aircraft behavior. Because joysticks differ in their behaviors, to make sure that aircraft comparisons are on an equal footing all test hand-flying in this article and in the reviews to come will be done from the keyboard with no modifications to the default keyboard response sensitivity.

The flight test sequence description below is very detailed, one of my goals being to acquaint people with the joys of serious IFR flying. You will have fun repeating the procedure yourself under autopilot control. At a later time, when you're feeling macho, you might see how much of the arrival and approach you can fly by hand. At the moment I'm really rusty so today I'll only be hand-flying a) the departure and b) the final approach beginning past the outer marker. (The point where the glide slope is intercepted.)

This is a very long article, perhaps a record setter, so for those of you who are already familiar with IFR and are easily bored I'll simply summarize what it takes to really put a jet aircraft through its paces:

Evaluate aircraft ground handling while taxiing to the active runway.
Evaluate takeoff performance, especially distance to Vr (rotation airspeed).
Evaluate manual flying characteristics during a noise abatement climbing turn.
Evaluate the autopilot's performance from course capture to station passage.
Climb to FL320 and then accelerate to a cruising airspeed of .81M.
Once stable in cruising flight, execute a (manual) procedure turn.
(Manually) initiate a nominal 2,000 fpm descent to 9,000 feet MSL.
(Manually) turn due east and maintain that heading for about 50 nm.
(Manually) make a hairpin 140 degree right turn with a heavy load of fuel.
(Manually) maintain the resulting heading while descending to 7,400 feet.
(Manually) intercept the 260 radial of the unusually placed DVV VOR.
(Manually) track in on that radial till the I-JOY localizer comes alive.
(Manually) jog left to intercept the localizer.
(Manually) execute the ILS approach.
Evaluate distance needed to slow the aircraft and then to turn off the runway.

Now, Easily Bored, in the immortal words of Chuck Berry in "Schoolday", close up your books, get out of your seat, down the hall and into the street. You may safely skip the remainder of this article.

As for the rest of you, will you please climb into your first officer's seats and strap yourselves in.

Okay ... If you already know a lot about IFR you should skip to the word "diagram" below. However, if you're new to IFR there is some reading you should do after you ride with me. If at any point I confuse you, somewhere in the set of articles linked to here will be the information you need, presented with greater clarity than I am capable of. Here's a list of authors and topics ...

Nels Anderson - - A comprehensive aviation glossary by the FlightSim.Com webmaster himself. Please look up every term you encounter that I fail to explain, yes? You're learning a new language and his glossary article will be your dictionary. Find it here.

Nels Anderson - - VOR navigation 101 and 102. Nels is a real world pilot of considerable experience. His Piper Archer N8439T was the foundation for the corresponding DreamFleet/Flight1 Piper Archer add-on product reviewed here His VOR navigation articles are here and here.

Ron Blehm, Andrew Herd - - More about navigation (article 1, article 2), supplemented by information about the navigation instruments found in modern airliner glass cockpits here.

Ron Blehm, Andrew Herd - - ILS approaches and ILS approach plates, here, here and here.

Neville van Eerten - - Graduate school: DME arc and NDB approaches.

Jacques Zahar - - Graduate school: How to fly a hold.

Andrew Herd - - Post-doc exercises: Four dangerous approaches explained in gory (no pun intended) detail, including Aspen airport, KASE, where I once successfully landed a lightly loaded 747 on my third or so attempt. I dare you. (1, 2, 3, 4).

Here's a diagram of my standard jet aircraft flight test route, crudely drawn on a stitched together FAA high altitude enroute chart. Please note that "CRIMMS" should have read "CRIMS". By the way, note that by international treaty all VORs have three-character identifiers, all intersections have five-character identifiers, and all ILS approaches have four-character identifiers beginning with I- as in I-JOY.

Flight Test Route

North up, aircraft at center, runway 25 to the Southwest.

Okay. Let's do a flight test of the FSX default 737-800. Mind you, this is not so much a review of the aircraft as a discussion of the flight test procedure itself. We begin with the aircraft parked at the default KDEN gate Gate A34 since this will provide us with a taxi route that is as short as possible to the departure end of runway 25.

After pushback, and after we exit the apron, our taxi route will require a 90 degree turn to the left and later a 90 degree turn to the right. There will be opportunities for braking, and for applying thrust to get moving again. All told this exercise will tell me most of what I want to know about the aircraft's ground handling characteristics, the rest of the information coming from a high-speed turnoff after the landing which will conclude the test flight sequence.

