Posts by Vic

Re: BUT...

Uh. Sorry, but a 777s flight controls are not computer assisted

Some most certainly are. I cannot tell you if all of them are. From the FCOM:

In the normal mode during manual flight, the ACEs receive pilot control inputs and send these signals to the three PFCs. The PFCs verify these signals and information from other airplane systems in order to compute control surface commands.

"ACE" is Actuator Control Electronics

"PFC" is Primary Flight Computer

that explanation is about as likely as truth from a politician.

Yup.

Vic.

Bah.

> the aircraft was clearly working its way North-East

North-West, of course. Sorry.

Vic.

If it was suicide or hijacking, why did they end up in the middle of the ocean so far away.

We don't know that they did; this is simply a *possible* sighting of something that could *possibly* have been aircraft debris in a location that MH370 could *possibly* have gotten to. And the first aircraft on-site has failed to find anything anyway.

I'm actually rather surprised they're looking down to the South; the aircraft was clearly working its way North-East before radar contact was lost. But then there is clearly quite a bit going on that we don't know about - for example, the NTSB tracks are somewhat detailed considering the paucity of data that's been released...

Vic.

Re: BUT...

why was the last ACARS transmission 12 minutes BEFORE the good night call to Malaysian ATC?

ACARS is generally scheduled. I don't know if the system would autonomously create more reports in the event if something going wrong - but the pilot has a button to push to create a new report on demand.

The fire theory doesn't work.

Not for me, it doesn't.

Also all Boeings (I think) have wheel well fire detection system and brake temperature monitoring

Not all 777s have nosewheel well fire monitorig - I've been reading the FCOM for some Qatar Airways aircraft, and they only have mainwheel well fire monitoring. I've no idea if MH370 had anything in the nosewheel well, but as I've said numerous times, even if there were a serious fire there that managed to incapacitate both pilots somehow, something would already have had to happen for the change of flightpath; this wasn't a simple divert (the FD would have taken the aircraft straight to the alternate).

Vic.

Re: Mobile phones don't have remotely enough range

Because pilots are trained to sort out the problem first and then communicate with ATC last. Aviate, navigate, communicate - in that order.

That's *generally* true, but pilots are also taught to get the Mayday call out when something goes wrong - SAR is very difficult if you don't know what you're looking for, as we've seen in this example.

The Mayday call is, IIRC, the only bit of the R/T exam that you *have* to get right first time or you will fail. Additionally, it's required during the PFL on the skills test, or you will fail. In an emergency situation, Commuicate trumps Navigate, and comes pretty close to Aviate.

The PIC on this aircraft had - what, 18,000 hours experience? He knew what he was doing. If he'd had the opportunity, he would have got the call or the squawk out. The aircraft clearly had no major handling issues, as we can see from the track it took.

Unfortunately the problem probably meant that the communication facilities stopped working as they turned around. So that's why there was no message from the pilots. They couldn't communicate.

One of the communication methods available to a pilot is to fly patterns - flying 2 left-hand triangles of 1 min per side would alert the radar controller that the radios had failed completely. We know the aircraft was capable of manoeuvring, so this procedure was available to the pilots, so long as they were conscious.

What about if they weren't conscious? That would prove that the FD had already been reprogrammed with waypoints that were not on the flight plan. So if this were the case, something nefarious had already happened.

Vic.

Re: The simplest explanation and confirmation bias

That incident tells us nothing about the ability of military radar but everything about the amount of misinformation that an embarrassed Navy / Government will spread in an effort to wriggle out of culpability.

And that, I'm afraid, is probably the most likely explanation for most of this incident.

Vic.

Re: My tuppence worth.

OK I was over dramatizing the dive bit, but you're not going to stay at 35000ft any longer than necessary and you are going to go for an expedited descent.

Sure - but if you come down at, say, 4000 ft/min, dropping 20000 ft is going to take you 5 minutes. If that descent is your plan to deal with the immediate decompression problem, you're all dead.

