Hi all,
I have a question regarding the system Logic of the Tank-to-Engine condition.
As we all know, when the amount of fuel in tanks 2 and or 3 is equal or less than the fuel in tanks 1 and 4, the related EICAS message appears to remind the crew to re-configure the tanks.
But what happens when a pilot is faster (or smarter ;-)) than the system logic? I tried it out with PSX and monitored the amount of fuel in the tanks. As soon as the amount in tank 2 and 3 were equal to 1 and 4 (13.2 tons), I quickly configured the tank to engine config and the EICAS message never appeared. This is what I expected.
Interestingly, I couldn't reproduce this in the PMDG product. Instead, I got a FUEL X-FEED message, although the amount of fuel was equal in all tanks. PMDG claims that the fuel indicated on the synoptic page is not accurate and that the O/J pumps have been switched off too early. They also state that the system logic requires pilots to wait for the FUEL TANK/ENG message to show up before they actually turn off the O/J pumps and crossfeed valves.
I couldn't find detailed information about this in my documents.
In reality, pilots will probably never spend unnecessary time to cheat with the system like I did ;.) - but what is the actual logic behind it? Some tolerance with the amounts (like it obviously is in PSX) or a strict equal tank condition?
Best regards,
Dirk
IIRC there is a bracket where no messages will be generated. It is to prevent spurious alerts.
e.g. Below 10T you get "TANK TO ENG", and above 12T you get "OVRD PUMP X" and "FUEL XFEED"
So if you switch when between these two levels, you will not get an alert.
However, in PSX, the logic seems to work that if it is NOT "TANK TO ENG" then it will alert after a few seconds for the other option.
I have raised this before, as sometimes PSX would request the OVRD PUMPS ON again even when the aircraft should be TANK TO ENG.
Note this was a while ago, so I'm not sure if this has been changed recently.
Hi Dirk,
in PSX (and certainly in the real system) there is a hysterisis of some kilograms to avoid nuisance messages. Remember, the message logic is sharp and the physical fuel volume-and-density is a big dynamic object; when the sensed volumes and sensed densities are dancing around a trigger value by just the width of a molecule you would get blinking nuisance messages.
Regards,
|-|ardy
In the Boeing Wiring Schematics, there is mention of a 120 second time delay for the FUEL TANK/ENG after the fuel has reached an equal quantity. Add this to pilot reaction time and I can see how the fuel quantities in the tanks will be well and truly balanced (i.e. minor fuel quantity measurement variations will be resolved). I guess it's more "hit and miss" if you reconfigure prior to the message appearing.
This 120 sec delay is the time required to burn approx. 300 kg in total. It makes sense as it works like a quantity based hysterisis. The gap of 120 sec on the time line is also a gap of 300 kg on the quantity line. Divided by the number of tanks, each tank will refer to a fraction of 300.
In typical cruise flight, if the message is delayed by 120 seconds, the message will appear when the inboard main tanks each contain 200 kg less than their respective onside outboard main tank.
So readers of the manuals shouldn't be confused if the text reads "equal or less". It's always way less than "equal". Or the message timer is started when the inboard tanks are 200 kg above, and at the timeout the tanks are equal.
QuoteIn typical cruise flight, if the message is delayed by 120 seconds, the message will appear when the inboard main tanks each contain 200 kg less than their respective onside outboard main tank.
That was my thought, too, and this prevented the flip/flop between messages. Pilot input required? When I checked fuel levels on the ground after a flight, I couldn't really say if this was so because of (normal) uneven engine fuel burn and APU usage.
For engineers... the logic is shown on SSM 28-41-02 Page 101 sheet 3 (although for some strange reason, the alert is called a memo message).
I just noticed in PSX that I'm getting the >FUEL TANK/ENG message with an inboard tank consistently above the adjacent outboard tank (by 0.1 tonnes). The F/O, if activated, also responds prematurely.
I can't remember if we discussed this during beta testing, but the logic is:
If the fuel in the inboard and outboard tank is equal (or less than) on any wing, the message will appear. You can reset the message by reconfiguring the fuel panel or by:
(during refuelling) increasing the level of the fuel in the inboard tank by 2000lbs more than the outboard... but it has to be greater on both wings.
This 0.1 on the EICAS is due to display rounding. The true sensed (and lagging) difference in the background is not higher than exactly zero.
Anyway, in the next update that 120 sec delay will rigorously take effect, and thus, in flight, the message will only appear when the difference exceeds ca. 200 kg.
QuoteIf the fuel in the inboard and outboard tank is equal (or less than) on any wing, ...
Is this pair-wise just left wing and just right wing? Or does "any wing" mean "anywhere cross-wise", e.g. M2 <= M4?
In PSX since version 10.0.0, tank-to-engine begins if ...
if M2 <= M1
or M2 <= M4
or M3 <= M1
or M3 <= M4
The first inboard tank to go equal with its adjacent (same wing outboard tank) tank will trigger the 120 second timer.
[(M2<=M1) or (M3<=M4)] plus 120seconds = message.
The reset is a bit more complicated than I thought and involves M2 and M3 reaching the VTO (max volume) of the inboard tanks plus 2000lbs
FUEL TANK/ENG Logic (http://www.members.iinet.net.au/~b744er@ozemail.com.au/PSX/FUELTANKENGlogic.gif)
Thanks for the replies.
But I think it's safe to assume that if the pilot doesn't necessarily has to wait for the TANK/ENG message to appear, correct?
If the pilot selects the tank-to-engine configuration at the appropriate time, he'll neither get a crossfeed advisory nor a TANK/ENG message.
After seeing so many discussions about alert inhibits, timers and other system logic stuff I really start to admire the engineers who implemented all this stuff roughly 30 years ago.