At the gate we take on a full load of fuel. This is deliberate because I want to make the aircraft unreasonably heavy for the approach and landing, and especially for the hairpin turn toward the FQF VOR. (When you're hand flying, this one's a [censored] if you're trying for precise navigation while attempting to maintain a precise altitude. In fact, a precise turn here can only be done manually, and it must be a steep turn, which is why I want the aircraft to be very heavy.)

Now I'm ready for cockpit prep. Before I can forget, regardless of aircraft type I turn on every available kind of aircraft exterior heat. In the case of the default 737-800 we do this from the overhead panel, turning on probe (pitot) heat, wing anti-ice and engine anti-ice. These settings ensure that Flight Simulator will not do us in as we get into the extremely cold air of Contrail Country, with the potential for icing of the pitot tube and other aircraft components. At this point I also resize the overhead panel so that if I pull it down later it won't obscure the autopilot controls.

Next, I preset the autothrottle to 240 knots IAS (indicated airspeed), the heading to 030 degrees (see below for the reason), the altitude to 13,000 feet, and the rate of climb to 1800 feet per minute. Clicking on the radios icon, I resize and reposition the radio stack over part of the EICAS so that I can leave it open all the time. (EICAS stands for Engine Indication and Crew Alerting System.)

Cockpit Prep Completed
With the radio stack available this way I set NAV1 to GLL 114.20 with a course of 000 (to), setting DVV 114.70 as the standby NAV1 frequency. I set NAV2 to CYS 113.10 because we'll need it later to identify the CRIMS intersection, assuming we elect to begin our procedure turn there rather then further north. In the NAV2 standby frequency window I set I-JOY 108.90 as a reminder of the ILS frequency we'll be using on the return leg's approach to KDEN runway 26.

Next I switch the MFD (multifunction display) to VOR mode, automatically getting rid of the airports overlay which shows up in map mode. (There are bugs in the FSX map mode display filters so I rarely use map mode, especially because there is no way to change the map scale.) I then click on the pedestal icon and manually position the pedestal in the upper left corner of the glareshield, shrinking it to a miniature. This will allow me to monitor the mechanical response of the thrust levers themselves, and it will give me an instantaneous visual way to preset power settings based on previously-determined thrust lever positions, assuming that I have taxied and/or flown the aircraft before. (The setting of principal interest to me is the one for level flight at 9,000 feet or so at an airspeed of 220 knots or so, allowing me to maneuver easily on approach without exceeding the speed restrictions for flight below 10,000 feet MSL (mean sea level). This power setting is usually in the vicinity of 60% N1, roughly the same as the 50-60% N1 required for taxiing.)

Now I push back and begin taxiing to the departure end of runway 25. Holding short, I apply five degrees of flaps and review the rest of the cockpit setup. I then set all of the visuals sliders to the left, except for frame rate, which I lock at 30. You see, my goal here is not to look at scenery but instead to evaluate the aircraft's performance on a typical computer. Locking at 30 fps with the absolute minimum of scenery ensures that the Flight Simulator flight dynamics model will be operating smoothly, even on a limited computer like the one I'm using today at my office.

Finally, I turn onto the runway and apply takeoff power, after which I oversimplify things regardless of aircraft type: If the aircraft is a typical jet fighter I rotate at 120 knots IAS (indicated airspeed). (I haven't flown a fighter in years but that's the number I recall.) If the aircraft is a relatively lightweight transport like our 737-800, I rotate at 140 knots. If the aircraft is a relatively heavy transport like a 747, I rotate at 160 knots. If it's something extreme like Concorde, or an SR-71, I rotate at 250 knots.

Anyway, I rotate at the "appropriate" airspeed for the aircraft type. We are now airborne and, as you know, during departure things can get very busy very quickly. Hand flying while maintaining the runway heading, once the gear is up I oversimplify again regardless of aircraft type, going from five degrees of flaps to two at an airspeed of 175 knots, to one degree at 190 knots, and to zero at 205 knots. (For extreme aircraft you are on your own. I have no recollection of what the generic flap retraction or extension schedules ought to be.)