Alternatively, you get everyone on O2 masks, and it just gets uncomfortable.

I worked on the design of Airbus avionics, so can only generalise about the opposition

Everything I posted is straight out of the FCOM. It's readily available, if rather long...

I do know that aircraft systems aren't designed for major structural failures

Aircraft are designed to be as resilient as is feasible. That means you're never going to recover from a wing falling off - but the 777 can lose substantially all of its fly-by-wire kit and still be controllable.

What I'm arguing is that as an alternative to hijack or a mad pilot, MH370 may have suffered a major structural failure that destroyed many of the systems (and their backups, and the backup's backups), but didn't destroy the aircraft. I can imagine all sorts of ways that this could have happened.

I can't. The evidence released so far shows an aircraft that was following a set of waypoints other than those corresponding to its flight plan, yet was uncommunicative. Someone programmed those waypoints - either during the initial stages of the flight, or as it progressed.

Maybe a bomb* that turned a freight container into shrapnel which then sliced through the several cables looms

OK, let's imagine that. How did that slicing through cables lead to new waypoints being added to the FD? How come so much of the aircraft was still functional (such that it continued to fly for hours) yet no communications were possible?

The first of these is essentially impossible - the chances of shorting *exactly* the right cables so as to have the same effect as pressing the FD programming controls on the flight deck is mathematically infeasible. The second *could* have happened if both pilots were incapacitated behind the security door with the deadlocks in place - but that would imply that the waypoints were already stored; the "hijack" (for want of a better word) had already happened.

Maybe something similar to flight 243.

You'll note that Aloha flight 243 got a mayday call out. This is the primary similarity I would expect between these two instances.

The question is, is such a massive failure possible?

No it isn't. The question is, could such a massive failure take out all communications and also re-route the plane to somewhere it wasn't supposed to be going whilst not causing sufficient damage to take the aircraft out of the sky? ITYF the answer would be a resounding "no".

Now I don't know what happened to MH370. But I do know it wouldn't just wander off in the way it did without human intervention - someone decided to change its destination without telling anyone. And that is suspicious.

Vic.

Re: The simplest explanation and confirmation bias

I discount the military radar data because I am familiar with such radar data and I have a good idea what it does or doesn't show.

I doubt that.

I went to Boscombe Down last year. In the radar room, we watched the parachute plane from Old Sarum drop its parachutists. You could count them out...

I suggest you read the accounts of the incident where the U.S.S. Vincennes shot down an Iranian airliner by mistake

That had nothing to do with radar problems and everything to do with the belligerence of the commanding officer.

I'm not saying that it's definitely a mistake. Just that we don't know until it's verified very carefully

Quite a lot is already verified. Australia seems just to have released some confirmation about what has been said, but I missed the article on account of my missus having an uncanny knack of talking through every single headline ever.

Vic.

I've seen an under-inflated nose wheel tyre suggested as a possible cause

I've seen Space Aliens suggested as a possible cause. I don't believe that one, either.

Overheated during take-off, smouldered for ~80 minutes, then started burning with a vengeance giving off thick black toxic smoke at the front of the plane.

The aircraft was at 35,000 ft. That means the pressure is very low - 26kPa according to the online calculator I've just tried. That gives us a ppO2 of about 0.05 - a raging furnace looks good on Thunderbirds, but really isn't realistic.

But even if it had occurred, the pilots would have put on their masks and selected positive pressure, in accordance with the FCOM. At that point, there would have been a mayday call, a mayday squawk, or a radar-visible emergency pattern. The aircraft would have remained under pilot control.

Could smoke be so thick that pilots could not see the instruments or anything outside the plane? They have Oxygen masks and smoke hoods, but no use if they are blinded.

The aircraft would have already been under FD control. They would be unlikely to need instruments at that part of the flight.

Question for a pilot: can auto-pilots deal with a stall if the pilots don't supply any inputs?

Trivially.

Vic.

