Navigational equipment required:

Standard basic avionics fit plus VOR indicator, DME, and ADF.

Aircraft:

This tutorial can be flown with the Microsoft 737 and I have included situations for that aircraft. If you are unfamiliar with the flight parameters for the aircraft I have included a crib sheet for the 737 here.

Note that in the situations, the fuel load has been cut to about 33% for the sake of realism, which will allow you to land without using up the entire runway.

Scenery:

There isn't any Kathmandu airport scenery available for either FS2000 or FS2002 as far as I am aware, though there is some available for FS98. For the terrain, you might consider using Eddie Denney's RealSceneFS mesh for FS2000. The FS2002 default mesh is excellent and gives a very good impression of why the circling approach at VNKT is something pilots don't look forward to, although the airport is subject to one of 2002's Jeppesen/mesh data conflicts and stands on a low mesa, which isn't the way things are in real life.

Downloads:

For FS2000 - kathmandu2000flts.zip - extract these files into your \FS2000\pilots folder

For FS2002 - kathmandu2002flts.zip - extract these into your \FS2002\flights\myflts folder

View the approach plates by clicking on either of the images alongside (note that there is only one Jeppesen plate)

Comments:

The tutorial assumes that you understand the basics of VOR and NDB navigation. I am assuming that you have flown the earlier approaches in this series and have learned the techniques explained there.

Kathmandu is treated with extreme caution by most pilots, and with good reason, because the airport has a dramatic location. The runway elevation is at more than 4000 feet above sea level and the airport lies in an oval shaped valley surrounded by mountains which rise on all sides to nearly 10,000 feet. The approaches are based on a stepped DME descent to a VOR situated 0.6 nm short of the runway and there is no ILS, so that flight safety depends entirely on the ability of the pilots to adhere strictly to the published approach profile even in the most trying conditions. In the most extreme case, an airliner full of passengers has to be flown within 1.6 miles of the threshold and down to a height of only 900 feet before the missed approach point is reached.

There is little room for error in bad weather and crews must be even more alert than usual, because this is not a procedure which can be taken for granted. Approaches in poor weather have led to many accidents, for example PIA Flight 268, where an Airbus crashed with the loss of all onboard because the crew mistook the altitude restriction at 8.0 DME for the restriction at 10.0 DME and flew into a hillside at about 7300 feet. It is quite easy to understand why mistakes like this are made, because the approach requires high descent rates (up to 2000 feet per minute) in a mountainous area at a time when cockpit workload is very high, without the visual stimulus of a glideslope to alert pilots to the risk of too fast a descent profile. To make matters worse, the missed approach procedure on the VOR 02 procedure must be one of the most complicated ever published, and if you do manage to land, the runway has a pronounced hump, although fortunately for us this is not modelled in Flight Simulator.

Take some time to study the approach plate series, which has been made available to us courtesy of the SAS Flight Support. Please click the top thumbnail to access pages showing the charts.

Jeppesen have also allowed us to reproduce an approach plate and you can view this by clicking the lower thumbnail. This approach plate and others for this area, is available on the SimCharts CDs.

Instructions:

Please make sure you have Indicated Airspeed (IAS) set. If you have True Airspeed set, the instructions which follow will make no sense at all. I would suggest that the aircraft is flown using the Mode Control Panel (MCP) heading bug, altitude and vertical speed windows and other functions - I am aware that some simmers like to fly lateral navigation using the yoke, but this is problematic in FS2002 and isn't realistic, anyway. Ensure that the EHSI and the RMI of the aircraft are aligned, which is the default situation. Also, make sure that you do not have gyro drift checked under the realism settings, unless you are proficient at making corrections for this. Do not use the GPS on pain of death.

The tutorials are written for the SAS Flight Support plates, but I have indicated where the Jeppesen plates differ. While the screen shots are taken for FS2002, the instructions work perfectly well for FS2000 and should be OK for FS98, although I haven't tested that version.

Last of all, before we get going, if any real-world ATPs who have flown this approach would like to correct me on any points of procedure, I would be only too pleased to hear from them.