QuoteIf the pilot selects the tank-to-engine configuration at the appropriate time, he'll neither get a crossfeed advisory nor a TANK/ENG message.
In a perfect world, yes. However, the fuel quantity system is not entirely perfect. Although there are multiple sensors in each tank at different points in the tank, if the aircraft attitude/speed changes quickly or the aircraft hits turbulence, the fuel quantity processor may register a false level during this X-FEED CONFIG ~ FUEL TANK/ENG transition.
There is also a delay on the X-FEED CONFIG message, so it may also help reduce message flip-flop.
On the Set/Reset AND-gate, shouldn't there be a feedback loop from the gate output back to the "SET" input, so that the SET input remains set when M2>M1 and M3>M4?
M2>M1 and M3>M4 must not reset the SET state. Only the VTO stuff may reset it.
Or is it sufficient if the symbology includes the word "SET"? Is this a standard word indicating that this input is self-holding?
And to add engineering hell, all wiring associated with the FQIS needs to be treated as if it was the fuse wire to the bomb. When installing SATCOM, for example, you need to keep your wires (audio, 429, power) 2 inch or so away from any FQIS bundle to avoid sparking between wires and potential induction of high voltages to the fuel quantity sensors and setting off an explosion.
Modern systems use fibre optics for the FQIS, and only for the FQIS... talk about precautions.
Hoppie
QuoteOr is it sufficient if the symbology includes the word "SET"? Is this a standard word indicating that this input is self-holding?
Most likely.
That VTO logic is interesting. I always assumed that you would get the TANK/ENG message if you loaded the tanks:
M1 10.0, M2 11.0, M3 11.0, M4 10.0...
and configured the aircraft to burn fuel from the inboards.
Of course, you would have to manually control the refuelling valves to get this configuration prior to flight. Normally, however, you would have the refuelling system carry out everything automatically and so you wouldn't get that configuration.
(EDIT: A PPRuNer has finally given us the answer we were looking for. He says it's normal for the inboards to be 1~200kg below the outboards when the message appears)
Ugh... more complications...
Just reported on PPRuNe...
In someone's QRH and I now see it in ours...
>FUEL TANK/ENG
Condition: One of these occurs with crossfeed valve 1 or 4 open:
*Main tank 2 quantity is equal to or less than main tank 1 quantity, or main tank 3 quantity is equal to or less than main tank 4 quantity
*On the ground after refueling, after initial electrical power established, or after CMC ground test; main tank 2 quantity less than or equal to main tank 1 quantity plus 500 kgs and main tank 3 quantity less than or equal to main tank 4 quantity plus 500 kgs
I didn't think of looking in the QRH for such detailed information.
So in these cases the reset occurs at plus 1100 LBS (500 kg) instead of plus 2000 LBS?
Perhaps these two different values are not really different; maybe they refer to the same hysterisis function as shown in that logic gates diagram. Taking all that sensor lag and sensor tolerances into account, those 450 kg (900 lbs) are just peanuts, aren't they? Subtract a few pounds from the 2000 value and add a few to the 1100, and consider the rounded indications on the EICAS, there won't be any dramatic difference. Aren't these values, 2000 and 1100, just approximations anyway?
My first thought was that the reset logic was unrelated. I thought this new logic might apply to the situation where the inboards were a few hundred kilos (but less than 500) above the outboards with the outboard tanks not necessarily full. A sort of a preflight reminder. This tolerance probably allows for APU fuel usage (and any messages related to the inboards being below the outboards as a result of APU fuel usage).
e.g..... M1 10.0, M2 10.5, M3 M 10.5, M4 10.0 ... TANK/ENG message appears, but not at M1 10.0, M2 10.6, M3 10.6, M4 10.0
Is that logic diagram very old? Perhaps in a later version the memo message was turned into an advisory and the hysterisis of 2000 lbs changed to approx. 1100 lbs -- and they forgot to update the diagram?
Edit:
That diagram doesn't specify whether those 2000 lbs refer to inflight or to on-ground resets. -- The other text "On the ground after refueling, ..." etc. seems to refer to on-ground operations only. However, on a normal flight the tank quantities will not increase; so after all this other text too actually refers to all conditions, including inflight, I think. It probably mentions refueling etc. only because such actions can increase the quantities, and the actual purpose of this text is probably just to tell the values. What I'm trying to say is that both the diagram and the other text probably refer to the same hysterisis function, and that the hysterisis probably makes no difference between inflight and on-ground conditions (see diagram), and that this "VTO plus 2000 lbs" thing is, perhaps, now to be replaced by "inboard tank (not VTO) plus 500 kg".
Wild guess... on ground = pitch zero, which leads to a slightly different fuel gauge indication than in cruise flight with pitch a bit up? Enough to cause head scratching "why does the TANK/ENG indication (not) appear"?
Hoppie
Quote from: Jeroen Hoppenbrouwers on Sun, 22 Apr 2018 14:07
Wild guess... on ground = pitch zero, which leads to a slightly different fuel gauge indication than in cruise flight with pitch a bit up? Enough to cause head scratching "why does the TANK/ENG indication (not) appear"?
Possibly. But why does the diagram not distinguish between air and ground? And why does that other text not mention the effects of inflight operations? Both documents only mention one value, not two values for air and ground. I think there is one approximate value somewhere between 1100 lbs and 2000 lbs, or maybe around 1100 lbs after a later update (and not referring to VTO) -- because the variable pitch attitude cannot provide a constant, precise reference level. Not only at cruise attitudes. Tank-to-engine may also occur at climb and descent attitudes.
|-|
Attitude should be compensated for by the multiple level sensors in every tank. There are 11 in each inboard tank, 6 in each outboard tank. However...