Re: Risk analysis needed

It's a risk analysis problem

Indeed it is.

What is the risk from an ACARS system which cannot be shut down in flight, but which is protected by a fuse and circuit breakers so it shuts itself down if it draws excessive current or gets too hot?

Small, but present.

What is the risk from a malicious pilot?

Very much smaller than the risk of fire from something that can't be isolated. And if you've got a fuse, your hypothetical malicious pilot can *still* disable the kit, with it being potentially rather harder to re-enable it than just turning the switch. So your "solution" solves nothing, but creates new problems in the process.

One thing to factor in is the much larger number of lives at risk on the ground from a suicide-pilot, compared to the fairly random crash location of a plane on fire that fails to reach an airport in time.

But if you've got a suicidal pilot, technical changes aren't going to make a blind bit of difference. The pilot can still point the plane at the deck, whatever radios are running. The aircraft is not rendered invisible by turning the transponder off - it just makes life a little harder for ATC.

Oh, and instead of meaningless pings, have the antenna transmit the plane's GPS coordinates even when ACARS is shut down!

If ACARS is shut down, nothing is transmitted. This aircraft only sent the satellite pings because ACARS wasn't *fully* shut down - the HF had been turned off, but the SATCOM left on. The reason the packets were empty is, apparently, because Malaysia Airlines hadn't paid for the satellite uplink, so the data packets couldn't transmit any real data.

Vic.

Re: My tuppence worth.

if there was a loss of cabin pressure they would have put the aircraft into a dive

They wouldn't.

If there was a problem with the cabin pressure, they'd have put everyone on oxygen and diverted to the nearest available airfield that could accomodate an aircraft of that size.

Putting an aircraft into a dive when you're already going hypoxic is a sure-fire way of destroying the ship; you're unlikely to control the speed particularly well, and losing 20,000 ft still takes quite a while. Hypoxia kills in minutes[1].

If the connections from the yokes to the flight computes were also lost by my hypothetical event, then they wouldn't have been able to do that. (The 777 is fly-by-wire with the computers controlling the hydraulics)

The 777 has three main surface control modes: Primary (with filght controllers taking input from the controls and actuating the flight surfaces hydraulically), Secondary (as Primary, but with reduced capability, and the flight surfaces actuated electrically), and Direct (Flight Computers are unused; the actuators are driven directly from the flight controls). There is also a cable back-up, but that is for maintaining straight and level flight whilst the flight crew get the electrics running to achieve one of the above modes.

Getting the electrics going is entirely feasible - there is the conventional two-bus AC layout with isolation relays and routing, and there are three primary generators and one ram-air turbine, and one of which could power the aircraft control systems. Two generators are required for full electrical capability, but the passengers can do without "Last Holiday" during a crisis.

Vic.

[1] Hypoxia causes loss of consciousness very quickly indeed - this is why the safety briefing tells you to make sure you've got your own mask on before helping others. If you're on O2, you stand a chance of saving someone's life by fitting their mask even after they've lost consciousness. If you've both passed out without a mask, you're both dead.

Re: The simplest explanation and confirmation bias

> The military radar data is nonsense.

[Citation needed]

Vic.

> I'll go for a fire in the electronics bay that caused the blackout

That's quite unlikely; the 777 has two main AC busses, powered from any of 3 main generators, with a ram-air turbine as a fourth option. Power can be routed from anywhere to anywhere.

In the event of an uncontrollable fire - to the extent that the offending components cannot be shut down and extinguished - it's very unlikely that the aircraft would have power for flight control or satcom. We can be pretty sure about the former (since the aircraft did turn, and the cable backups are only intended to maintain straight-and-level flight while the electrics are restarted, according to the FCOM), and we know the latter.

So an amount of the electrical system was alive. The 777 has many radios - typically 3 VHF, 2 HF and 2 SATCOM. In the event that there was a big fire, I think it unlikely that someone was able to turn the aircraft around, but couldn't get off a mayday call, a mayday squawk, or trigger the ELT beacon.

