Posts about: "Auto-pilot" [Posts: 34 Pages: 2]

Ozegrade3, Biggles78 and all,
I agree, the more of the history that we can write down somewhere, the better....
Just look at the "Did You Fly the Vulcan?" thread here on PPRuNe....

A chance remark by M2dude reminded me of something I meant to write about sometimes... and that has barely been mentioned in the various Concorde stories.
It's the huge gap between the prototypes on the one hand, and the pre-production and production aircraft on the other hand.

It's not just the visor, or the shorter tail.

In my own "field", the AFCS (autopilot, etc.), there was not a lot of similarity between the prototypes and their successors.
The prototypes were "proof-of-concept", designed in the early to mid 1960s.
The pre-production aircraft were designed in the end of the '60s and already close to the production aircraft in most respects.

Some of this difference was due to the very sudden and rapid evolution in electronic technology, with the arrival of the integrated circuit in particular.
The microprocessor - in a way just a large integrated circuit - didn't arrive in time... I don't think there was a single microprocessor on board Concorde until the days that they had to fit TCAS (in the '90s, IIRC).

I'll have to see how to do it.... maybe write it off-line and post snippets on here, then move it into a blog or suchlike?

Galaxy Flyer
We never had a case of lost pressurisation, ever. The cabin windows had dual laminated panes; an inner pressure pane and an outer thermal pane. We had dual systems that kept the cabin at a max diff' of 10.7 PSI, the engine bleeds pushing about 200lb of air per minute into the cabin. This allowed you to fly the cabin at an altitude of around a 6000' maximum only, right up to TOD. If you HAD to fly subsonically, the ideal was Mach 0.95 at FL290. (Subsonically the aircraft had similar range to normal, but took well over twice as long of course). If however you had to shut down an engine, your range deteriorated quite dramatically, and a diversion was usually sought.

It's great that Bellerophon is posting here again; we need a steely eyed Concorde pilot's input here (not just the boffins/nutters and nerds [that's me ]. To touch more on a couple of his valid points;
Fuel burn: The aircraft would naturally require less fuel as she became lighter and as a consequence gently climbed to maintain cruise Mach number, this is what the engine control system was doing all the time, even though the throttles were wide open it was 'tweaking'.. BUT, the decreasing IAS as you climbed, due of course to the reducing density, just like any other aircraft meant that drag was reducing too, so it was a combination of both of these factors, reducing weight and reducing drag.
Flying controls: It was a slightly weird but wonderful arrangement; pilots inputs would move a servo valve in the hydraulic relay jack, the jack would move in response and drive both a resolver AND mechanical linkages. The resolver ourput was sumed with the flying control position resolvers, and the error signal was fed into an autostab' computer, where it was summed with stabilisation demands (primarily axis rate and acceleration). The autostab computer would the directly drive the surface, and the reducing error signal would reduce the demand etc. While all this was going on, the mechanical linkages would slavishly follow, but as long as you were in FBW (what we used to call 'signalling') mode, these mechanical inputs were de-clutched at the PFCU, so did nothing at all. Only if there was an EXTREMELY unlikely failure of BOTH FBW channels would these inputs be clutched in and the flying control group (rudders, inner elevons or outer and mid' elevons) would then be in Mechanical signalling. The system redundancy was checked after engine start on every flight. But to reinforce what Bellerophon stated, there was no mechanical reversion here; without hydraulics you had nothing. Another aside here; the designers, being paranoid like all good designers (no offence Christiaan ) were worried what would happen if the controls would somehow jam up. A jammed mechanical flying control input run itself would have no effect on FBW operation whatsoever, due to spring boxes being fitted to the runs. A 'Mech Jam' light would be set, together with a separate red light and audio warning, but this was all. But to completely protect against the aircraft was fitted with a Safety Flight Computer (SFC) system. The idea was, if a control axis (pitch or roll only) jammed up, the captain could press down on a switch light set between the two halves of his control wheel, (at the centre of the 'W') and the Emergency Flight Controls would activate. Strain gauges at the front of the control wheel, two sets on each control column for pitch and roll axis, would input into an SFC that would covert the control force into an elevon demand. These commands were then fed into the autostab' computers, and hence directly into the controls. (A little like L-1011 CWS in a way). There was a little test button that was used to test this system, again after engine start. So although the controls were jammed, the aircraft could still be flown. (Never used in anger I'm pleased to report).
But there was a problem; if this system was inadvertantly used, the results could have been catastrophic, as the system was extremely sensitive indeed, and full elevon movement could be enabled with only moderate effort. Because of this hairy prospect some safeguards were obviously put in place. The first safeguard was an interlock in the autostab' engage logic; If the switchlight had been inadvertently selected beforehand (the light was green by the way) you would not be able to engage pitch or roll autostab's (both channels too) so you would not be going flying until that was fixed. The second safeguard was a little more subtle; A plastic, frangible cover was fitted over the switchlight, unless the captain pressed reasonably hard the cover would prevent the switchlight from being pressed. At least that was the theory, in practice this little bit of plastic could be a pain in the ass . It was carefully fashioned, and I seem to remember BAe charging the airlines a few hundred pounds each for these things. If some wally fitted the cover upside down (and unless you were careful it was easy to do) THE THING WOULD NOT BREAK!! I remember at Fairford in 1976, G-BOAD was on pre-delivery flight testing, and the late great test pilot John Cochrane was doing a test of the system. The cover on this occasion HAD been fitted upside down, and of course he could not plunge his thumb through it and engage the EFC button. After trying everything, in the end he removed a shoe, took out his pen, and smashed the plastic cover until it broke. (It's OK, the autopilot was engaged at the time). Unfortunately, his combined shoe/pen emergency device also wrecked the switchlight as well, so the system still could not engage. (There was only a switchlight on the captain's side). After he landed and he confronted us all with his dilemma, he was shaking; not with rage but with laughter. (This was the great John Cochrane, sometimes the dour Scotsman but he was always able to see the lighter side). After that event, careful instructions were issued regarding the fit of the cover, and it was modified and made a little more frangible.