The schematics are old and the QRH is new and we don't know what remains of the old logic. Perhaps someone can make sense of this version (even more complicated) in the Fault Isolation Manual:
A. Fault Isolation Procedure
(1) This EICAS message alerts the flight crew to configure the fuel system to establish a tank-toengine
fuel-feed configuration.
During flight, when the fuel quantity for the inboard main tanks becomes equal to the outboard
main tanks, the EICAS advisory message FUEL TANK/ENG will show (M1 = M2 or M3 = M4).
The flight crew action is to manually close crossfeed valves 1 and 4 and put the M2 and M3
override/jettison pumps to the off position. This will initiate a tank-to-engine fuel-feed
configuration until the end of the flight.
BAB ALL; AIRPLANES WITH FQPU BLOCK B UPGRADE (SB 28-2182)
(2) The >FUEL TANK/ENG message contains logic to prevent fuel-feed configuration messages
from showing on short flights during takeoff and initial climb. The logic allows the airplane to be
established in an early tank-to-engine configuration when the inboard main tanks are within
1000 lbs (455 kgs) of the outboard main tanks.
Additional logic will maintain the fuel-feed configuration established on the ground for 8 minutes
after the airplane transitions from ground to air. Specifically, for 8 minutes after the transition
from ground to air, the status of the tank-to-engine alert bit shall remain the same as its status
prior to the ground-to-air transition. This will prevent the >FUEL TANK/ENG and >X-FEED
CONFIG messages from showing during takeoff and initial climb.
BAB ALL
(3) The EICAS message >FUEL TANK/ENG (ADVISORY) shows when all of these conditions are
true:
(a) The No. 1 or 4 crossfeed valve(s) are not closed.
(b) All main tanks contain more than 2000 lbs (910 kgs) of fuel.
(c) Jettison mode is not active.
(d) The TANK TO ENGINE ALERT discrete is set (2-minute time delay).
(4) The TANK TO ENGINE ALERT discrete is set when these conditions are true:
(a) The airplane is not in refuel mode (the refuel control panel door is closed and/or the
airplane is in air mode).
(b) The M2 fuel quantity is less than 32,000 lbs (14,500 kgs) or the M3 fuel quantity is less than
32,000 lbs (14,500 kgs).
NOTE: M2 fuel qty is less than M1 VTO+2000 lbs (910 kgs) or M3 fuel qty is less than M4
VTO+2000 lbs (910 kgs).
(c) One of these conditions is true:
1) The M2 fuel quantity is equal to or less than M1 or the M3 fuel quantity is equal to or
less than M4.
BAB ALL; AIRPLANES WITH FQPU BLOCK B UPGRADE (SB 28-2182)
2) The M2 fuel quantity is not more than 1000 lbs (455 kgs) greater than the M1 fuel
quantity and the M3 fuel quantity is not more than 1000 lbs (455 kgs) greater than the M4
fuel quantity and one of these events occur:
a) The fuel quantity processor unit is powered up
b) The CMC ground test >FUEL QUANTITY IND is performed
c) The refuel panel door is cycled from open to closed and in ground mode.
zOMG now I opened the can of worms...
Easy. Except for the 28-2182 upgrade all these conditions are already implemented in PSX 10.32.
So the VTO+2000 stuff remains.
The 28-2182 upgrade just allows 1000 lbs more on top of the traditional "equal-at-any-level" condition for 8 minutes after lift-off. During these 8 minutes Dirk can set tank-to-engine even earlier, at 1000 lbs above equal level :-)
But the message will occur after 8 minutes. *
It seems when near tank-to-engine conditions, the tank-to-engine config is safer than letting the inboard quantities go lower by the O/J pumps, and getting X-FEED and O/J messages during initial climb.
|-|
* Edit: I was about to write that it will not occur as during these 8 minutes 1000 lbs will be burnt. But I just realized that the tanks cannot get equal when the x-feeds are closed and the outboard pumps are running.
Quoteb) The CMC ground test >FUEL QUANTITY IND is performed
Is this test manually performed before each flight, or even automatically? In other words, can we assume that it is performed before each flight
in the virtual world of any simulator?
|-|
Definitely not routine. Logic only applies if Avi's CMC add-on is activated ;)
The FQIS may go through some kind of self test on power up, but this will be different from a CMC test.
This makes me sad. I just implemented that 8 minute stuff :-)
So do they call maintenance if they know they will be within 1000 lbs and the 8 minute thing is to be activated?
Not sure I'm with you.
The reference to the CMC ground test is in an "OR" section. It would be normal for the other two ((a) & (c)) to occur prior to a flight.
Quote2) The M2 fuel quantity is not more than 1000 lbs (455 kgs) greater than the M1 fuel
quantity and the M3 fuel quantity is not more than 1000 lbs (455 kgs) greater than the M4
fuel quantity and one of these events occur:
a) The fuel quantity processor unit is powered up
b) The CMC ground test >FUEL QUANTITY IND is performed
c) The refuel panel door is cycled from open to closed and in ground mode.
EDIT: I'm not sure about (a). I don't know if they're talking about the FQPU/FQIS being powered up from a dead ship, the CB's being cycled, etc. I guess the key word is "up"... rather than just "powered".
I still don't know how to read the logic (e.g. That reference to VTO is confusing).
P.S. I guess if you were transitting an airport where there was no fuel (and had enough for the next flight), you might not get any of those three triggers. Even so, I doubt anyone would be aware of the ramifications. No one would ask maintenance to carry out a CMC test before the next flight. Having said that, a fueller might hook up the refuelling truck out of habit (before the engineers stopped him) and save the day :)
QuotezOMG now I opened the can of worms...