Vic.

Re: Mass murder and elaborate suicide...

> then he helped it along, perhaps by cracking the doors open

How's that achieved, then?

Vic.

Re: Tinfoil Hat?

Now what's the possibility that a some weird bug or posibly data corruption triggered a set of circumstances where the system shut down pilot control and switched everything irrevocably to auto

*Very* unlikely. On commercial aircraft, the pilot has veto over most systems. Only military jets (with unstable airframes) lock out the pilots from the flight control surfaces (fo obvious reasons); civilian kit may supply envelope shaping, but the pilot can override (and may be forced to - as in the AF447 disaster). FADEC is about the only full-authority system in play here - but as we can see from the long flight-time, this wasn't an engine failure...

Very far fetched I know - but impossible?

Yes. Impossible.

Vic.

Re: Facts

Amongst other things...

> Aircraft Communications Addressing and Reporting System (ARCAS)

ACARS.

> Pilots are not normally trained to disable these systems

*I* know how to disable ACARS on a 777, and I'm not ATPL. I've never even set foot on a 777 flight deck. It's a trivial matter - Google will show you how to do it.,

> A pilot changes the Transponder code to 7500 for Mayday

7500 is Hijack. It's 7700 for Mayday.

> Freescale worked on computer viruses & anti-virus software

Yeah, I've got a roll of tin foil here. It's only really intended for cooking, but you might want to take advantage of it...

Vic.

Re: Another interesting hypothesis

> I really have no clue.

Really?

Vic.

Re: So what is going on?

I suspect some significant amount of vegetation removal (not to mention repaving) would be necessary before the runways are again suitable for use.

I flew over Greenham Common the other day - it's still a monumental gash in the landscape. I don't imagine you'd need to remove any vegetation.

I've no idea if you can land a 777 on a broken runway without significant damage. But then if someone was planning on stealing this plane, they might have found a runway[1] that wasn't destroyed...

Vic.

[1] They exist, even in this country. The chart shows many "disused airfields", and most of them are still viable.

Re: So what is going on?

But I think we can now see that there is a much greater risk to the safety of both passengers and the rest of us on the ground, in allowing it to be disabled.

No, I can't see that at all.

The power switches are all on the flight deck behind a locked door. Only the flight crew and specific members of the cabin crew can get through that door. All crew know intimately that, whatever the threat they might be facing in the passenger cabin, opening that door to an attacker means they will all die.

So the risk from allowing a transponder to be disabled - and remember, this is only secondary surveillance; it doesn't prevent primary radar contact - is only from members of the aircrew. And if they are prepared to disable aircraft systems, they could just as easily pull the fuses.

Disabling the transponder and the ACARS is extremely suspicious, but it doesn't turn a 777 into a stealth jet. It does prevent the 7500 hijack code from being transmitted - which would cause the aircraft to be highlighted on every radar screen in range. But even in the event of a cockpit battle between the pilots, turning the transponder on again is just as easy as turning it off.

Something odd happened on that flight, and I can't help but think that one or more of the pilots was involved. But we're not going to find out until the CVR is found.

Vic.

Re: The Most Bonkers Explanation award goes to...

> That twitter feed is bizarre

I'd say it goes deeper than that; she appears to have a rahter tenuous grip on reality...

Vic.

Re: So what is going on?

How does one infer from the transmission that the satellite picked up that the search corridor is an arc stretching both north and south (geometry pedants you know what I mean) from the last known position?

The report on the telly says that the satellite can determine the angle to the transmitter, but has no range information, and there's only a single satellite involved - leading to an equi-angle arc on the surface.

Why is there a need to have the functionality to allow an identification beacon on a civil aircraft be switched off in flight?

The transponder must be controllable while the engines are running - to do otherwise is likely to make the SSR display unreadable in the vicinity of a busy airport.

