ATCC Controllers' Read Binder...

NOTAMS, FAQs and other info for users of ATCC

October, 1997

#1: New ZLA Maps

The real-life sectors 38 and 19 have changed their radar maps to show solid lines for the sector boundaries, instead of dotted lines. This was done to save memory in their aging, 16MB RAM mainframe. If you are interested in continued realism, click here for the new ZLA radar maps (MAPS.ZIP), save it to your \ATCC directory, then use an unzip utility to extract them. Why each center seems to have their own map format, nobody knows, but this is the latest.

Along the lines of new add-ons, Dragon Systems (www.dragonsystems.com) has very shortly (or has already) released their continuous-speech voice recognition package, selling in the U.S. $150-200 range. After our talking pilots upgrade is completed ("coming soon"), we will get to work on a driver for the voice recognition, followed by new, "official" sectors.

#2 Deals Everywhere!

If you have been certified at one or more sectors in ATCC, you are a "controller" now, so you might as well know the truth: real-life losses of separation ("deals") happen all the time, and nobody knows how to stop them! (Numbers are on the order of 500+ per year in the U.S.). Full-time trend analysts produce comically random graphs that ultimately show almost zero correlation between the occurence of deals and time of day, season, time on position, complexity of traffic, or any other factor that might contribute to the problem.

Various changes in practices and procedures have been attempted, to try to reduce the frequency of these "operational errors." One cyclical stop-gap measure has been to have a D-side at the sector at all times, with the reasoning that an extra pair of eyes and ears can only help. As it turns out, it also means having somebody to talk to, share jokes with and get distracted. Also, some controllers tend to be more bold and daring when somebody else is watching closely, so there are both advantages and disadvantages to permanent D-sides.

"Miscommunication" is likely the bottom-line cause of the vast majority of "dumb" errors. You tell an aircraft to turn "left," when you really meant to say "right." Or you say "flight level two two zero" while you think you are saying "flight level two zero zero." Or maybe you see aircraft A and B are in conflict, plan to descend aircraft A, but look at B's callsign in the datablock as you issue the descent.

Certainly fatigue and lack of alertness (see August issue, #6 ) can play a large role, but even the most careful, precise, ultra-alert controller will make the occasional mistake. What can be done? Humans make mistakes. Fortunately, radar separation is just one of several safeguards against collisions, which include cockpit radar (TCAS), pilots' eyes, and the big-sky theory. It is unlikely all will fail simultaneously, and in fact there have been no mid-air collisions between two Center-controlled IFR aircraft in the U.S. (Collisions with VFR aircraft have occured, though--those little I's and V's are very real!)

Good scanning techniques and a constant re-evaluation of what is happening in the sector can help catch "dumb mistakes" before they become a problem. Good short-term memory also helps. For example, you may see in your scan that you have an aircraft climbing to an interim (temporary) altitude of FL330 for crossing traffic at FL350. You may remember that was your plan, to stop the aircraft underneath the other until they pass, then continue his climb. But if you cannot specifically remember issuing FL330, and cannot specifically remember the pilot reading it back, re-issue it ("just to verify...maintain flight level 330."). The pilot may sound annoyed, "uhhh, yea, that's what you said the first time, FL330." But why risk it? It may be one of the one-in-ten-thousand-times you meant to say FL330, and put FL330 in the datablock, but actually said FL350.

On the other hand, at some point you have to trust your memory, that the pilot did indeed read back FL330, otherwise you will go crazy trying to verify everybody's altitude again. But, if you are ever unsure, re-issue the altitude "just to be safe."

#3: Deals in ATCC

For comparison, a real-life controller trains and works for 500-1000 hours on the six sectors before becoming an FPL, and that is after spending a similar amount of time doing just D-sides (which mostly just let you watch how other controllers work). If you spend 300-500 hours with ATCC either working traffic or training at the 80-100 level (with master level set at 100), you should get a similar level of experience as a real-life (new) controller. Of course, you will not get one-on-one instruction ("On-the-job training"), but in most cases the real-life instruction just consists of a bored controller who is told to watch you, and who yells at you if you make a mistake, and writes "no discrepencies noted" if you don't. Controllers aren't teachers, and teachers aren't controllers -- you sink or you swim, and are mostly on your own. Out of the 500 or so hours this author spent training at radar positions (after about 500 on the D-sides), he received only three moments of practical advice: "Keep the scan moving," "Get rid of aircraft as soon as you can," and "The sooner you start something, the less you have to do." The rest was just figured out as time went on, and new situations were experienced.