I think it was me who opened the can of worms ... perhaps you just gave the open can a kick ;)
But anyway, thanks for the insight to everyone! Learning something new every day.
Cheers,
Dirk
Quote from: John H Watson on Mon, 23 Apr 2018 12:41
The reference to the CMC ground test is in an "OR" section.
Ah, right. In the back of my head there was an "AND". All fine now :-)
Quote from: John H Watson on Mon, 23 Apr 2018 12:41
EDIT: I'm not sure about (a). I don't know if they're talking about the FQPU/FQIS being powered up from a dead ship, the CB's being cycled, etc. I guess the key word is "up"... rather than just "powered".
In my current model you can cycle D5 and D6 on P6 to repower the FQIS in flight at any time, and this way you can restart the 8 minute timer before its time-out.
8 minute stuff is now in PSX 10.33:
http://aerowinx.com/board/index.php?topic=4191.0
QuoteAdditional logic will maintain the fuel-feed configuration established on the ground for 8 minutes
after the airplane transitions from ground to air. Specifically, for 8 minutes after the transition
from ground to air, the status of the tank-to-engine alert bit shall remain the same as its status
prior to the ground-to-air transition.
I don't see how cycling the CBs will recreate a ground-air transition. Won't the FQPU just see Air straight away? Is the FQPU just seeing air/ground relays or something else?
Two of the three OR conditions (power-up, CMC, refueling) include no on-ground condition in the text. Any of these three conditions can set a certain data bit which can inhibit the X FEED CONFIG message. This bit will be reset when the aircraft has been in the air for 8 minutes. The timer starts when the aircraft is airborne.
CMC implies on-ground because CMC tests usually cannot be performed in flight. And refueling implies on-ground too (it's even explicitely mentioned). But the power-up does not imply on-ground conditions. So the power-up (CB cycling) can set the bit again, even in flight. And when the bit is set and the aircraft is airborne, the 8 minute timer restarts.
That's how I understand it. The effect is not dangerous. It's just a play with the system to illustrate the principle.
Quotea) The fuel quantity processor unit is powered up
b) The CMC ground test >FUEL QUANTITY IND is performed
c) The refuel panel door is cycled from open to closed and in ground mode.
We could also assume this text is based on another text or diagram and perhaps has lost some logical parantheses in the final copy; maybe it originally went like this:
(The fuel quantity processor unit is powered up
OR The CMC ground test >FUEL QUANTITY IND is performed
OR The refuel panel door is cycled from open to closed
)AND in ground mode
It's strange anyway that it mentions the ground mode in the refueling condition; has anybody ever refueled the 747 in flight?
Cheers,
|-|ardy
QuoteIt's strange anyway that it mentions the ground mode in the refueling condition; has anybody ever refueled the 747 in flight?
I've seen the refuelling panel under the wing left open due to miscommunication between engineers and refuellers (but noticed just prior to pushback). The panel is hinged on the forward edge. When fully open, the door is latched with a mechanically resistive device (or some kind of spring clip, I can't remember). Also at the fully open position, a magnetic switch is activated (which signals "fuelling in progress"). The door may be blown away from the fully open position (depending on the strength of the wind and the condition of the resistive device/clip). This should end the "fuelling in progress" logic. Perhaps, if the resistive device/clip was faulty (too tight), the fuel door may not move away from the fully open position until the aircraft lifted off. i.e. it would satisfy the open to close cycling logic in flight.
(Aside) Sometimes if the magnet part of the magnetic sensor falls off, we have to fit a temporary magnet to fool the magnetic switch during refuelling. If the magnet is left on, it can create all kinds of chaos in flight.
QuoteAny of these three conditions can set a certain data bit which can inhibit the X FEED CONFIG message.
Sorry, I'm still trying to wrap my head around this stuff. The new logic is designed to
prevent (inhibit) both messages (>FUEL TANK ENG and X FEED CONFIG) appearing during takeoff and initial climb, yet all the logic seems focussed on enabling the messages ;D
(2) The >FUEL TANK/ENG message contains logic to prevent fuel-feed configuration messages
from showing on short flights during takeoff and initial climb. The logic allows the airplane to be
established in an early tank-to-engine configuration when the inboard main tanks are within
1000 lbs (455 kgs) of the outboard main tanks.You obviously want to take off with no messages and have no new messages appear for the first 8 minutes after the ground to air transition (if at all). Does the new inhibit logic just apply to immediately after liftoff until 8 minutes has elapsed? (or is there a ground element?). e.g. if the tanks start out exactly equal (and the tanks are in a TANK to ENG configuration... and the engines are individually started 1~4, but #2 has start problems and takes a while to start (Meanwhile #1 is burning fuel and is now below #2). Do we now need to start talking about inhibiting the X-FEED CONFIG message or should that remain active? Maybe I should post the logic for that (from the FIM)?
Quote from: John H Watson on Tue, 24 Apr 2018 06:43
The new logic is designed to prevent (inhibit) both messages (>FUEL TANK ENG and X FEED CONFIG) appearing during takeoff and initial climb, ...
Both messages cannot appear at the same time anyway -- regardless of the upgrade.
There is a main-tanks-equal point on the timeline.
Before this point, only the X FEED CONFIG message can appear (if fuel feed is closed to early).
After this point, only the TANK TO ENG message can appear (if fuel feed is still open).
When we have 1000 lbs more on the inboard, we are before the point. Before the point, we only need to inhibit the X FEED CONFIG message. Before the point, the TANK TO ENG message will not appear anyway.
When we have 1000 lbs more on the inboard, and we close the fuel feed, we get the X FEED CONFIG message. We need to inhibit that. Just that.
Thanks... getting clearer now.