Talk Through:

Situation 1: VOR 286 non-precision approach -

FS2000 - Select situation Kathmandu2000circling

FS2002 - Kathmandu2002circling

This is a circling approach, which means that the published procedure will place you in a position from which you can make a standard overhead join, assuming conditions are VFR. The problem with the approach lies in the topography - the one thing you wouldn't normally choose to do is fly an airliner in the circuit in tiny valley surrounded by mountains, but you can take some slight comfort in the knowledge that the missed approach procedure is a simple "turn 'em and burn 'em" straight out climb to a hold at the LDK NDB. Remember things could be worse - you could be doing this one in real life.

The situation places you a few miles out from, and descending towards, the IGRIS intersection, identifiable as D15.0 KTM on the 106 radial from the VOR. Check the MCP - you need to begin a descent to 10500 feet, with Nav1 tuned to KTM on 112.3, and the ADF to the LNC NDB on 252. You are flying with the weather as it was on August 24th, 2001, and I strongly advise you to tune to a local ATIS and reset your altimeter as you go through 18000 feet, or alternatively you can do it the lazy way and just hit B on the keyboard. Anyone who has trouble with this concept may like to try the Aspen approach I did a while back (-: A rate of 1500 feet per minute should get you down easily; make sure to adjust your heading to keep the CDI centered - or select NAV mode on the MCP. Reduce power so that when you reach the target altitude speed has fallen to about 190 kias and drop 2 degrees of flap when the airspeed is on the button.

Take a minute to check out the plate as you descend and work through the checklists. You need a reported ceiling of 5150 feet to even contemplate this procedure, which we will assume we have been given - incidentally, don't fly this with ATC enabled, or they are likely to give you vectors which are nothing to do with the approach procedure and you are likely to end up decorating a mountainside.

From IGRIS, there is a slow descent to ECHO at D10.0 KTM, during which time you can establish the aircraft in approach configuration. Then from ECHO, there is a steep section with a gradient of 5.4 degrees to the LNC NDB at D6.5 KTM. This is an important fix, as you want to adhere exactly to the track shown on the profile; drift too far left and you risk hitting the two isolated peaks shown on the plate, too far right and you will overfly the airport, leaving you without enough room to get around. From there you continue a fairly steep descent until you reach the MAP, which on this approach is at 5600 feet, 1300 feet above ground level at D1.0 KTM. The plate makes it clear that you must reduce speed to a max of 180 kias below 10500 feet, which is a departure from the normal 250 kias restriction - the reason for this being that it would be impossible to execute some of the procedures at a faster airspeed. Screenshot 1 shows the panel at IGRIS.

In a sense I was quite fortunate with the weather for this approach, because it is just good enough to provide glimpses of the mountains sliding past, while at the same time it is just bad enough to make the flight IFR down to just before the MAP. I flew this one several times and you should get in, no problem.

Screenshot 2 shows the cockpit at ECHO. The aircraft is rather nose high because I didn't set any flap on the way down and so the autopilot is compensating by altering the angle of attack. Incidentally, you might like to take a look out the window at this point, or to pan around in spot plane view, because if you are using FS2002, or have mesh installed in FS2000, you will appreciate how tight this particular way in is. There is one route through the mountains and if you don't stay exactly on course, you won't be in it. While we are on the subject of difficulties, this is the best approach for appreciating Kathmandu's sloping runway, which sadly doesn't exist in Flight Simulator, but then neither do the flocks of birds which enjoy loitering on the airport. So cheer up, things could be worse.

Screenshot 3 shows the cockpit just after passing LNC. The descent rate I needed to get the aircraft down on time was around 2500 feet per minute at 180 kias. Now this is OK with a pressurised aircraft, but reduces the margin for errors, as the plane is virtually standing on its nose. One way around this would be to reduce speed and drop more flap, which would allow you to reduce the descent rate some and would also have the benefit of giving you a little more time to check things out - so feel free a descent at say, 160 kias after you have done this the first time.