With the aircraft cleaned up I continue to hand fly, beginning a fairly steep simulated noise abatement turn to the right. KDEN is out in the boonies so strictly speaking this isn't necessary, but I'm evaluating aircraft performance and noise abatement turns are a fact of life for departures from KJFK, for example. Anyway, my goal here is to assess the stability of the aircraft in a bank of 20-30 degrees to the right while climbing at 1500-2000 feet per minute.

Can I remove my hands from the controls and have the aircraft continue the climbing turn I have set up? For the default 737-800 the answer is yes, if the bank angle is about 25+ degrees and the climb rate is about 2000 fpm. The bank angle does slowly increase, but I elect to ignore this since there will not be enough time before our next maneuver for the bank angle to become unacceptable.

The takeoff and subsequent stabilization of the aircraft's climb rate has taken us well past the 180 radial of GLL. Radials are always named as bearings from the relevant VOR. Thus the 180 radial of GLL points south away from the station. We want to track this radial north, so our course must be the reciprocal of 180, which is 000, but as we do we will still be on the 180 radial until station passage, after which we will be on the 000 radial. If this isn't clear, stop here and read webmaster Nels Anderson's VOR navigation primers.

Anyway my turn to the right is to a heading of 025-030 or so, my goal being to intercept the target 000 northbound course at a roughly 25-30 degree angle. I get us established on the intercept heading of 030 while still hand flying. Then I engage the autopilot, selecting NAV and altitude hold, or whatever the particular aircraft's autopilot settings need to be to achieve the desired effects of a course of 000 and an altitude of 13,000 feet at a climb rate of 1800 feet per minute. (On our FSX default 737-800 panel, NAV is called VOR/LOC.)

When altitude hold is engaged during the departure phase of this test flight, if the frame rate had been locked to a low value such as 10 fps, the FSX built-in autopilot would have tended to oscillate around the target climb rate of 1800 feet per minute, though the excursions would have damped out within a few cycles. With our sliders-left 30 fps philosophy we see no such behavior because of the smoothness which with the flight dynamics model is operating.

Soon we are climbing easily, and as we pass through 10,000 feet the ATC (air traffic control) airspace speed restrictions come off and I advance the autothrottle setting to 270 knots, another aircraft-independent oversimplification. With NAV mode already selected, we will soon test the autopilot's ability to capture and hold a radial, and selecting altitude hold will test the autopilot's ability to capture and maintain both a climb rate and an altitude target without porpoising.

Okay ... With the departure completed, now we can finally sit back and relax, waiting to see how the course and altitude intercepts unfold.

By now we are beginning to approach our northbound course of 000 as defined by the 180 radial of the GLL VOR. The altitude intercept goes beautifully with no sign of porpoising. On the intercept of the 180 of GLL the autopilot does fairly well. At a distance of about 25 miles from GLL it begins hunting to either side of the target radial, but the initial excursion is less than a dot and the hunting soon damps out. (Each dot represents five degrees of deviation from the commanded course. The resulting position error is small when you are close to a VOR, very large when you are distant from it.)

By the way, the thin vertical white line represents the aircraft's current heading. It is pointing straight up because the MFD is in VOR mode, and it would point straight up in MAP mode too. Were it not for an apparent bug in the MCP we would be able to put the MFD into PLN (plan) mode, which would give a north up presentation. The whole MFD presentation would then be rotated thirty degrees clockwise. I mention this because PLN mode does work in other aircraft so you should be aware of the difference between MAP (based on aircraft heading) and PLN (always north up).

During this period I dial 23,000 into the autopilot's altitude window. The aircraft begins to climb again, this time with no sign of porpoising, even at 10 fps as shown by an earlier test flight.

Intercepting the 000 Course at a 30 degree angle
By the way, the 13,000 and 23,000 altitude settings simulate a controller-ordered ATC step climb to cruising altitude in the Denver area. Once you are comfortable with IFR, you really should try flying online. I will be writing about this at a later date but in the meantime you might want to visit www.vatsim.net, sign up, and simply jump right in. Stay out of really busy airspace till you know what you are doing, but don't hesitate to fly from any secondary airport to any other secondary - - you'll be given all the help you need, including vectors to final approach if you request this.

Anyway, during the climb the airspeed holds fairly well at about 265-270 knots IAS. During this phase of the flight, at constant IAS our Mach number gradually increases as the air becomes less dense. When we intercept .60M, at around FL230, I will switch from KIAS hold to Mach hold, this being another simplification I always make for non-extreme aircraft.