After almost 7 years of real-life controlling, and about 120 hours of "working" ATCC sectors, this author has an 85 rating and one deal in ATCC, ironically at one of the sectors he worked in real-life, ZLA 38 ("I just didn't see it!"). Unless you are a real Center controller, you may certainly have more deals than this, but if you ever make it to the 300-500 hour level (including time spent training), you will be a true "professional" and should have no more, ever. (Though see #2 above).

#4: Article 65

The controller is responsible for all activity in the sector, and if there is a deal, he will take the blame. With their jobs ultimately at stake, very few controllers will tolerate somebody else telling them what to do with their traffic, especially when it is busy and the controller has formulated a shaky plan he is struggling to make happen. If you were working a very busy sector 97, for example, and received a call from sector 73 telling you to "Hold UAL243 at ETX" because they can't take him right now, the proper, expected response would probably be an obscenity. Nobody can tell you how to run your sector! Forget sector 73. Hand him off to sector 93. 93 can't take him? Try 98. No? Hold him at PTW. Or FJC. Or anywhere you want! 73 can't tell you how to run your traffic. It's your sector, your responsibility, and you run your own traffic!

Nobody can tell you what to do with your traffic, that is, except a supervisor. A supervisor may override you at any time, and whatever he says to do, you must do. If a supervisor tells you to "descend this guy to FL230," you must descend him to FL230, even if that causes a deal. If the supervisor says to slow UAL243 to 250 knots, you must slow UAL243 to 250 knots, even if there is a DC-10 immediately behind him, coming up his tailpipe! If a supervisor ever tells you to issue a particular clearance, and you don't personally agree with it, you must issue it but invoke "article 65" (of the union contract), which states the supervisor now assumes all responsibility for the operation of the sector. Even if the particular command does not cause a deal, it may disrupt the general flow of the sector, and was not your plan originally, so you cannot be held responsible for future deals that may occur.

Naturally, supervisors can't (and don't) know what is going on with every aircraft in every sector, so it is extremely rare for them to order you to issue a particular command. It has happened, though... in a recent deal at L.A. Center, a supervisor looked over the shoulder at a situation he thought would not work, and told the controller to turn one aircraft 40 degrees to the right, which the controller promptly issued--which then put the aircraft head-on with another the supervisor didn't know about, causing a deal. The deal was blamed entirely on the supervisor.

That can be a very difficult situation for a controller--follow the orders of the supervisor, and possibly have a head-on collision, or disobey the supervisor, and be fired. Too bad! Take some more antacid tablets. Also, your request for vacation has been denied.

Fortunately, most supervisors know better, and leave controllers alone. They learn who they can trust in the busy sectors, who will need a D-side, and who needs extra watching. They'll bite their lip as they watch controllers run aircraft 5.1 miles apart, and if there is a deal, there is a deal. What can they do?

#5: Track Jumps

You are happily moving traffic in your sector, everybody's separated, and you have a J-ring around one aircraft just to make sure. The other aircraft is 6 miles away to the side, on a parallel course, and has been that way for the past minute, so you go ahead and climb him. The blips move -- and suddenly he jumps 3 miles sideways. Instant deal! Then the blips move again, and they're back to 6 miles apart. Next update, 6 miles again. You are taken off the sector.

This is the infamous "track jump," where a bad radar hit (from the computer switching to a different radar site) causes the blip to move up to 3 miles from where it actually should be. It can cause the snitch to start blinking, or can just cause general confusion. If you do have a deal because of a track jump, though, it is usually written off as a "bad radar hit," and discarded.

Usually, track jumps happen in particular locations in certain sectors, so after awhile you learn where to maybe use 8 miles instead of 5 miles, just to avoid the trouble of being yanked off of position and having to wait while the radar data is collected to get out of your "deal." In L.A. sector 4, for example, there is a huge track-jump area from GMN to about 5 miles north, so you may have a steady 8 miles apart between departures climbing out, then they both hit the track jump spot, one jumps 3 miles forward and the other 3 back, and now you have 2 miles, same altitude. A little scary, sometimes. Then they jump back into position, 8 miles again.