Biggles78
This is the minimum Mach number that can be flown with the existing CG. (which would be around 59%). Just as the CG indicator (not shown in this photo) gave minimum and maximum CG for a given Mach number, the Machmeter gave a reciprical indication also). You can also see that as the aircraft is not flying at Vmo any more, being at Mach 2 cruise, that the VSI pointer is now away from the orange and black Vmo bug. At our 'not so coffin corner', now that the aircraft is at maximun alllowable altitude, Vmo would naturaly coincide with Mmo; the orange and black Mmo bug being shown at Mach 2.04. This really superb photo taken by Bellerophon gives a graphic illustration of what the panels looked like at Mach 2. Note that the with the TCAS VSI Concorde retained it's original linear VSI also. (Miust have beeen the only aircraft flying with FOUR VSIs. (The originals had to be retained due to the fact that the autopilot Vert' Speed Mode error was derived from the indicator itself. As far as TCAS goes, R/As werer inhibited above FL300 (on acceleration this would coincide with the aircraft becoming supersonic, and the mfrs would not countenance the aircraft doing extreme manoeuvrs as a result of TCAS RAs at supersonic speeds).
The tail wheel was lowered for all 'normal' gear cycles (not stby lowering of free-fall). It was designed to protect the bottom the nacelles in the case of over-rotation, but in practical terms the thing was a waste of space (and weight) and a simple tail skid (used on the prototypes) would have sufficed. Any time that the tail wheel contacted the ground, it would ALWAYS collapse, damage the tailcone structure and in fact aforded no protection whatsoever. Fortunately these events were EXTREMELY few and far between. The biggest problem with the tail wheel was a major design flaw: On gear retraction the assembly would retract in sequence with the nose and main gear, and as it entered the opening in the tailcone, it would release over-centre locks that were holding the spring-loaded doors open. The doors would then firmly spring shut behind the gear assembly and finish the job. UNFORTUNATELY this was a very poor design; if for any reason one of the two doors had not gone over-centre on the previous gear lowering, it would be struck by the retracting tail wheel gear and cause structural damage to the local skin area, that would have to have a repair done. Unfortunately these events were not quite so rare, and several measures were tried to reduce the chance of this happening. Although not a safety issue, it was an issue that was a total pain. (As a matter of interest, G-BOAC had this happen on one of it's first test flights out of Fairford in 1975).
Nick Thomas
As ChristiaanJ said, the last two BA aircraft WERE lighter than the others, and would be preferred aircraft for certain charters. But that is not to say that any aircraft could not happily do ANY sector. We fortunately had no distorted airframes in the British fleet, so this was never an issue. There was very little spread, regarding fuel consumption between different engines; one of the best parts about the Olympus 593 was that it hade very little performance deterioration with time, it was an amazing piece of kit.
Time on wing for the engines was a real variable. Each engine was built up of modules, each one of these had a seperate life. In the early days of operation, time on wing was quite poor, and MANY engines would be removed on an attrition basis. One of the early failure problem was the fuel vapourisers inside the combustion chamber were failing, taking bits of turbine with it!! A Rolls Royce modification that completely changed the design of the vapouriser not only solved the problem completely, but also increased the performance of the engine. As the engine matured in service time on wing greatly improved, and in service failures became a thing of the past. A 'trend analysis' was done after each protracted supersonic flight, where engine parameters were input into a propiatry RR computer program, that was able to detect step changes in the figures, and if this were the case, more boroscope inspections were carried out. The OLY time on wing was nothing compared to the big fan engines, but the conditions that it operated under bore no comparison. Not really sure about absolute figures on this one Nick, I'll ask one of my Rolls Royce friends and see if I can find a figure.