Shortly before we reach GLL, ATC clears us upstairs to our cruising altitude of FL320, so we will not need to level off at FL230 after all. However, things get busy again because while we were receiving this clearance and twiddling the target altitude setting, we also had to monitor the airspeed in order to make the switch from KIAS hold to Mach hold as discussed in the previous paragraph.

The next area of investigation will be what the FSX built-in autopilot does as we approach and then pass over the GLL VOR. Will it do the right thing? No, it will not. Instead of simply maintaining heading during station passage it begins to "chase the needle". The swings get wider and wider until finally the autopilot loses course lock altogether and it becomes necessary for me to revert to manually-adjusted heading hold in order to get back on course. I am able to accomplish this by the time we are 20 or so nautical miles north of GLL. (20 nm.)

[Note that the FS2004 built-in autopilot exhibits also exhibits bad behavior during station passage, but unlike the FSX version the FS2004 autopilot does eventually recover course lock, though it takes a while for the oscillations to die down. Nevertheless, if you are willing to tolerate the FS2004 behavior, you will tolerate the FSX behavior.]

For simplicity I have shown our procedure turn to the south as beginning at the CRIMS intersection, as might be the case if we were flying a general aviation aircraft. We would fly to the CRIMS intersection as follows ...

At the CRIMS intersection
If I hadn't done it earlier, I now switch the MFD to VOR mode and then click the center button to expand the presentation to a full 360 degree compass rose. The NAV1 GLL bearing selection is now depicted by a white-colored directional arrow which straddles the entire compass rose. Because we are now north of GLL, and because our desired course is the 000 radial of GLL, this arrow is pointing straight up, indicating "from" (white arrowhead pointing south) because we have passed GLL. NAV2 is tuned to CYS, the bearing arrow shown in green. This is kind of presentation is different from older mechanical "steam gauge" HSIs (horizontal situation indicator) in which NAV1 will be presented as a white single-line mechanical arrow while NAV2 will be presented as a white double-line arrow.

CRIMS is on the 125 radial of CYS, so all we have to do is wait for the green tail of this arrow to arrive at the 125 degree (from) mark on the compass rose. Now you know why Microsoft sold you two navigation radios.

Ordinarily, once at CRIMS we would execute a procedure turn. The purpose of this maneuver is to get us headed south on the course we had been on when the procedure turn was initiated. The required bank angles will vary with airspeed but there are rules of thumb for this and other aircraft evolutions. A good source of such rule of thumb information is at http://www.scenery.org/faq_aviation.htm

However, on today's test flight we can't turn at CRIMS because we are in a jet and we haven't reached our cruising altitude of FL320 yet. Thus we are going to have to run well north of CRIMS.

Some older aircraft, like the 727-200, have difficulty maintaining a thin air climb rate of 1800 fpm. Accordingly, since we're not supposed to know anything about the aircraft we're flying, at FL260 I reduce the climb rate to 1,000 fpm, and I reduce it again at FL300, this time to 500 fpm. At FL320 I might further reduce to 200 fpm, or even 100. At some point the aircraft will no longer be able to climb. This will be its service ceiling for the current weight of the aircraft. Anyway, these step reductions in climb rate will allow us to maintain airspeed without overstressing the engines or stalling out.

The aircraft finally reaches FL320 and levels off when we are roughly 70 nm north of GLL. However, I know from having made several of these test flights that the FSX built-in autopilot sometimes glitches on thin air altitude intercepts if the climb rate is perhaps 1,000 fpm or greater. Disengaging and then reengaging altitude hold seems to fix the problem, but it tends to recur, perhaps as we are transitioning from one weather "tile" to the next. Therefore you might want to set the rate of weather change to low, or even to zero. In the end, having had enough of this nonsense, I simply engage static real-world weather so that weather refresh becomes a non-issue.

With all the autopilot altitude glitching, we don't become stable till we are 130 nm north of GLL. Having earlier advanced the autothrottle to .81M before the glitching set in, I now decide to run north for another 20 nm to make sure that things truly are stable. We will then be ready for our procedure turn. Here's how that evolution is done ...

In a slower aircraft we would begin by making a 45 degree turn to the right and then holding the resulting heading for one minute. In aid of this there is a chronometer at the left side of the 737-800 instrument panel. Its CHR button, in the upper left corner of the instrument, first starts, then stops, and finally zeroes the time sweep.