Fortunately, when a target jumps to the side suddenly, the radar often misses the altitude ("NONE"), so the snitch never goes off. But it is part of the radar you have to get used to. (ATCC doesn't simulate track jumps, but there is the occasional lost altitude ("NONE") displayed).

#6 Histories

You should consider the "HIST" control on your radar to be a very useful tool. The "VECT" control will extend a projected route of flight, of course, but it is just a computer estimate, and we all know computers make mistakes. Much better for determining if two aircraft will eventually come together is to move up the HIST display to 5, which will show you the REAL directions of flight.

Histories will also show if an aircraft is in a turn. It is rare, but occasionally a pilot will mis-program his automatic navigation ("LNAV") with a wrong fix name, or the wrong departure route, and when he flips on the autopilot, the aircraft suddenly makes a 90 degree turn toward the wrong point. Because the VECT display is a mathematical average of previous histories, a sharp turn won't show up initially, and after 30 seconds it would appear only as a slight turn (because of averaging), and wouldn't be obvious for probably another 30 seconds.

Looking at the history displays, though, you can know almost instantly. If you get one radar hit off-course, it may be a track jump (see #5), but if you know it's not in a track-jump area you can be pretty sure the aircraft is turning. A second radar hit (total of 12 seconds later) will confirm it, or after the first off-course hit you can ask the pilot to verify his heading.

You should incorporate the histories and the current position in your scan of each aircraft, which means you should have at least two histories displayed, or better three. One history is the "legal" minimum, and does make things less cluttered, but often it is harder to tell if an aircraft has taken a wrong turn until it is too late.

#7: Wrong Frequency

This topic has been covered before, but is common enough to mention again. Occasionally aircraft will check on to your frequency that don't belong, or which you had previously handed off. This is somewhat common in reality, as so many sector frequencies are just one digit or transposed number away. You should ask their position with SP, and pull up their datablock with <F7>callsign, as in <F7>AAL1025. Then click up the RANGE until the datablock is visible. Tell them to switch to whichever sector is nearest to their position. If they come back again, requesting lower, just give them lower (like 3000) to keep them quiet, then ignore them. In reality they would return to their old frequency, but due to a minor glitch they sometimes lose their old frequency and come to you for help. But, of course, you can't always help them. Ignore them, and eventually they will go away.

#8: Airline Differences

Different aircraft have different abilities, of course; different pilots have different moods (which you can sense in their choice of words and how they say it), and likewise different airlines have different policies and procedures which all affect how the aircraft will move or perform in a given situation.

We can't name names, but one regional airline's pilots are always cheerful, and when you tell them to expedite, they'll give you 3,000 feet per minute in their 737's (maximum rate). Another airline you may tell to expedite, which they'll acknowledge, but then you'll see absolutely no change at all. One airline you can tell to "maintain maximum forward speed," and they'll push their 737's to the absolute max, 325 knots. Another you'll tell "maximum forward speed," and you ask them a little later what they're doing, and they'll say "280," in an MD-80 which can easily do 330. If you assign 330, they'll do 330, but if you say "maximum forward speed" they'll ignore it.

You just learn which are the "good" airlines and which are "average" (there are no bad airlines in the U.S., really). With a "good" airline you might decide to ask him to expedite his climb to get above another aircraft, but with the "average" airline you'd just stop underneath until they pass. If you have two aircraft in an even tie for the airport, you'd probably make the "good" airline #1, all else being equal, knowing you can ask them to keep going as fast as possible, and they'll do just that. The "good" airlines will tell you they've hit some turbulence and need to slow down, for example, while the "average" airlines will just slow down on their own and may not tell you until you have a 100 knot overtake.

You cannot give preference to one airline over another, but when performance and expectations are a factor (as they often are), you can't help but take the differences into account. Likewise, if it is a strange airline (which you know probably isn't familiar with the area), with pilots who barely speak English, keep the instructions simple, speak a little more slowly, and be sure to clarify any unusual instructions or routings. Most controllers would feel comfortable with 6 miles parallel spacing between two UAL 757's flying into ORD, for example, but would probably leave some extra room if one was a foreign airline they'd never seen before. You can never be too safe!

END

The Read Binder is updated at the beginning of the month. All information is for use with Xavius Software's Air Traffic Control Center^TM only, is the opinion of the author(s), and does not necessarily reflect the policies or practices of the U.S. Federal Aviation Administration or Federal Aviation Service. Send your questions or comments to support@xavius.com and we'll be glad to help!