Apart from the scramble to get down to altitude (and remember that the figures on the profile are minima, you can go higher if you want, but if you can't get down fast enough at the bottom, it may mean you won't be VFR at the MAP) everything else looks OK. The ADF is swinging as the aircraft crosses LNC, and it is time to tune it to 358, to pick up the LDK NDB, where the missed approach hold is set.

Screenshot 4 shows the aircraft going VFR as it nears KTM. If this hadn't happened, we could have flown all the way down to the minimum of 5600 feet at D1.0 KTM (at which point, incidentally, the threshold would be visible just ahead of the starboard wingtip, easy to miss in marginal weather), but as it is the runway is clearly in sight and so a circling approach can be made.

Whatever you do, resist the urge to pull a sharp right and slap her down. Apart from anything else, the plane is too close to the threshold of 02 to make short finals, given that doing so would involve a turn of more than 90 degrees, right on the end of the runway, which is hardly a safe option with only 140 IAS on the clock... OK, it can be done if in Flight Simulator if you have the reflexes of a cat, but just don't try it with real passengers, please (-:

A proper landing would involve descending to 6300 feet (the threshold elevation of rwy 02 is 4313 feet), making a 180 degree turn to cross over the end of runway 20 at 5300 feet, then turning through 90 degrees shortly after crossing the centerline and flying downwind parallel to the runway on a heading of 200 degrees, dropping 30 degrees of flap as you do so, before turning base and descending to land on 02. The tricky bit of this approach in Flight Simulator is getting your height right and making the turn onto finals at the right moment. In a real plane you can keep half an eye out the window as you do this to time your turn, but unless you are running something like WideView on multiple PCs, you won't have this luxury in FS, although opening another view window can help. In FS2002 there is a conspicuous road which makes a good visual reference point and as long as you begin your turn short of the road you should be able to get down without overshooting the centerline, but whatever you do, make sure that you are at about 600 feet above ground level as you straighten up over the extended runway centerline, or you may have to go around.

I agree that flying a circuit isn't the sort of thing you don't often get to do in a passenger jet, but then you don't usually get to land places where an ILS isn't provided either. For interest's sake, I notice that China Southwestern operates a flight from Lhasa in Tibet, to Kathmandu, using a 757 - it isn't unreasonable to assume that that flight might use this approach if conditions were favorable. Anway, I included a flight plan in the zip, which works in both FS2000 and FS2002 - enjoy the flight from Lhasa. I would suggest using a wee bit more flap than usual on departure. The exit is a tiny bit steep.

Situation 2 - VOR 02 non-precision approach.

FS2000 - Select situation Kathmandu2000MAP

FS2002 - Select situation Kathmandu2002MAP

This is the most commonly used approach to Kathmandu, and consequently the one on which most of the accidents have happened. If you study the plate, it isn't too difficult to understand why. The procedure begins at the NOPEN intersection, and after a gentle descent to D10.0 KTM, there are gradients as steep as 6.6%, before the profile flattens out towards the MAP, which once again is at D1.0 KTM. SAS Flight Support have taken the unusual step of publishing an extra sheet dedicated to the approach profile, which is presumably intended to keep their crews and passengers alive. This additional profile is well worth a look, because it compares a "flat" 4.7 degree descent to the published procedure, which includes a variable rate descent designed to give the crew as much opportunity as possible to go VFR, while keeping the plane as far away from the mountains as possible.

And, while I hate to say it... this might be time to take a look at the missed approach procedure. Who knows, we may need it...

Okay. The situation places you inbound to the SMR VOR on G336 at FL270. You have just changed from Calcutta to the Kathmandu FIR and you have been cleared to descend to FL150 at SMR. The weather at Kathmandu is reported as broken cloud with the base above the 5150 foot minimum and so you have informed the cabin staff that you are about to begin your descent, and reset your altitude window to 15000. Your route takes you to SMR on 357 degrees, which you must cross at 15000 feet, then you transition to RATAN on 054 degrees, finally altering course to NOPEN on 022 degrees, the IF for the approach, which you should aim to reach at 11500 feet. From NOPEN, you maintain the same course, descending to the 9500 feet to cross the FAF at D10.0 KTM (if none of this makes any sense, please go back and work through the introductory articles in this series).