After one minute on the 045 heading we would then start a standard rate turn to the left and continue that turn to the reciprocal of the original heading. A standard rate turn is 360 degrees in two minutes, or three degrees per second. The bank angle in degrees required to achieve this turn rate varies with airspeed, the rule of thumb being to divide the airspeed in knots by ten and then to augment the result by 50%. In a general aviation aircraft flying at, say, 200 knots, the required bank angle would therefore be roughly [200 / 10] + 10 = 30 degrees.

Outbound Procedure Turn Leg Completed
However, in jet aircraft, at high altitude cruising airspeed I can't pull this trick because at 300 knots IAS the required bank angle would be 45 degrees ( [300 / 10] + 15), while most transport aircraft autopilots limit the bank angle to a maximum of 30 degrees. Thus the only way to compensate for this while flying on autopilot is to extend the outbound leg to two minutes or so, which will allow for the inevitable too-wide left turn to the south, letting the autopilot do the turn itself at the maximum available bank angle of 30 degrees.

(By the way, if you can hand fly this procedure turn to within 250 feet of our cruising altitude you will be doing extremely well. I find it difficult to remain within 1,000 feet. C'mon - - take the challenge.)

Note that on our default 737-800 you must use the outer right portion of the heading dial to maximize the bank angle limit. Also, with any autopilot, at this point in our test sequence if you simply dial in the 180 heading, the aircraft will turn right rather than left, the

Procedure Turn Completed
autopilot having been designed to seek the shortest turn that will meet the stated target of a 180 degree heading. Accordingly you must walk the heading control into the left turn until the aircraft has passed through 000 degrees. At that point the ambiguity is resolved and you can set and forget the 180 heading.

What counts for me here in this situation is not whether the procedure turn is executed with the utmost precision but rather whether I have managed to get turned around without substantially over- or under-running the new course of 180. Precise procedure turns were important during the days before VORs, and they are important today for NDB approaches, but we are so far from GLL that there will be plenty of distance in which to recover from any course error.

In spite of the improvised two minute outbound leg, the result is a procedure turn completion that brings us back within a dot of our target course of 180. (Note that course is where you want to be over the ground, while heading is the direction you are pointed.) By the way, you did remember to change the OBS from 000 to 180 during the procedure turn, yes? Otherwise the needle would now be luring you away from the desired course instead of steering you toward it, yes? Think about it.

We are now about 170 nm north of GLL, too far north to receive DVV. By occasionally flipping the NAV1 switch we will be able to tell when we come within DVV's range. Once we are within range we leave DVV active, and we see that we are now 130 nm north of it.

At present we are at FL320. I would like to have descended to 9,000 feet by a point 30 nm north of DVV, where we will begin our turn to the east. This trip down to BOD (bottom of descent) will require an altitude loss of 23,000 feet. Given a descent rate of 2,000 fpm this will take [23,000 / 2,000] = 12 minutes or so, which is 12 / 60 = .20 of an hour. Since our groundspeed today is 550 knots (90 knot real world weather tailwind component!), the descent will require about .2 x 550 = 110 nm. (There are rules of thumb for this stuff but I want you to see the principles involved. Anyway, the thumb rules don't consider headwinds or tailwinds.)

If we were to consider only these issues, the descent should therefore begin 30 + 110 = 140 nm north of DVV. However, in the immortal words of the usual late night TV product announcer, that's not all. We will need at least 10 nm to slow the aircraft from .81M to .60M, maybe 20 nm if we are in a "slippery" aircraft like the 727-200.

Since we supposedly are unfamiliar with our aircraft, we'll average the two numbers and allow 15 nm to get the ship slowed down. This now puts TOD (top of descent) at 30 + 110 + 15 = 155 nm north of DVV, a good 25 miles behind our current position of 130 nm north of DVV.

Thus there is trouble in River City. We have several choices regarding being late in starting our descent.

Our first choice is to cancel this review. However, even though doing this would save me the time of finishing this piece, and you the time of reading it, I'm not going to do it because I need the shameless self-promotion which each published article offers.