On the ITVV Concorde DVD the Captain explains that during the cruise climb at mach 2 the auto throttles were armed and would be used as required. Then during descent the throttles were gradually pulled back whilst the autopilot was given an IAS of I think 350 knots. Therefore the plane would have to descend to maintain that speed. He explained that the power settings were chosen to ensure that there was sufficent cooling etc.
My query is if an autoland was to be undertaken was the auto throttle system able to maintain the required landing speed or would the pilot have to manage the throttles? I guess that decreasing or increasing the pitch to control speed when on the glide slope would not be a good idea.
Thanks again
Nick

Approach handling was an interesting exercise - being so far down the back of the drag curve (over 100kts below best L/D) very accurate thrust handling was required.

The autothrottle was always used for approach if servicable, except for a two-engined approach, and was absolutely excellent. There were two, operating in parallel, and if the speed was more than a couple of knots out for more than a few seconds it was unusual. In IAS ACQ mode there was an active input from the INS which tracked grounspeed and so enabled anticipatory throttle movements during wind changes - if an autoland runway was available the preferred method of dealing with an approach likely to experience windshear was to carry out an autoland. (I think I speak for everyone, though, if I express a preference for the just going somewhere else option).

In Supercruise one autothrottle only was rearmed if the associated autopilot was engaged - it wasn't active but was available to cater for sudden drops in temperature which could cause unpleasant sudden high rates of climb if the temp shear was sufficient.

The rest of the flight - basically manual thrust, usually with the throttles fully forward.

Maybe one day we'll start making aeroplanes with such sophisticated systems again...........there's still lots that hasn't been hinted at on this thread

M2dude ,

Thanks for bringing up the story of the temperature shears at low latitudes, saves me some explaining !

The phenomenon was not really "discovered" until the route proving started.
As you say, it meant abrupt changes in Mach and Tt.
Since the pitch autopilot in, say, 'Mach Hold', had only one way to cope with those changes : pitching up or pitching down, this resulted in abrupt climb and descent manoeuvres that were totally inacceptable.

It was not only fixed by modifiying the intake controls.... !
It also led to a fairly major mod in the AFCS, by bringing the auto-throttle into the loop.
Instead of letting the pitch A/P take the aircraft into a zoom climb to try and stay within the "speed limits", it woold be the autothrottle pulling the throttles back.
After a fair amount of flight tests, this became the final form of the 'Max Cruise' mode.

I know... I should have kept a diary.
All this happened 35 years ago, so I can't put exact dates, or even specific aircraft, into the story. At least I still have some Concorde doc, that at times allows me to refresh my memory.

But one thing stayed in my memory over the years ...

The initial A/P+A/T 'Max Cruise' mod meant a major modification of the autothrottle computers. One of the circuit boards in particular was totally "butchered" on-site (Fairford), with well over a hundred track cuts and new wire links, not to mention the number of component changes (to give an idea to the 'experts', these were double-sided PCBs of about 15x20cm).

We got the job done (4 boards : 2 boards per computer, 2 computers), got the computers tested and on the aircraft. Knowing full well how easy it was to introduce faults and problems during such a modification (a dud solder joint could be enough), we expected to see them back within days, or at least within a few flights.

Well..... those computers left the lab with each about a thousand hours "on the clock" (they have little elapsed-time counters that indicated hours under power, NOT flight hours).
The first time we saw them back (for a minor mod, not a fault), both clocks showed over 10,000 hours !

