>P.S. Off topic folks, but if you think the FMC is complex, >try juggling the fuel on this aircraft... 9 fuel tanks! On a normal flight; turning off the center tank pumps (and stab pumps if it's a long flight), and later on closing the crossfeed valves and turning off the override pumps ("fuel tank to engine" configuration), doesn't seem that complicated. ;-)And you don't even have to do anything until you see the EICAS messages! :DMartin767 fetishistIt's a lot like life and that's what's appealing

>On a normal flight; turning off the center tank pumps (and >stab pumps if it's a long flight), and later on >closing the crossfeed valves and turning off the override >pumps ("fuel tank to engine" configuration), doesn't seem >that complicated. ;-) And what do I do with the "AUX" tank? ;-)(Actually, I meant from a designers point of view, not the pilots' ;-). The new tank has some unusually complex logic for vents/valves and pump.>And you don't even have to do anything until you see the >EICAS messages! :DTrue. The same should apply to the new aircraft. Unfortunately, there are a few teething problems ;-)Cheers.Ian.

I guess shes a wonderful and a beautiful brand new B747-400, maybe ER? the shape of the screens are just nice and sharp~:-sun1 http://www.airliners.net/open.file/298578/L/

Hmm okay, here's my point of view, speaking of experience in the 757 with the B744-type autopilot, but not the Pegasus system. It is, however more modern than the Pegasus, as I understand it.TOD is calculated backwards from EOD (end of descent, which is either the runway, or the lowest altitude restriction in the flight plan (this can be the intercept altitude for a glide slope).Initial TOD will give an idle descent to the nearest altitude restriction. The FMC cannot give a "fixed" V/S value to the auto flight system since IAS will increase first to maintain Mach number (TAS and GS increase) and after that IAS will remain constant while TAS and GS decrease. So it's a vertical path that is calculated with projected groundspeed and the v/s at that thrust setting with the specified Mach number or indicated airspeed. Vertical speed is still dependent on IAS and not on ground speed, since it's feet per minute and not feet per distance.The FMC will calculate deceleration phases using its performance database. It will do so for ANY deceleration. So if you put 10 waypoints in your descent with 10 different airspeeds, the FMC will calculate all those decelerations. Based on gross weight, a certain deceleration rate will be adhered to, from experience, rougly 1 knot per second. If you're heavy, this gives you a higher v/s than if you're light. The path will flatten until the FMC projects you will have the desired airspeed, from which point the nose will drop and the path will become steeper to hold that airspeed.As an example, see the picture I made below. The FMC starts all the way on the right at EOD. The BIG numbers on the CDU window are mandatory, the smaller ones are calculated by the FMC. Same with the altitude/speed readouts on the bottom of the graph, the numbers in parentheses are not fixed, it's just what they happen to be at. The FMC will start calculating at EOD and then calculates a deceleration path that happens to come over OA at 156 knots at that 2500 feet restriction. Then it draws a power on path at 250 knots from ARTIP. Before ARTIP, everything is idle. It first calculates a deceleration path from 312 to 250 knots, then the rest is the calculated idle descent path. The whole approach is done in VNAV PTH without use of speed intervention, but with manual flap setting of course.If you're too high in the first part of the descent, use speed intervention (340 knots) to steepen the descent and bring the nose down when it accelerates (6000 fpm+). Cancel the speed intv when you're slightly below your path again so the aircraft can decelerate and go back to VNAV PTH, otherwise it will stay in VNAV SPD until you tell it otherwise, right through the path!Any questions? :)Iz

Iz,Very nice dePICtion of the decent in VNAV. I often get busy logging fuel or chatting with friends over the text window in VATSIM and find myself sliding a little past TOD before getting the MCP altitude reset and reselecting VNAV. PIC is very forgiving since it does not compute TOD at idle thrust and will pretty well bring you back down to the descent path. While I occasionally use (increased) speed intervention in level flight, it had not occured to me *:-* that it would be a very handy tool to steepen the descent and help get back on track. This would be a nice alternative to setting VS. As you say, you would still need to monitor when you had intercepted the path to close up the speed window but I like this method much better. Thanks!Do you guys really go to/or past -6000 fpm? I do it in the sim sometimes but thought it would be completely out of order in real world.

Great pic, Iz. I've often wondered about this fixed mach segment. I've also wondered how the aircraft slows down between waypoints: Does it slow down gradually (fixed deceleration and thrust) between the waypoints... or, when it can, slows down just before the next waypoint (if it has a lower speed restriction). Previously I thought that power might be applied constantly between ARTIP and OA. Your diagram shows otherwise. Presumably one method is more efficient than the other (or too hard to compute!).Of course, different FMC software may change all this(?) Thanks, Cheers.Ian.

Ian, yeah that's the key, efficiency, obviously. That's what VNAV is for, instead of fixed descent angles. Regarding speed changes, the aircraft will go as fast as possible at any time. More correctly stated, it will fly as close to the most efficient speed as possible. This is determined with the CI (cost index), unless you override the speeds by manually selecting them in the FMC VNAV pages. So when the aircraft accelerates, it will do so as soon as possible, right after passing the lower speed restricted waypoint. When decelerating, it will do so as late as possible before the restricting waypoint.It's quite clear that the FMC will use fixed acceleration/deceleration, as I said, about 1 knot per second. During climbout through FL100, it will use this acceleration rate with the CLB (or CLB 1 or CLB 2) thrust setting, resulting in 1000 fpm climb when you're heavy and 2500 fpm when you're light.During normal descent, it's always that the higher and faster (CI) you are, the better. So in my example from ARTIP @ FL100 to OA @ 2500', it's better to keep the descent path relatively flat and at higher altitude, then decelerate at idle thrust, than making a steeper descent at idle or lower thrust and then decelerate with thrust applied. Those are the only two options to do it and the latter is less efficient.Mike, yes it's not that strange to momentarily exceed 6000 fpm in the descent as the aircraft accelerates.I never override the CI ECON speed of around .80/312 and if for some reason I have to descend faster, I'll punch the speed selector and dial it up to Mmo/Vmo (the aircraft will automatically keep about a 5 knot margin between the barberpole and its commanded speed).The biggest altitude loss is actually achieved by the plane accelerating to 345 knots. The nose will drop a few degrees, you'll actually feel the "gravity" decrease to less than 1g (some cabin attendants have noted that they almost went airborne but that's a little exaggerated) and vertical speed will increase dramatically. It's all done in a smooth motion though. This is exactly the opposite of what happens when you fly at a high speed and pull the nose up sharply. Vertical speed will spike initially as the flight path vector is pointed upward and lift is increased dramatically but will settle down as kinetic energy (speed) is converted into potential energy (altitude).Back to the descent, after you reach the new higher speed, the new descent angle will also be steeper but vertical speed will not be that dramatic anymore.Iz

Iz,It is great to see another point of view. It helps toward my further understanding of the 767 FMS system logic.Regards,Cliff