Our second choice is to travel back in time. Does this sound ridiculous? Well, it's not. The FSX version of Pete Dowson's FSUIPC has the classic Dowson AutoSave logic built right into it. With FSUIPC4 installed and registered, all we have to do is click Flights, then Load, and then select an AutoSave flight which will roll the clock back by perhaps ten minutes ... Oops, I have FSUIPC4 installed on the office XP machine, but not here at home, where I have resumed writing this piece on the Vista machine, so ...

Our third choice is to increase the descent rate. We could recompute everything from scratch but let's not, if only because my calculator just died. (Truth.) I know there's a calculator built into Windows, but real men need only guesstimates, so let's instead observe that our overrun of TOD is about [25 nm / 130 nm] or roughly .20. Thus we need to increase the descent rate by about .20 x 2,000 = 400, giving a new descent rate of 2,400 fpm.

Okay ... I dial the authrottle speed window back to .60M. Then I cancel altitude hold. Then I preset 9,000 feet into the altitude hold window and -2,400 fpm into the rate of climb window. Then I wait ...

The aircraft reaches .60M when we are about 115 nm north of DVV, just about 15 nm from where we began the speed reduction, exactly per my estimate. Now I reengage altitude hold, beginning the descent proper. Then I wait some more ...

The default 737-800 is nicely stable in this situation. The rate of descent is nailed, and by FL230 or so the airspeed has come back to .60M after an initial excursion to .62M as we started downhill. As the air density increases, so does the indicated airspeed. When we intercept 270 knots we will switch from Mach hold back to KIAS hold, the reverse of what we did on the way upstairs.

We make this changeover at about FL220. However, the airspeed then gradually increases to 280 knots, presumably due to our relatively rapid descent. At about 13,000 feet our Denver sector approach controller clears us down to 9,000 but requests a speed reduction from 270 knots to 240. We change the speed mode setting accordingly. As we pass through 10,000 feet we are in violation because we have only slowed to 260 knots. Were this not a test flight I would have used the spoilers to get a quicker speed reduction.

Anyway, as we level off at 9,000 feet we are 45 nm north of DVV, not 30 nm. What went wrong? My calculations were approximations, but they were reasonable ones. Being off by 5 nm would be perfectly normal, maybe even 10 nm. But 15? No, something unusual must have happened ... Oh yes, I know ... Inspecting the MFD I see that our former 90 knot tailwind has turned into a 20 knot headwind, invalidating our assumption about groundspeed during the descent. Mystery solved.

While I'm waiting for approximately 30 nm to show up on the NAV1 DME, I preset 090 into the heading hold window, though I remain in NAV mode in order to hold to our southbound 180 course ...

... And when the NAV1 DME shows 33 nm from DVV, I click on heading hold, causing the aircraft to roll into our turn to the east. As the turn proceeds I rebuild the radio frequencies setup. Now I set NAV1 to FQF 116.30 with a course of 228 (to), and with a standby frequency of 114.70 for the DVV VOR. I set NAV2 to DVV 114.70, still with I-JOY 108.90 as the standby frequency. (My purpose here is to be able to form a continuous mental picture of where we are relative to DVV at the same time that we are maneuvering to get on the intended FQF radial.)

With the turn complete and our heading now 090, I begin the wait for the CDI (course deviation indicator) to show that we are approaching the desired FQF radial. If you're as impatient as I am, and as untrusting of electronics reliability, you may want to monitor our progress re FQF by reducing the course window number down from 248 to some lower value which proves that the needle is actually working ... But don't forget to restore the course window (OBS, omni bearing selector) to 248 at some point, yes?

Okay ... With the commanded course restored to 248, the needle begins to move when we are about 45 nm from DVV. However, if we're not quick to respond we may run to 50 nm from DVV, or even further. Anyway, whenever we do observe the needle moving, I then walk the commanded heading number down till we are past the ambiguity angle. At that point I will simply set the heading value to 248 and let the autopilot complete the turn on its own. Once the turn is complete I will remain in heading hold mode, accepting whatever course error exists.

Hairpin Turn Underway

Now I command a 500 fpm descent to 7400 feet, which is the altitude for intercepting the I-JOY glide path. Then I flip the NAV1 switch to make DVV 114.70 active, I dial 260 as the (to) course, and I set I-JOY 108.90 as the standby frequency. Next, I flip the NAV2 switch, thereby making I-JOY 108.90 active. (Trust me, you'll understand why in a moment.)