Please make sure you stick to the airways - if you get tempted to stray off course, a glance at the Minimum Safe Altitude circle should cure any latent wanderlust. You definitely don't want to go too far northwest even at FL330 unless you are a dedicated mountain watcher. Incidentally, at this stage, if you are running FS2002, you can enter 112.3 into the standby freqency for Nav1 and 112.9 into Nav2. Not only will this allow you to change the Nav1 frequency quickly at SMR, setting up Nav2 to the VOR's frequency will allow you to use the RMI's DME to estimate when you are approaching RATAN, although you should make the turn when you intercept the KTM radial, rather then waiting for the DME to give exactly the right reading from SMR. The reason for turning on the radial intercept rather than the DME reading is that if you turn even slightly early or late at SMR, the distance to RATAN will be different to the one on the chart and as you can see from screenshot 5, it isn't desirable to fly on too far here.

So, begin your descent to SMR, at about 1800 feet per minute, 280 knots. At about 3.5 DME from the VOR (remember you are 15000 feet up, so you will never get closer to it than about 3 miles), alter course onto 054, towards RATAN, keeping the bank angle to about 15 degrees - with a bit of practice you should be able to make a perfect intercept. Now, if you have already set the course window to 054, you should have no trouble flying out on the radial. Reset the altitude to 11500 and keep descending, reducing airspeed to 180 at about 12000.

If you look at the STAR plate, you can see that the leg is 14 nm long, so if you have FS2000, tune Nav1 to the KTM VOR on 112.3 and dial 022 into the course window to set up the intercept. If you have FS2002, you can just switch the frequencies on Nav1 and watch the DME on the RMI. Don't get misled by the "41" next to the track on the FSS plate - that is the MEA, not the distance to go, which is 14 miles - and bear in mind, as I hinted above, that the most important thing is to intercept the radial, not to turn exactly on the correct DME, although it is desirable to do both.

Screenshot 5 shows the aircraft making the turn on the intercept of the 202 radial from KTM. Now we are inbound. Don't the tops of the mountains look lovely as they shyly peep out of the clouds all around you?

The NOPEN intersection lies at D16.0 KTM, and you have to get down to 11500 by then, so check that you are on target. Remember that you must keep under 180 kias below 10500, so this is a good time to begin reducing speed in stages. After NOPEN, you have to descend to 9500 to make the FAF at D10.0 KTM. Don't be afraid to use the airbrakes to get your speed down and whatever you do, make sure you drop five degrees of flap as you get down to 180 kias.

Screenshot 6 shows the situation at the FAF. There does seem to be a lot of cloud about, and tower confirms that visibility is deteriorating pretty fast, although their opinion is that conditions will still be VFR at the MAP, when the aircraft gets there.

On the basis of their advice, and your assessment of the situation, you decide to make the approach. As you near the FAF, reset the speed to 140 kias and drop the gear and lower the flaps in increments to 40 degrees. Bearing in mind our groundspeed will be about 160 knots, a descent rate of up to 1800 feet per minute will be needed until D5.0 KTM, so set the first altitude restriction of 8200 feet in the window and dial in a vertical speed of 1700 fpm, which should be about right for the 6.1% section of the profile. Once we reach the shallower section, after D5.0, we can slow up to a more leisuredly 5-600 feet per minute. The best tip I can give you here is to ensure that the aircraft is in landing configuration before you reach the D10.0 KTM. If you are still getting dirty as you pass the FAF, then the aircraft will tend to balloon and you will need to set descent rates of over 2000 feet per minute to get down.

But the cloud continues to close in, and at D6.0 you realise that it is unlikely you will be visual at the MAP. At D5.0 KTM, you alter the descent rate to 700 feet per minute, and set 5800 feet in the altitude window without much confidence. Note that you may go briefly visual somewhere on the approach, but don't rely on it.