Just an excerpt from the FIM to remind me....
(2) The TANK TO ENGINE ALERT discrete is not set when any of these conditions are true:
(a) The airplane is in refuel mode (in ground mode and the refuel control panel door is open).
(b) The M2 fuel quantity is more than 32,000 lbs (14,500 kgs) or the M3 fuel quantity is more
than 32,000 lbs (14,500 kgs).
NOTE: M2 fuel qty is more than M1 VTO+2000 lbs (910 kgs) and M3 fuel qty is more than
M4 VTO+2000 lbs (910 kgs).
(c) The M2 fuel quantity is more than M1 and the M3 fuel quantity is more than M4.
(d) The M2 fuel quantity is more than 1000 lbs (455 kgs) greater than M1 and the M3 fuel
quantity is more than 1000 lbs (455 kgs) greater than M4.
I see now at (a), that the message can appear in the air with the fuelling door fully open (and other inhibits are not applicable). This is not a likely scenario, though.
Quote from: John H Watson on Tue, 24 Apr 2018 07:04
(c) The M2 fuel quantity is more than M1 and the M3 fuel quantity is more than M4.
(c) is only valid when the discrete is not set yet. Once it is set, (c) cannot reset it. Only the VTO+2000 condition can reset it. (c) is part of the hysterisis.
QuoteI see now at (a), that the message can appear in the air with the fuelling door open (and other inhibits are not applicable)
Understood. But why should the on-ground condition only be valid in connection with the refueling action? Aside from the fact that an inflight timer reset triggered by the refueling system is neither dangerous nor probable. Why should the on-ground condition not be connected with the power-up action as well? That action is even more likely to happen in flight (momentary DC bus failure).
I bet the on-ground thing is a "text layout" mistake and should actually refer to all three OR-conditions, not just to the refueling action.
By the way, if the fueling door closure is the trigger point in time where the 1000 lbs excess is measured, the final difference may be higher than 1000 lbs because, as we know, there is a lag in the sensed quantity indication. When the discrete has been set while the difference was 900 lbs, and the sensed difference now increases to 1100 lbs, you can now reset the discrete by repowering the FQIS. The repowering event measures the difference again, and it now sees it's more than 1000 lbs.
I'm wondering exactly what happens as a result of these three things (especially with the old software):
Quotea) The fuel quantity processor unit is powered up
b) The CMC ground test >FUEL QUANTITY IND is performed
c) The refuel panel door is cycled from open to closed and in ground mode.
With a), does powering up the FQPU
on the ground (only) make the cards go through a self test
and stop the message appearing/reappearing? (even if the logic says it should)With b),
on the ground (Ground Tests need the airplane to be on the ground), does it clear messages
and stop them reappearing?With c), the message is inhibited with the fuelling panel open. What happens when the door is re-closed? Does it re-arm the message
or does the message normally stay away? Is the new logic piggybacked on the old logic and designed to inhibit the normal(old) behaviour of the FQIS logic in response to these events.
Re a) Note that some aircraft computers behave differently on power up. e.g. when the Yaw Damper is powered up on the ground it runs through a test cycle. When the Yaw Damper is powered up in the air, it doesn't go through a test cycle (it gets straight to work).
Hope this makes sense ;D
(EDIT Disregard... If the airplane behaved this way with the old logic, it would be dangerous... The pilots wouldn't get the message during their preflights.
Quote from: John H Watson on Tue, 24 Apr 2018 11:18
I'm wondering exactly what happens as a result of these three things ...
It sets a discrete in the memory of the EIU (or elsewhere). This discrete inhibits the X FEED CONFIG message. Upon lift-off an 8 minute countdown starts. When the countdown is done, the discrete is reset, and the X FEED CONFIG message is no longer inhibited.
That's how I understand it.
This upgrade is just designed for the range between 0 and 1000 lbs. Outside this range it has no effect, and all will work as usual.
I was just trying to find a reason for why the text only said "on the ground" for the refuel panel door. I'm not entirely convinced that it's a bracketing problem.
Anyway, here's the FIM tests for the >X FEED CONFIG
841. EICAS Message >X FEED CONFIG (ADVISORY) - Fault Isolation
A. Fault Isolation Procedure
(1) The EICAS message > X FEED CONFIG (ADVISORY) will show when one or more of these
conditions are true:
(a) The No. 2 and/or 3 crossfeed valve(s) are closed with the flaps in the retracted position (10-
second time delay).
(b) The No. 2 and 3 crossfeed valves are not automatically commanded to the closed position
(takeoff command) with the flaps set for takeoff and the airplane in ground mode (one
minute time delay).
(c) The No. 1 and/or 4 crossfeed valve(s) are closed and the TANK TO ENGINE ALERT discrete
is not set (two minute time delay).
(2) The TANK TO ENGINE ALERT discrete is not set when any of these conditions are true:
(a) The airplane is in refuel mode (in ground mode and the refuel control panel door is open).
(b) The M2 fuel quantity is more than 32,000 lbs (14,500 kgs) or the M3 fuel quantity is more
than 32,000 lbs (14,500 kgs).
NOTE: M2 fuel qty is more than M1 VTO+2000 lbs (910 kgs) and M3 fuel qty is more than
M4 VTO+2000 lbs (910 kgs).
(c) The M2 fuel quantity is more than M1 and the M3 fuel quantity is more than M4.
(d) The M2 fuel quantity is more than 1000 lbs (455 kgs) greater than M1 and the M3 fuel
quantity is more than 1000 lbs (455 kgs) greater than M4.
(3) Corrective Action
747-400
(a) If the EICAS message >FUEL X FEED CONFIG (ADVISORY) shows, then do these steps to
clear the message:
1) Close the refuel control panel, P42 (if open).