Ah, those were the days

CJ

ChristiaanJ
aaah yes, Max Climb/Max Cruise modes. I'd not forgotten this my friend, I was going to say a few words about that in a future post, but maybe we can do that now. (And I'd love to hear more of your comments on this here too, ChristiaanJ). The intake and autopilot modifications were in a way complimentary it's true, but really dealt with separate problems, at least in my view:
The intake control unit software change (a change to the control law that limited engine N1 as a function of intake local Mach number, Mo, and inlet total temperature, T1) was able to put an absolute limit on aircraft achievable Mach number during Mmo overshoots, but it would not PREVENT Mmo overshoots occurring altogether, it was more of a safety brake. This particular overspeed problem manifested itself well before route proving, and in fact the intake system 'fix' resulted in the Thrust Auto Reduce System being deleted, electronic control boxes and all. The TAR system was fitted on all development aircraft equiped with the digital intake system, and it tried (in vain) to limit extreme Mach overshoots. The production aircraft retained the TAR wiring and locked out circuit breakers, as well as two vacant spaces on the electronic racks. The prime reason for all these efforts were that some of the rapid excessive Mach overshoots quite often drove the intake into surge; the modification to this N1 limiter control enabled engine mass flow to be controlled in such a way that these surges could be prevented during temperature shears. The aircraft Mach limit was an extremely useful fringe benefit.
The AFCS mode change from what was Max Op and Max Op Soft (always loved that name) to Max Climb/Max Cruise was at a stroke able to deal with the regular Mmo overspeeds that kept on occuring during, as you say, the route proving trials of 1975, when British aircraft G-BOAC and the French aircrfraft F-BTSD carried out pre entry into service evaluation flights, SD sadly was the aircraft that was tragically lost at Gonez in July 2000). The Max Climb/Max Cruise AFCS mode combo is a mode like no other that I've personally seen before or since anywhere, (it for instance resulted an elsewhere taboo; an autopilot and an autothrotte working together IN A SPEED MODE).
This problem encountered primarily at lower lattitudes, (for example, G-BOAC doing route proving flights out of Singapore), occurring initially as the aircraft reached Mach 2. It was termed 'the insurmountable problem', but the AFCS designers (such as ChristiaanJ) fortunately did not have 'insurmountable problems' in their vocabulary. The issue was that the aircraft would have been climbing rapidly at Vmo of 530 KTS, with throttles at the gate as usual, At exactly 50,189' we hit what was known as 'the corner point' in the flight envelope, where 530 KTS IAS equated to Mach 2 exactly. Max Op mode would then 'let go' of the Vmo segment, and try and control the aircraft to Mach 2. (As the aircraft climbed, Vmo itself would progreesively decrease in order to equate to Mmo, or 2.04 Mach). But in very cold conditions, the aircraft still 'wanting' to accelerate, and the simple Max Op/Max Op Soft modes just could not cope with gentle pitch changes alone. The problem became even bigger during the cruise/climb when severe temperature shears occured, and routinely regular Mmo exceedences occured. Something had to be done, and something WAS done and how; enter Max Climb/Max Cruise. It was really a classic piece of design, where the aircraft would do the initial supersonic climb in Max Climb mode. This mode itself was relatively simple, in that it was more or less a Vmo -Vc hold mode. That meant that the difference at selection between indicated airspeed, Vc and Vmo would be maintained, with a vernier datum adjust to this being available. In practice this mode was selected pretty much at Vmo, so datum adjusting was not always required. Now comes the clever part; the autothrottle, this would operate in standy mode at this point, just waiting there doing nothing, with the throttles at maximum as before. So the aircraft would now climb as Vmo increased to 530 KTS, and then following a now constant Vmo of 530 KTS until the magic 'corner point' (51, 189' remember). Now all hell would break loose; the mode would automatically change to Max Cruise, the autothrottle would also be automaically selected to Mach Hold mode (initially datumed here to Mach 2) and the throttles would retard, attempting to hold this Mach 2 datum, and the autopilot is commands a 'fly up' signal, over a 20 second lag period to 600'/minute. Now comes an even cleverer (?) part; the autothrottle Mach Hold datum is gradually increased over a 100 second period towards Mach 2.02, and so in stable conditions the throttles would now gradually increase again until they once more reach the maximum limit. At this point, the autothrottles now come out of Mach Hold mode and back into the waiting in the wings standby mode. The autopilot would now cancel it's 600' fly up, demand, returning to a datum of Mach 2. There was a little more complexity built in also, where the difference between the 'commanded' and actual vertical speeds offset the autoplilot Mach 2 datum. This would apply whether the autothrottle had cut in (+600'/min demand) or with the throttles back at maximum (0'/minute demand. A positive climb error tweaked the cruise Mach up slightly, a negative error (eg. in a turn) the converse was true. The effect of all of this complexity was that the aircraft itself could 'scan' until it settled at a point where the throttles could be at maximum, and the speed between Mach 2 and 2.02. On the North Atlantic, with warmer ISA temperatures, there was usually just the initial routine with the autothrottle as you hit the corner point. However at lower lattitudes (eg. LHR BGI) there could be a few initial autothrottle intercepts before things settled down. This whole incredible routine completely took care of the insurmountable problem, a problem that was shown not only to be insurmountable, but was put to bed forever, by people like ChristiaanJ.
I hope that my explanation here does not sound too much like gibberish.