Screenshot 7 shows the MAP. Note that I have elected to fly the approach in Nav mode on the MCP and have committed the cardinal error of not backing up the aircraft course with the heading bug. The aircraft is slightly high at 5200 feet, one nautical mile from the Kathmandu VOR and with nothing on the clock but the maker's name, there isn't anything but cloud up front. Given that there is only 900 feet between you and the threshold, it is time to clean up and get out - this is where you can make a horrible mistake if you didn't set the heading to back up the course in Nav mode, because when you hit the heading key so that you can use it to alter course, the 737 will fly off on the last heading you set, which happens to be 054 - straight into a hill. First priority is to increase airspeed to 180 kias - then, as the 737 begins to accelerate, raise the gear. Once you are up to speed, then begin a climb on 022, passing overhead KTM to D2.0 KTM on the far side of the airport. Pick up the flaps in stages to five degrees after you reach 180 kias and reset the course window to 289, which is the outbound course to the hold.

Now we have to fly the DME arc. Begin a fifteen degree turn, watching the DME as it begins to increase, and adjusting your rate of turn when it hits DME 4.0. Needless to say, it isn't easy to get this right, and in real life, you would fly this one using the procedure programmed in the FMC, but if you don't happen to have one of those, there is nothing like trying the hand-rolled version. Assuming you can keep the bank rate constant, and you get the rhythm of your mouse clicking right, you shouldn't have too many problems getting round. The alternative is, of course, to leave the aircraft on autothrottle and fly it manually, but whatever you do, don't get into too steep a turn, as you aren't that far off your stalling speed.

You can keep climbing in the turn, but I would restrict your rate initially to around 1500-2000 fpm, just to make sure you don't run out of airspeed. Although a textbook missed approach would have you fly the entire arc at 4.0 miles, the key point is to ensure that you don't stray outside it, because ground rises to above the mimimuns all around the east side of the airport and some of it is uncomfortably close to the arc itself.

Now you have to intercept the 289 radial. This isn't as easy as it sounds, because the intercept occurs very close to the VOR, and the CDI will begin to wander almost the instant you straighten up and want to use it, so just establish yourself on a heading of 289 and reset the ADF to 258, which is the frequency for LDK, the NDB on which the hold is based.

Maintain a heading of 289 until you are far enough from the VOR to know which side of track you are, and watch the ADF needle, which will also give you a clue which side of track you are on. While you are outbound, you can increase your rate of climb and when you reach 10500, increase speed to 220 knots and clean up some more flap.

The most difficult part of this procedure is undoubtedly the DME arc, which presents a high workload even in Flight Simulator. Just imagine what it is like doing it for real... but the advantage of the simulation is that you can do the procedure again and again, until you get it right. If you can fly this one perfectly, there isn't much going to faze you. Just for kicks, I have included a situation for the VOR 02 approach where the weather isn't socked right in, and you can fly it by selecting situation Kathmandu2000vor02 for FS2000 or Kathmandu2002vor02 for FS2002.

Situation 3. SID rwy 02.

FS2000 - Select situation Kathmandu2000SID02.

FS2002 use Kathmandu2002sid02

Now you've got all that practice flying DME arcs, it seems a shame to waste it, so we are going to fly the DHARKE 1E SID out of Kathmandu. This the most complex SID I have ever seen constructed off a single navaid and once you have figured the procedure out, it will stick in your memory for a long time, as an elegant solution to a difficult problem. No doubt connoisseurs of SIDs speak of this one in hushed tones - "Yep, but just think of that one Ragu did for VNKT, Jed. Come up with a better one than that!" In fact, if anyone can come up with a better one than this, I will feature it in a special article in the series.

We had better get on before I break into song. The reason why DHARKE 1E is done the way it is will be obvious the moment you finish loading the situation and take a look out the window. You can climb a fully loaded 737 straight out over that hill up front, but if an engine failed on the way they would be posting your remains back to mom in an envelope, so a circling departure is pretty much the only way out.