2) Put the applicable crossfeed valve(s) in the correct position:
a) Open the No. 2 and 3 crossfeed valves if the flaps are retracted (not in the takeoff
position).
b) Make sure the No. 2 and 3 crossfeed valves are closed (automatic operation) when
the flaps are extended for takeoff and the airplane is in ground mode.
c) Open the No. 1 or 4 crossfeed valve(s) if the inboard main tank(s) contain more fuel
than the outboard main tank(s).
BAB ALL; AIRPLANES WITH FQPU BLOCK B UPGRADE (SB 28-2182)
d) Close the No. 1 and 4 crossfeed valves if: The M2 fuel quantity is not more than 1000
lbs (455 kgs) greater than M1 and the M3 fuel quantity is not more than 1000 lbs (455
kgs) greater than M4 and one or more of these events have been accomplished:
<1> The fuel quantity processor unit is powered up
<2> The CMC ground test >FUEL QUANTITY IND is performed
<3> The refuel panel door is cycled from open to closed and in ground mode.
(b) If the airplane is in the correct system configuration and the EICAS message >FUEL X
FEED CONFIG (ADVISORY) continues to show, then do these steps:
1) Confirm the correct airplane configuration per the above logic.
2) Make sure a magnet is not installed on the fueling power control switch (S681) (P42
control panel).
NOTE: The magnet may have been left on the S681 control switch after a refuel
operation. On the ground, with the magnet attached to the S681 control switch,
the FQPU will be in refuel mode, regardless of refuel door open/closed
position. With this condition, the >FUEL X FEED CONFIG (ADVISORY) will
continue to show until the No. 1 or 4 crossfeed valves are opened. This
condition can cause an incorrect preflight fuel-feed configuration (No. 1 and 4
crossfeed valves open and main tanks even).
3) Make sure the refuel panel door is closed and properly latched.
4) Use the input monitoring system to check the REFUEL DOOR OPEN discrete: E/15/355/
11.
a) Bit 15 = one (DOOR OPEN)
b) Bit 15 = zero (DOOR CLOSED)
5) If bit 15 is incorrect for the state of the door (open/closed), then do a check of the wiring
from the S681 control switch to the FQPU, M7860 (SSM 28-21-08) (WDM 28-21-23).
6) If the door open discrete is OK, then coninue.
(c) If the >FUEL X FEED CONFIG (ADVISORY) continues to show, do these steps:
1) If the EIU DISAGREE (STATUS) also shows, first do the corrective action for EIU
DISAGREE (PAGEBLOCK 31-61-00/101 Config 1).
2) Look for one or more of these CMCS messages:
a) 27600, 27602, 27627, 27628, 27679, 27687, 28676, 28677, 28774
Quote from: John H Watson on Tue, 24 Apr 2018 11:44
(c) The No. 1 and/or 4 crossfeed valve(s) are closed and the TANK TO ENGINE ALERT discrete
is not set (two minute time delay).
Does that mean that the X FEED CONFIG message is still enabled until the 2 minute countdown is over?
I thought there are three phases:
Phase 1: Tanks not equal, X FEED CONFIG msg enabled, FUEL TANK ENG msg disabled.
Phase 2: Tanks equal, X FEED CONFIG msg no longer enabled, 2 minute timer running, FUEL TANK ENG msg still disabled.
Phase 3: 2 minute timer passed, FUEL TANK ENG msg enabled.
In phase 2 both messages are disabled, I thought. The above quote suggests that the crew will always see a message, either X FEED CONFIG if they act too early, or FUEL TANK ENG if they don't act:
Phase 1: Tanks not equal, X FEED CONFIG msg enabled, FUEL TANK ENG msg disabled.
Phase 2: Tanks equal, X FEED CONFIG msg
still enabled, 2 minute timer running, FUEL TANK ENG msg still disabled.
Phase 3: 2 minute timer passed, X FEED CONFIG msg disabled, FUEL TANK ENG msg enabled.
Don't forget that the US President's personal 747 can be refueled in the air. Since it is fairly old, I wonder if much common software or hardware is shared by commercial 747s and "Air Force One"? Air Force One is based on a 747-200B.
Interestingly, the proposed replacement Air Force One (also a 747), may not have in-air refueling capability:
https://www.military.com/dodbuzz/2017/09/26/congress-questions-air-force-ones-lack-of-refueling-capability
I forgot "Air Force One" indeed :-)
Any clues here?
(Old software/hardware)
X FEED CONFIG logic diagram (http://www.members.iinet.net.au/~b744er@ozemail.com.au/PSX/xfeedconfig1.gif)
These logic lines that start with the word "120 SEC" come from nowhere. The diagram doesn't indicate what condition will start the 120 SEC timer.
For the FUEL TANK ENG message we know that the timer is started when the "MAIN TANKS EQUAL" signal is set (from the SET/RESET OR-gate output).
For the X FEED CONFIG message, which has its own 120 SEC input to another AND-gate, I don't know what condition will start that timer. That AND-gate has an inverted input from the "MAIN TANKS EQUAL" signal. Now is the timer starter linked with a point before the invertion or after the invertion? If the starter sits before the invertion, it starts running when the tanks become equal (inflight scenario); if it sits after the invertion, it starts when the inboard quantities become higher (refueling scenario).
If we assume that the designers wanted to have a delay for inflight scenarios (not refueling scenarios), the timer starter probably sits before the invertion, i.e. at the same spot where the other 120 SEC timer is started.
QuoteThese logic lines that start with the word "120 SEC" come from nowhere. The diagram doesn't indicate what condition will start the 120 SEC timer.