EXWOK
I think you've guessed right as far as my identity goes; it's great that it's not just Concorde pilots I can bore the socks off now
PS. I bet the ex-SEOs LOVED your comments

Dude

stilton

...In your superb photograph ... thank you ... both RMI'S shown seem to be showing a heading of 230 while the heading on the HSI shows 220 with the annunciation TRUE above the HSI...I am thinking that the RMI's are showing magnetic headings ?...

Correct. The RMIs show 230\xb0(M) and the HSI shows 220\xb0(T).


...were you operating on TRUE headings in this case as it seems...

Yes.

The HSI had earlier been selected to show TRUE data, by moving the RAD/INS switch (out of picture) to INS. This changed the data inputs to the HSI from Magnetic to True, but did not affect the data input to the RMI compass cards. The data input to the RMIs was always magnetic, supplied by one of two compass systems; normally #2 compass system fed the Captain\x92s RMI card.

The red/black flags visible in the VOR RMI indicate that both the VOR stations selected are out of range.

The flags visible on the HSI indicate:

• TRUE.......True compass data (not magnetic) is being displayed. (TRUE/MAG)
  
  
• INS..........An INS (not a VOR) is supplying Nav data to this HSI. (INS/RAD)
  
  
• 1............#1 INS (not #2) is supplying Nav data to this HSI. (1/2)
  
  
• HDG.........The steering index shows selected heading (not track). (HDG/TRK)
  
  
• LIN...........The beam bar is displaying linear (not angular) displacement. (LIN/ANG)

... what was the reason for this ?...

When flying long legs over areas of the world with few/no radio facilities, aircraft generally navigate from one Lat/Long waypoint to the next Lat/Long waypoint by using their INS systems, rather than tracking from radio beacon to radio beacon. The North Atlantic Track system is good example of this, as is this route down to BGI. Because there is no useful magnetic/radio information to display in such areas, True (INS) information is usually displayed, which brings with it two main advantages.

Firstly, you get to monitor how well the autopilot is doing, because the beam bar in the HSI should always be centralised if the aircraft is on track.

Secondly, as the aircraft changes course over a waypoint, you get the chance to check the Initial True Track it then takes up corresponds to the ITT pre-calculated on your flight plan. This is an important check (called the \x93Waypoint Change Drill\x94) in BA, that you have not mis-programmed the INS waypoints!

This procedure, of navigating in True, is by no means a Concorde procedure, all long range BA aircraft utilise it, and it is in common use in most airlines and aircraft flying long range routes around the World.

Best Regards

Bellerophon

The PROM board is from an AICU, the other board is from an AFCS computer.
Could be Autostab, Lateral Autopilot or Trim, since they all used exactly the same technology, board size, etc.
Looking at the board, I think it's 'Lat A/P' but I can't be certain.
No excuse needed about missing the "central connector" !
It was really a "one-off" feature, invented by Bendix, and abandoned afterwards. I'm not even sure the early A300B AFCS computers, that used much of the Concorde technology, still had them.

See my post on the subject. I doubt there was any core memory at all in the production aircraft.

CJ

Nick Thomas
[QUOTE]Going back to expansion and paint. With the aircraft expanding approx 6 inches and a temp change up to 127`c, I guess a special kind of paint; able to withstand such adverse conditions; must have been used? When deciding on the paint specification was any consideration given to the overall weight of the paint?[/QUOTE
Can't remember much about paint spec's, but a lot of experimentation/trial and error was carried out with different paints until the right one was found. I remember when G-BOAD was delivered, that copiuous sheets of paint had peeled off in flight. Finally a superb polyurithane paint was found that did the trick perfectly.
Yes Nick, the life of the airframe was limited by the number of supersonic cycles, however modifications carried out extended the life of the airframe significantly. (and more were planned).
And the 'hat in the gap' stories are quite true.