You *could* make the argument that the transponder should be running when there is no weight on the undercarriage - but that adds a lot of complexity to the aircraft (i.e. makes it more prone to failure), and the only thing it really obviates is the pilot deliberately switching it off (which he could still do by way of pulling the fuse)

Given that there is one known received transmission, can its timestamp be used to identify likely blips in backround static that other radio-frequency sensitive systems on the planet might have recorded, to enable even approximate triangulation?

No. Aircraft radios are AM or SSB. You're not going to gather any information from the noise.

This whole affair is very saddening. There is a specific transponder code to notify ATC (silently) of hijack; the transponder being turned off rather strongly implies that someone on the flight deck knew that. Given the locked door, that itself rather implies pilot collusion :-(

Vic.

Re: Tinfoil hat

> US law enforcement are the only people on the planet who still think 'Lie Detectors' actually work

...And the audience on Jeremy Kyle, of course...

There was an excellent apocryphal story about an interrogation when the cops involved didn't have a polygraph, but they did have a photocopier. They had written the word "lie" on a piece of paper and put in into the machine. Every time the suspect answered a question, they'd push the "copy" button. He cracked...

Vic.

Re: We need something more simple than webbrowsers

> I mean we (normal people) are not using webmail since it's far to insecure,

I do...

> Imagine we had some trivial "GUI over IP" protocol

ssh already does X forwarding. Has done for years.

It's very useful - but generally rather slow. Most users will not want to use it.

There are also security issues to consider - do you reallly want to send all your keystrokes in real-time to a server you don't control?

Vic.

Re: Do I actually have to talk to anyone...

> Communicate with the beer bringer

With the what?

Vic.

Re: My town has it's own brewery

So does mine - the Dancing Man brewery.

I think their beers are gorgeous, but they tend towards the hoppy, so probably not to everyone's taste.

Vic.

Re: London's a pain in the arse to get to...

> I've not heard of butchers hook

It's only just opened. I went there last night and was suitably impressed by the beers.

SO18 1NN, if you're interested.

Vic.

Re: Oh dear.

Vic, your last comment is the important one: "the difference is almost always irrelevant in power stages".

Of course. That's why I made it...

I'm sure you're aware that your claim that class A stages are entirely linear isn't accurate: no audio output stage is purely linear.

Class A stages are very, very close to linear - close enough to say that they are indeed so. Class B is nothing like as close.

But the main point here is that no audio output stage is as perfect as it ought to be

No, the main point is that the imperfections very rarely matter. Distortion added in the speakers/headphones and in the room will dwarf the distortion from the output stages.

But why? Resolution and sampling frequency can easily be increased to stupid levels (from the audio point of view). However, in doing so linearity and other aspects of fidelity might well be compromised.

No, linearity will not be compromised to a greater extent than losing that resoution provided your ADC is monotonic. And if it isn't, you've got far bigger problems than word length and sampling rate.

The reason for using as much resolution as possible is that noise always increases - so if you're post-processing (e.g. mixing) two signals with uncorrelated noise x, you'd expect a resultant noise level of sqrt(2) * x simply from the combined noise level. So you minimise your recorded noise (i.e. the value of x above) and keep your interim value quantisation noise to the minimum. You can always shed resolution later (at mix-down); you cannot create it from nowhere.

It's the same reason an old-style studio would master to 2-inch tape, not compact cassette.

Surely the thing to consider is how it sounds?

No. If your recording equipment colours your sound, you've already got fidelity problems. The recording/mixing/processing stages should be as transparent as possible. This means generating as little noise as possible, and in the digital domain, that means high sample rates and high resolution.

If a given combination of resolution and sampling frequency produce results audibly indistinguishable from the best available analogue technology, maybe the thing to do is improve linearity and noise (etc) before bothering with top-line buzz-numbers?

Why? Analogue recording technology pushed the capabilities to the limit to achieve acceptable performance. Digital techniques allow *much* higher fidelity at negligible cost; it would be foolish to restrict those capabilities just because the previous technology couldn't match them.