You have to wonder why they simply didn't put a timer in the circuit instead of an AND gate. Timer symbols can have a "0" or "1" symbol indicating a delay on a particular input. What would a delay on all inputs do to the logic? (Give the pilots a reprieve from any messages for a few minutes?)
Is there any 747 pilot on the forum who has ever set the tank-to-engine configuration at a time when the inboard quantities were within 0.0 to 0.2 tons below the outboards, i.e. before the FUEL TANK ENG message appeared? (It only appears when more than 0.2 tons below, not when equal.) And when you did so, was the X FEED CONFIG message shown until the differences were greater than ca. 0.2 tons?
I conclude from the complete diagrams that there is no point in time where neither the X FEED CONFIG nor the FUEL TANK ENG message is shown. Either one will always appear. The crew can't hit the right moment. They can only act too early or too late. No neutral phase in between. And the change-over point is never at the tanks-equal condition as most manuals suggest, but 2 minutes later.
|-|ardy
This will probably be no help whatsoever, as no rigorous investigation. but I have occasionally set tank-to-engine before the EICAS message appears, (when the tanks were clearly balanced and I was impatient). Usually then get no EICAS message. Occasionally I have subsequently got Crossfeed Config message, and had to briefly go back to the previous pump config. But I didn't notice exact tank quantities. I'll try to be more methodical, maybe balance 2 to 3 and 1 to 4 before tank-to-engine, to get ideal conditions, then investigate.
Thank you. I think the balance is not required for the message logic. Just one inboard needs to be 0.1 tons lower than its onside outboard. At that point the timer has been running for about 1 minute, and thus the X FEED CONFIG message should still appear when setting tank-to-engine.
maybe by the time I had completed moving the switches, the conditions were then correct to not then give the X-FEED CONFIG message ?
Yes, maybe there's another, undocumented delay on top of the 2 minute delay, just before the final message trigger. Perhaps 10 or so seconds ...
I've seen EICAS itself doing delays of 7-10 seconds after I flip the bit going to it. So unless EICAS has the 120 second timer, I would not be surprised to see an additional bit of delay, to know for sure that the 429 word on the bus is stable.
SELCAL though needs about 1 second only, if at all.
Hoppie
Philosophical Comment: This thread makes me wonder if some of the "message-timing logic" that isn't absolutely critical for safety wasn't just thrown together to get the job done on the real aircrafts and software.
Cheers,
Torrence
Quote from: Jeroen Hoppenbrouwers on Thu, 26 Apr 2018 23:11
I've seen EICAS itself doing delays ...
You mean the symbol generator of the EICAS screen? Or the EIU box?
If I read the diagram correctly, the first three logic gates to the left are in the fuel quantity system, and the rest to the right sits in the EIU, including the 120 sec delays.
|-|ardy
QuoteIf I read the diagram correctly, the first three logic gates to the left are in the fuel quantity system, and the rest to the right sits in the EIU, including the 120 sec delays.
Correct. I joined two diagrams together, one of the FQIS, one of the EIUs
Quote from: Hardy Heinlin on Fri, 27 Apr 2018 05:59
You mean the symbol generator of the EICAS screen? Or the EIU box?
The EIU. I never talk directly to the display units. I think the symbol generator is just dumb and paints immediately whatever is sent to it.
Does anybody know whether the symbol generator contains the message texts, or the EIU? In other words, whether the whole message is a single symbol?
Hoppie
Will anyone disagree if I add those 2 minutes to the X FEED CONFIG message logic as well, as indicated in the diagram? I'd like to get this part finished.
This will not only inhibit the FUEL TANK ENG for 2 minutes after the tanks became equal (implemented in PSX 10.33); it will also show the X FEED CONFIG message when tank-to-engine is configured during that same 2 minute phase.
|-|ardy
Quote from: John H Watson on Thu, 26 Apr 2018 01:54
QuoteThese logic lines that start with the word "120 SEC" come from nowhere. The diagram doesn't indicate what condition will start the 120 SEC timer.
You have to wonder why they simply didn't put a timer in the circuit instead of an AND gate. Timer symbols can have a "0" or "1" symbol indicating a delay on a particular input. What would a delay on all inputs do to the logic? (Give the pilots a reprieve from any messages for a few minutes?)
Reconsidering these "open-end" signal paths of both 120 sec delays ...
It could be that these delay paths are started by an undocumented "extra feature" (instead of being directly started by the output of the fuel quantity system). This "extra feature" might be something completely different; e.g. related to power-up, air-ground, refueling, jettison, whatever. If one of these systems can start the delays, the 120 seconds may already be passed when the fuel quantity system sends the main-tanks-equal signal to the EIUs.
This may also explain why they didn't put the timer in the circuit but installed an extra path into an AND gate.
|-|ardy
Quote from: John H Watson on Sat, 21 Apr 2018 13:51
(EDIT: A PPRuNer has finally given us the answer we were looking for. He says it's normal for the inboards to be 1~200kg below the outboards when the message appears)
If it's the whole range from
1 to 200 kg, there must be something special that does
not always delay the signal by 120 seconds.
I think I'll remove that delay from PSX until we know for sure what condition starts the delay. It has no hysterisis function anyway. The hysterisis is provided by the reset logic of the "equal+2000 lbs" stuff.
QuoteA PPRuNer has finally given us the answer we were looking for. He says it's normal for the inboards to be 1~200kg below the outboards when the message appears)
As we know, cockpit observations are often generalisations until we start focussing on the subject. Was the pilot really looking at same wing differences or opposite wing differences.
Does the text fully agree with the logic diagram or is it too vague? (excluding the service bulletin)
What do you think about my new theory that the 120 sec timers might be started by something not related to the FQIS output?