ChristiaanJ

Quote:

Many years later, I discovered that several airline Concorde pilots did not even know the function existed....

This was the real beauty of the autostab' on all 3 axis; you could just safely take it all for granted. The Mach 2 engine out case was a classic, as not only would the aircraft yaw towards the dead engine but there was an adverse roll input, where the wing on the same side would LIFT due to the excess intake air for the failed automatically being 'dumped' through the now open spill door. If for any reason the aircraft HAD been under manual rather than autopilot control, then life without autostab would be rather uncomfortable to say the least. And putting further Concorde's achievements in terms of stability; the world's only previous large delta winged Mach 2 aircraft, the B58 Hustler, had the slightly awkward feature in the case of an outer engine failure at Mach 2, in that the yaw forces were sufficient to tear the fin off. This happened on more than one occasion during service life of the Hustler, but engine failure (or far more likely a deliberate precautionary shut-down) although hardly a non-event in the case of Concorde, it was routinely dealt with without drama or danger.

Dude

ChristiaanJ
During AUTOLAND a flare manoeuvre was instigated by the Pitch Computer at 50' radio, where a fairly simple flare law was invoked. I seem to remember that the law , which used a combination of radio rate (from the RadAlt) and vertical acceleration (from the INS) gave you a commanded height rate of 10'/second at 50', exponentially reducing to 1.7'/second at point of main wheel touch down.
The autoland on Concorde was both extremely accurate and reliable, and an awful lot of guys said they hated using it 'because it can land the aircraft better than I can'; their words NOT mine. (Personally I never bought that one, the guys were just modest as far as I was concerned ). This in my opinion is an absolute testament to the AFCS designers; ChristiaanJ and his colleagues at SFENA and GEC Marconi.
To give the complete final approach story; as the aircraft tracked the glideslope in LAND mode, the autopilot G/S deviation, like most aircraft, was geared as an inverse function of radio altitude, and at 75' radio this deviation was flushed down the loo altogether, leaving the A/P to hold radio rate for just a few feet. At 50' the flare was instigated, and at around 35' DECRAB was commanded, where the yaw channel would use a rudder input alone to 'kick off drift' and align the aircraft with the runway centreline. (Concorde did not employ a fwd slip manoeuvre in crosswind conditions, being a slender delta). The 'final' command was at 15' radio, when the autothrottle smartly retarded the throttles. (The Pitch Computer flare law of course continuing to control decent rate all the way down). On touchdown the autopilot would be manually disengaged and the nose gently (usually ) lowered to the ground. (Concorde was only designed and certified as a CAT 3A system, so there was no automatic rollout guidance. However there was a runway guidance symbol on the ADI, which used a combination of Localiser deviation and lateral acceleration, to give you runway rollout track).
Now the flare law was tested every autoland, at G/S capture, and failure of this test resulted in the loss of LAND 3 status on the landing display panel. The most common defect of all with the Concorde autoland was in fact failure of the flare test, when at G/S capture, the previously illuminated LAND 3 indication would drop all of it's own to LAND 2. A simple changeover of autopilot paddle switches would nail the offending Pitch Computer, which would then be replaced before the next trip.

Dude

Brit312 ---- that reminds me....

On the two prototypes, the AFCS controller was located on the centre pedestal, and had two big rotary multiposition switches to select the autopilot/flight director modes, plus the autopilot, flight director and autothrottle engage switches.

On the preproduction and production aircraft, the design was totally changed and became the big box with rows of pushbuttons and little handwheels, to set things like altitude and speed, that one now sees at the top centre of the instrument panel, just under the windscreen and above the engine instruments.

It was not until years and years later that I had an occasion to visit a VC-10 cockpit .... to discover how similar the VC-10 AFCS controller was to the prototype Concorde one.
No wonder... both were made by Elliott.

CJ

Hi Steve, good question. You could not prime Alt ACQ directly from MAX CLIMB/MAX CRUISE, but I'm sure the guys would confirm that FL600 was nonetheless set in the AFCS altitude window. If you did come close to FL600, then you would get an altitude alert audio and visual warning anyway, but the crew would obviously know anyway. The guys will have to confirm this but I think that ALT HOLD would be selected at this point, and the autopilot would now constrain the altitude by varying pitch attitude. You would normally be flying in nice cold conditions for this to occur, hardly ever on the North Atlantic, but on the LHR-BGI sector it would occur as often as not. (As usual, apologies to my all flying pals if I'm talking rubbish again ).