With the new radio setup complete I command an airspeed reduction to 180 knots. As we pass through 205 knots I command one notch of flaps. At 190 knots, a second notch. At 180 knots, a third notch, giving us five degrees of flaps if we are flying the FSX 737-800. We reach this condition when we are about 40 nm from DVV, allowing us plenty of time to catch our breath. Had we instead been on the official approach we would be rolling out onto the localizer course only about 15 nm from touchdown, similar to the high-pressure Stadium Approach into KLAX. This is another jet transport hand flying challenge - - try it.

At this point please pause the simulator and think slowly and carefully about what we are doing. If need be, review the preceding paragraphs, and most definitely look at my flight path diagram again. In addition, do Flights/Save now and name the situation so that you'll be able to return to it as many times as you need to. (When you restore the situation you'll find that the MFD was not saved correctly. You will have to twiddle the switches to get back into VOR mode.)

Now ... Even though we are not yet at the 260 course to DVV, I am nevertheless setting us up to begin the ILS approach to runway 26. We cannot receive the I-JOY localizer this far out which is one of the reasons the FAA placed DVV on the KDEN field, DVV having a sufficiently high signal strength such that we can receive it at quite a distance even at our low altitude of 7400 feet. (We will actually receive a different localizer, I-FUI, before we receive I-JOY, so don't be fooled by this. I-JOY itself will not be received until we're about 29 nm from DVV.)

The placement of DVV apparently is a compromise between the requirements for an approach to runway 26 and the requirements for various other KDEN approaches. In particular, with the compromise, simply tracking all the way in to DVV on a course of 260 would put us too far north of the I-JOY localizer to be able to execute a successful localizer intercept.

This is why we want I-JOY 108.90 to be the active NAV2 frequency. It will allow us to see when the localizer comes alive, even though at that moment we will not be aligned with the localizer. However, when the needle does come alive, all we will need to do is a) flip the NAV1 switch to make the I-JOY frequency active, and b) use the heading hold control knob to jog us to the south until we are roughly aligned with the localizer. At that point, if we have been flying precisely, we will be able to engage approach mode and let the autopilot take us to the OM (outer marker), where the glide slope will be intercepted automatically.

Capish? If not, my apologies, but do go back and review.

The rest of the approach is left to you the student as a homework exercise. If you learn to execute the entire test flight as described in this article you will understand almost every aspect of IFR flying, the only important things missing being NDB approaches and holding patterns, discussions of which were linked to near the front of this article.

Approach plates, airport diagrams, and charts for SIDS (standard instrument departures) and STARs (standard instrument arrivals) can be obtained through www.airnav.com, which serves as an easy-to-use portal into the FAA's USA chart database. These are high resolution charts which are downloadable, viewable and printable pdf files. The high-res approach plate for the I-JOY ILS can be downloaded here, and the high-res KDEN airport diagram can be downloaded here.

In low resolution these charts look like this:

I-JOY approach to KDEN runway 26

KDEN airport diagram

Mike McCarthy
mike@pcgamecontrols.com

Addenda ...

My first FSX computer was the ordinary Windows XP office machine I used to develop the FSX Nirvana articles. (Start with Mike's Musings #5 here and read backwards through the series.) This machine is typical of the recent computers of yesterday, having a 2 Ghz CPU, 1 GB of memory, a 256 MB GeForce 6100 graphics card, and a 7200 rpm hard drive. I was able to make this machine run FSX surprisingly well by applying the folder/file name defragmentation strategy and then balancing the system, both operations as described in the FSX article series.

The second FSX machine is one I bought recently for purposes of AirBoss™ product test. Typical of what's now coming off the assembly lines, it's installed at home so that I now have FSX test capability at both locations. The new machine is a Windows Vista computer with an AMD 2.2 Ghz dual-core CPU. Like the office machine it too has 1GB of memory, a 256 MB GeForce 6100 graphics card, and a 7200 rpm hard drive.

Neither machine is of high-performance. Both machines have had all of their boot-time startup services disabled except for their Nvidia display drivers. Both have 4095 MB PageFile.sys files. Thus the only significant difference between these two average computers is single-core-with-XP versus dual-core-with-Vista. I therefore expect to be able to make some comparative statements about the two FSX environments in the relatively near future.

There ... Award yourself an Air Medal for having read this piece all the way through.

[ Back | Home | Main Menu | Logout | Help ]