24 bit for mixing and mastering makes good sense if you look at it, but better than 16 bit for final distribution and reproduction?

I wasn't talking about the final, down-mixed product - if you look at my post, you'll see I said

capture, processing and mix-down should be performed at the highest resolution/sampling frequency available. But once mastered, CD-quality is usually plenty good enough as long as you've done a completent job in production

Vic.

Re: Oh dear.

I've read convincing articles showing that actually, class B amplifiers (when properly implemented) provide the best practical audio output quality

The articles might have been convincing, but they weren't accurate. A properly-implemented class A amplifier will always have higher fidelity than a properly-implemented class B amplifier because it doesn't have the transition at the zero crossing; class B amps aren't entirely linear, whereas a class A is (or at least, it should be).

That said, the difference is almost always irrelevant in power stages.

analogue to digital (for processing, storage, and distribution) and back again to analogue (for playback) is the best way to record and reproduce music - provided that you perform all steps thoroughly competently.

Absolutely so. Digitisation introduces a small amount of noise due to quantisation, and any post-processing will introduce a little more for exactly the same reason - but the total noise is *dramatically* lower than that added by an analogue storage/post-processing environment.

But I always cringe when I see CDs with "digitally remastered" on the cover; it inevitably means "blown out by some YTS monkey who doesn't give a shit".

This is how come Decca managed to make some really rather excellent audio recordings back in the 1950s

The biggest problem with such recordings is the durability of the media. My mate has a huge stash of recordings[1] predominantly from the '70s and '80s. On the quiet bits, you can hear the print-through...

24 bit/44.1kHz in the studio is indistinguishable from analogue tape for mixing and mastering.

Sounds reasonable.

There's little evidence that higher sampling rates than CD standard produce audible benefits, at least not when properly competent A/D and D/A conversion is involved

Don't agree there, though - capture, processing and mix-down should be performed at the highest resolution/sampling frequency available. But once mastered, CD-quality is usually plenty good enough as long as you've done a completent job in production. Such a task is much rarer than it ought to be - but isn't improved by better kit...

VIc.

[1] I made some of them...

Re: The Tintinnabulations Of The Ad-Copy.

> the more you pay, the better it sounds.

A mate of mine used to use the tag-line "the more you drink, the better I sound" :-)

Vic.

Re: Another one?

There is some software that captures vinyl at 192 khz before the pre-amp, and then applies the RCA curve on playback, but it's OSX only.

And there is software that does that that isn't OSX-only.

JAMin is GPL and multi-platform. And has a RIAA filter plugin shipped with it.

Vic.

Re: Golden ear set again

There's no snake-oil about Class A amplifiers... it's still the best way to design an electronic amplifier if all you care about is output quality.

Absolutely so. But I very rarely come across anyone who cares enough about the improved quality over class B amps to pay for the electricity...

Class D might well be the way to go in the future - there is necessarily some distortion added, but the idea of digital-to-the-output-stage is very appealing.

The old "modern music is so distorted anyway" has been made since the days of Motown, and it's still not true

Perhaps. But the modern "production" standards of a big smile curve and <2dB dynamic range still makes everything sound like shite, though.

Vic.

Re: It's bad procedure to take off using your backup systems

> GPS does not indicate airspeed

No, it indicates groundspeed.

> it's airspeed that generaties the lift to keep the aircraft in the air.

This is true.

But there's a fairly simple delta between groundspeed and airspeed, and that is W/V. And if you're landing anywhere with A/G or better - as you would be in a jet - you'll be given that on final, if not before. ATSU is always helpful...

> If airspeed indication is unreliable pilots fall back to flying on known pitch and power settings

Not on final, we don't. PAPIs, VSI, eyeballs are all good. Trying to land on pitch/power leads to a crash, a bounce or a "CFIT"[1].

Vic.

[Yes, I have passed now]

[1] Without the "C"...