Dude

Hi Dude,

Thanks for the reply, I never knew you couldn't prime ALT ACQ from MAX CLIMB / MAX CRUISE.

I was thinking though that if you had ALT HOLD selected then the autopilot wouldn't be able to vary the pitch attitude and it would hold at the ALT HOLD level. Once there it wouldn't be able to use pitch to increase or decrease speed if necessary. I can see how the autothrottles could hold the speed back but if it were to get warmer then even at full power it may not be able to maintain speed at a fixed altitude.

I'm sure there is a simple answer which will leave us saying "oh yeah, how obvious, didn't think of that" Just strange it doesn't seem to be written anywhere.

Thanks,

Steve.

spfoster
Hi again Steve. At FL600 your aircraft is going to be relatively light, and your autopilot is set to ALT HOLD mode. As the aircraft continues to burn fuel it will 'try' and climb, but the ALT HOLD mode will of course slightly reduce pitch attitude to keep prevent the aircraft from climbing. The autothrottle in MACH HOLD mode can keep us from straying from Mach 2, large positive temperature shears were to my knowledge almost unheard of, and so any speed reduction was exceedingly unlikely. We really need a pilot's input here to further clarify the realities of LHR-BGI supersonic cruise; that's the way the AFCS was designed to do things however.

Dude

I was wondering if it was possible to hand fly Concorde when she was supersonic? Therefore would the failure of one or both autopilots mean that you would have to divert to the nearest airport? Also how long after takeoff would it normally be before engaging the autopilot?
Thanks again
Nick

I should let the pilots on this thread answer, really,
But I would say yes , just tedious, and needing the PF to pay far more attention minute-to-minute to all the various basic flight parameters, and fly them, rather than just monitor the autopilot doing the job.

Basically, no.
The autopilots were quite reliable, and what's more, they were essentially independent, so the probability of both failing during the same flight was pretty remote.
If one dropped out in flight for whatever reason, you'd engage no. 2 and continue.

The only situation where losing both autopilots within a minute or so of each other would be critical, would be during a Cat.III autoland, and just before that you'd run an autotest of both computers. The probablity of then losing both almost at the same time during those last few minutes was in the order of 10E-09 or less, and indeed never happened.

Can one of our pilot friends oblige?

CJ

PS I should add, that normally only one of the two autopilots was engaged, with the other powered but inactive, and IIRC, with AP1 active, AP2 would refuse to engage until you disengaged AP1.

Only in LAND mode could both APs be engaged at the same time, with normally no.1 flying and no.2 as a "hot" standby.
The system was referred to as "fail active", in that no.2 would already be synchronised to what no.1 was doing, and would take over totally automatically, without a hiccup (except an "oh merde" from the pilots, probably).

Quoting from memory.

ChristiaanJ
Your quote from memory is correct, although Land PRIME only was required for the second autopilot to be engaged.
Autopilot disengagememnts in Land mode werer in fact extremely rare. (Love the 'oh merde'bit though). The most comon autoland problems were the loss of Warning and Landing Display LAND 3 annunciation. Most problems were due to a failed flare test in the Pitch Computers at G/S capture and failures in the BCII inertial comparator.