The plate looks completely incomprehensible until you realise that the intention is for the aircraft to carry out two ascending circuits of the airport, using the same left hand 4 mile DME arc based on the KTM VOR. After takeoff, the SID begins with a turn to port at D3.0 KTM, then you fly right around the arc to pass over KTM, but instead of following the runway direction, you take the 038 radial from KTM to D2.2 and then fly another DME arc (higher than the first), before intercepting the 310 radial from KTM, where you turn to starboard onto the 289 radial from KTM. This takes you to the LDK NDB, which must be crossed at not less than 10500 feet. The key to flying this rather neat procedure is to tune both Nav radios to KTM and to use the RMI to fly the arc and the HSI to intercept the radials. Committed the procedure to memory? OK, let's go.

The weather I have chosen for the flight isn't too bad, chiefly because it allows you to appreciate just how dangerous the departure would be if the SID wasn't followed exactly.

Use flaps 15 and retract the gear once you have a positive rate of climb. If you take off with the autothrottles armed, all you have to do is flick in the IAS button and you can forget about the power setting - there is plenty else to do, because I do not recommend flying this one on autopilot until you have completed both the DME arcs.

Climb until the DME reads about 2.9 and then make a seriously steep bank to port - at least 30 degrees will be necessary to begin with, because otherwise you will trespass outside the arc, which is a no-no. The danger here is that flying a fully loaded plane close to its stall speed in a climbing turn is not an insurable risk, so the best policy is to climb out as steeply as you can to DME 3.0, then flatten off some as you begin the turn.

As soon as you intercept the arc, begin to roll the aircraft level, or you will turn inside the arc and find yourself flying directly towards the VOR with no hope of intercepting the outbound radial on the second pass - you can tell if this is happening because the DME will rapidly fall. Screenshot 10 shows how steep the initial bank angle must be, though as I say, it would be sensible not to climb the aircraft as steeply as I did in the turn. Once you are flying the arc, keep an eye on the EHSI so that you don't miss the CDI beginning to move as you near the 310 radial - this marks the point at which you complete the first turn around the arc and tighten the turn to intercept the 038 radial outbound from KTM.

The 038 intercept throws up its own set of challenges. The first is that the only way you have of knowing you are passing over the VOR is bymonitoring the DME or picking a landmark and watching it (easier said than done in FS2002, even if you are VFR) - and then you have to align yourself with the 038 radial in the very short space of time available between the CDI stabilising after passing over the VOR and the aircraft reaching DME 2.2, at which point you have to pull yet another turn to intercept the 4.0 nm arc again.

Bearing all this in mind, the secret of flying the second arc is not to make such a steep bank as you did going round the first time. 2.2 DME on the 038 radial puts you that little bit further out and means that 25 degrees or so of bank should be enough to pull the plane round onto the arc - once again, as you begin to parallel the runway, take care to roll the aircraft a little more upright, or you will get too far inside the arc once again, although this time it isn't so crucial that you get it right, because you don't have to pass over KTM.

Screenshot 12 shows the 737 at the 310 radial, having completed the second arc. At this point you can set the course to 289 and intercept that radial outbound to the Dharke NDB at D13.5 KTM. Check out your altitude at this point, because you should be at a minimum of 10500 feet when you cross over the NDB.

The 289 intercept is probably a good moment to quit flying the plane by hand and engage the autopilot, but one final point is that it is worth using heading mode to alter course just slightly to starboard, so that the aircraft intercepts the radial at a 30-45 degree angle, rather than 90 degrees, as it naturally would if you just switched to NAV mode when you crossed the 310 radial. Sure, the autopilot will intercept the radial at right angles, but the plane has a tendency to do a messy series of lazy esses while it locks onto the centerline and in some cases it might fly straight through just to spite you.

And that is Kathmandu - a uniquely challenging set of procedures, I think. I learned a lot from flying it and I hope that you did too.

Andrew Herd
andrew@flightsim.com