QuoteThe hysterisis is provided by the reset logic of the "equal+2000 lbs" stuff.
Sorry.. I'm still trying to catch up here. I'm not 100% sure how that reset/set gate works... or if it is a true reset/set gate.
http://www.members.iinet.net.au/~b744er@ozemail.com.au/PSX/FUELTANKENGlogic.gif
This is how I see it:
The "VTO" reset is a fixed fuel quantity value. i.e. inboard qty = outboard VTO qty plus 2000lbs (not just equal qty plus 2000lbs). When any inboard tank fuel qty goes below (or is equal to) outboard VTO plus 2000, zero logic outputs from the "VTO" gate to the inverter on the next AND gate and (assuming there are no huge 2000lbs variances) that next AND gate will always have an OK signal on the bottom leg. This leaves the M2<=M1 or M3<=M4 logic to output a tanks equal signal. I can see that
minor variations in quantity due to turbulence can produce a varying TANKS EQUAL output to the EIU (where the mysterious 120 second timer exists). That's why I think the timer is required (at least for the TANK/ENG message).
The VTO gate is an AND gate. I don't know if that adds an extra layer of complexity(?).
"One to two hundred kg" may very well mean "one hundred to two hundred".
"One to two hundred kg" may very well mean "one hundred to two hundred".
Hoppie... 1kg is not visible ... nor 10kg ;)
QuoteIf it's the whole range from 1 to 200 kg, there must be something special that does not always delay the signal by 120 seconds.
Just making sure.
That's a good point, Hoppie. I thought he meant to refer his "1", which cannot be seen on the EICAS, to the theory in the manuals that say "anything beyond equal". Sort of a poetic "1". But I think he's no poet and actually meant to say "100" to 200 because that can be seen on the EICAS indeed.
So I agree. I bet he said (and meant to write) "one hundred to two hundred".
Quote from: John H Watson on Wed, 2 May 2018 06:28
QuoteThe hysterisis is provided by the reset logic of the "equal+2000 lbs" stuff.
Sorry.. I'm still trying to catch up here. I'm not 100% sure how that reset/set gate works... or if it is a true reset/set gate.
I think it's a reset/set gate. Otherwise the VTO+2000 stuff wouldn't make sense. It's the only gate on the diagram that has SET/RESET labels. So they must have a special meaning. Those +2000 lbs provide a safe hysterisis zone.
Another theory: As this diagram is a compilation of two documents -- FQIS and EIU -- maybe they refer to two different versions. We have two hysterises here, one using 2000 pounds, the other using 120 seconds. Perhaps there was a version where just pounds or just seconds were used. So now in this compiled diagram one of the two methods may be a relict from an older version and should be removed.
Sorry, I thought this kind of expression was universal. e.g. 2~3000 usually means 2000~3000. Just a shorthand way of writing it.
The fault isolation manual* text is a written expression of all the logic and contains both the VTO and the 120 seconds. I don't think one set of logic is obsolete. Perhaps there is some situation which requires both. Might the VTO logic stop the messages appearing and disappearing all the time during refuelling? In the real world, during refuelling, the tanks are not refilled at the same rate.... Fuelling favours the tanks closest to the tanker (but imbalances are from time to time corrected).
The 120 second logic may be just to help with quantity fluctuations around the time the tanks reach the same value in flight.
*Note that the FIM is up to date as it contains the additional new logic.
Quote from: John H Watson on Wed, 2 May 2018 15:35
Might the VTO logic stop the messages appearing and disappearing all the time during refuelling? In the real world, during refuelling, the tanks are not refilled at the same rate.... Fuelling favours the tanks closest to the tanker (but imbalances are from time to time corrected).
So it could be that the 120 sec delay for the X FEED CONFIG is in fact designed for the refueling process, not for the fuel burn process(?)
Every time I look at the diagrams I seem to read them differently, but no, I don't see how 120 seconds would be long enough to take care of inboard/outboard imbalances during refuelling. I thought the VTO might offer longer term message inhibits (in combination with other parts of the logic). Anyway, looking at the FIM again, I see that the logic inhibits both messages during refuelling anyway.
In a YouTube video, a pilot said selecting the crossfeeds on during preflight "arms" the logic, but I don't know if this is part of the old or new logic or even if this is correct. The pilots do seem to turn on the crossfeeds irrespective of load. I don't know if this is simply a procedure to help the pilots remember to turn on the crossfeeds or if it does arm something.
Most likely I will be talking nonsense here (I am not very technically minded...), but could it be possible that these 120 seconds in the drawings only apply to aircraft that use main tank 2 for APU fuel?
If the APU is used for 20-30 minutes before Take-Off, main tank 2 will contain between 100 - 150 kgs less fuel than main tank 3. In flight, main tank 2 will reach "equal to or less than (main tank 1)", when main tank 3 quantity is still above the fuel quantity in main tank 4. Depending on duration of APU usage on the ground and fuel flow in flight, it will take a minute or more before main tank 3 also conforms to "equal to or less than ...", thus satisfying > FUEL TANK/ENG.
Regards,
Simple Simon
Quotebut could it be possible that these 120 seconds in the drawings only apply to aircraft that use main tank 2 for APU fuel?
Nice theory. Note, however, that our fleet has both types, but they seem to have the same logic in the FCOM and Maintenance Manuals. Also, 150kg is not a lot in the real world. Engines burn fuel at different rates. I may even have seen 100kg difference between random tanks directly after refuelling.
As an afterthought... I seem to recall (vaguely) that the APUs which use both inboard tanks still tend to favour the left inboard tank (slightly). Perhaps it's because the fuel has to travel a greater distance from the right wing. The fuel line to the APU still goes via the #2 tank.
Independent verification required ;)