Dude

I have to admit that some of the subsonic fuel burn figures for Concorde were truly eye watering, and without massive engine and airframe modifications there was precious little in service that could be done to improve things. Paradoxically improvements to the supersonic efficiency of the powerplant were easier to implement, and several modifications were implemented, tried or proposed to improve fuel burn:
Way back in the late 1970's we did a major modification to the intakes that increased capture area by 2.5% and gave us typically a 1.6% improvement in trans-Atlantic fuel burn, and although this was our biggest performance improvement modification, there were more:
The famous elevon and rudder trailing edge extension modifications (that due to poor design, produced in later life the water ingress induced honeycomb failures) together with the re-profiled fin leading edge modification, I never saw the performance gains quantified (anyone have any ideas?).
Can anyone here remember the riblet trial? In the mid 1990's Airbus supplied 'stick on' plastic riblets, applied to various areas on the under-side of the wing on G-BOAG. These riblets had very fine undulations moulded into the surface; the idea being that as the air flowed through and around the riblet patches, boundary layer turbulence, and hence induced drag would be reduced. Now, the performance gains (if any) were never quantified, mainly because the riblet patches either peeled off or the surface deteriorated with the continuous thermal cycle. (I was over in JFK when the aircraft first arrived after having the riblets fitted, and as the crew were trying to proudly show me these amazing aerodynamic devices, they were sadly embarassed, as several had dissapeared in the course of a single flight).
There was one modification, proposed by Rolls Royce in the late 1990's that did have quite a lot of potential; this was to increase the engine N1 by around 1.5%. This would have had the effect of increasing engine mass flow and therefore reducing the drag inducing spill of supersonic air over the lower lip of the intake. Depending on the temperature, the performance gains were in the order of a 1.5% improvement in fuel burn at ISA Plus upper atmosphere temperatures ('normal' LHR-JFK) to none at all at significant ISA Minus temperatures (LHR -BGI). The modifacation had been trialed on G-BBDG before her retirement in the early eighties, and was proven in terms of performance enhancement and engine stability. In order to keep TET at the pre-modification level, there was a small increase in N2 commanded also. (The higher N1 required an increase in primary nozzle area, reducing TET). The main reason for the modification not being implemented was one of cost; The Ultra Electronics Engine Control Units were analog units, and the modification was a simple replacement of two resistors per unit. However because ultimate mass flow limitation was also controll by the digital AICU (built by British Aerospace Guided Weapons Division) the cost of getting a software update for this exremely 'mature' unit was found to be prohibitive.
A certain 'brainy' SEO and myself were working on a modification to improve fuel burn on ISA minus sectors. The idea was to force the autopilot, in Max Cruise at low temperatures only , to fly the aircraft close to Mmo, rather than at Max Cruise speed of Mach 2 - 2.02; this would have given us gains of up to 1%, depending on the temperature. The basic electronics involved for the modification were relatively straightforward, but it was never pursued due to the complexity of dealing with temperature shears and the cost of certification.

Dude

Hi DavvaP, and welcome. As far as ice on the wing goes, I'm sure as any of my pilot friends here will agree that she was treated just like a subsonic in that regard; any ice or snow build up on the surfaces of the wings would not be tolerated and would have to be removed before flight. (She may have had a revolutionary wing design, but still this was a wing nonetheless ). She would also require pre-flight chemical anti-icing/de-icing treatment from a ground truck just like the rest, in shall we say, 'less than tropical conditions'. (Winters in Prestwick during crew base training... such fond memories ). As far as active ice protection on the wings, there was a highly sophisticated Lucas electrical 'spraymat' system fitted, but only the wetted areas of the wing, forward of the engines were 'covered'. Two digitall cyclic timers (CTPUs) would automatically regulate cyclic switching on and off of 115 VAC for various load areas of the wing at a time at pilot pre-selectable intervals (2, 4 or 8 seconds). Also as part of this system, there was continuous de-icing for certain other load areas too, so you had a mix of cyclic and continuous de-icing in operation. The whole idea here was to prevent chunks of ice entering and damaging the engines, the only other areas of this electrical de-icing system were the intake lips and side-walls and also the D Box area above the auxilliary inlet vane, built into the spill door. (This would only operate if the auxilliary inlet door itself was open). The whole shooting match would automatically switch itself off, for obvious reasons, above a TAT of 15\xb0 C. (ie. the vast majority of the flight). The only other de-icing system (apart from the galley drain masts) was on the engine inlet guide vanes, but this was purely pneumatic and again would swith itself off above 15\xb0 C.
I think you will find that precious little of Concorde is now not generally available in the public domain, some control software and laws are still I would expect covered by some sort of patent. (That is why when I publiished here the engine 'E Schedule' graphs I deliberately deleted the equations for the various running lines.
Your efficiency question was a valid one; as IAS and Mach number increase the aerodynamic drag (in all it's forms) will generally increase, but the efficiency OF A WELL DESIGNED powerplant wil also increase, and Concorde was definately no exception here. The real beauty of Concorde was just HOW MUCH the powerplant efficiency increased with increasing speed and more than totally eclipsed the aerodynamic drag rise with this increasing speed. At supersonic speeds, the closer you could fly to Vmo/Mmo the lower the fuel burn was. (Especiall true at Mach 2, although the autopilot would hold you Mach 2 (ish) in Max Cruise mode, flying closer to Mmo, Mach 2.04, would save fuel, assuming the static air temoerature was low enough to sustain this). This fact (along with about a million others) produced what we all like to call 'The Magic of Concorde'

Best Regards
Dude