Malaysian Airlines MH370 contact lost

[Vinnie Boombatz]

Some Graphics in a BBC Story

BBC News - Search for MH370 switches underwater

(Image removed -- click on above link to see source)

[olasek]

.And there is no reason to doubt the Chinese in this matter, unless you (or someone) dislike Chinese and are allowing that to cloud their judgement

I do doubt them very much, they don't tell what kind of technology they are using, allegedly they use some 'hand held' devices, one sound engineer familiar with the technology to search for ping signal burst in laughter when found out about it. Anyway, they show zero transparency. I doubt whatever they heard had anything to do with the signal source picked up by Ausies. Chinese may be too eager to feed their domestic propaganda market - this could cloud their judgement so there are very god reasons to doubt them.

[mm43]

Originally Posted by JamesCam

Anyway, even a decent LC oscillator will have better frequency stability over varying supply voltages than the performance described in the OP's post!

I'm sure it does, though the spec would indicate that stability is not a huge problem, but rather the lack of stability and frequency accuracy will help in identifying more than one ULB in the area. How about an oceanic mid-air! Potentially you may need to identify 4 ULBs.

[porterhouse]

If there is one thing the Chinese cannot afford is to do is to release some info/data which proves to be false later on.

I don't think so, they already released quite a few info that proved to be false, news is changing so fast here nobody remembers what/who claimed a few days ago. Plus they can always spin it to their advantage - we were here, we participated, etc. it is mostly about how looks on their domestic market.

[hamster3null]

Originally Posted by JamesCam

I find it extraordinary that the ping frequency apparently changes so dramatically with power supply voltage! Surely the device is crystal controlled? what use is a device like this if the frequency one is looking at is effectively unknown. To detect the signal at a distance one needs to use very narrow bandwidth filters of the order of only a few Hertz: here we are talking about a shift of 4.2KHz, well over 10%...

I don't believe this equipment performs this badly, given it's designed role. was the Ocean Shield following something else? can someone quote the source of the "manufacturer's" comments?

James

It's more of a "crystal" than "crystal-controlled", I think. It's an IC generating short square-wave "chirps", ~10 ms long, modulated with a 37.5 kHz wave, which then drives a piezoelectric crystal that's cut with the same resonance frequency. The whole device is very simple, can fit in the palm of your hand and weighs 200 grams including battery and casing. Narrow bandwidth is not a high priority target. Length of the chirp by itself spreads the spectrum by 100 Hz. Vibration response of the crystal to such a short chirp may be pretty complicated too. ULB manufacturers only guarantee frequencies up to +/-1 kHz, even for a battery at full charge.

My understanding is that underwater sounds can reflect back off the bottom and / or thermal layers in unpredictable ways. There are tales of submarines being tracked from more than a thousand miles away under the right acoustic conditions. I have no problem imagining that a pinger's signal could turn up three hundred miles away. Also, the only way I could imagine parts of the plane coming down 300 miles apart is if it was high up on re-entry al la space shuttle. I also can't imagine any fish making a love call on 37khz or so with a pulse repetition frequency similar to an underwater beacon.

Reflection can happen, but attenuation is attenuation and you can't get around that. At 37 kHz in seawater, it's ~5 db/km. A pinger that starts off at 160 db will be certainly below ambient noise even 30 miles away.

Submarines are a different story. Attenuation is orders of magnitude lower at lower frequencies.

Any repetitive source of white noise with the right repetition period would sound like a pinger to a narrowly-tuned pinger detector device.

[amizaur]

My understanding is that underwater sounds can reflect back off the bottom and / or thermal layers in unpredictable ways. There are tales of submarines being tracked from more than a thousand miles away under the right acoustic conditions. I have no problem imagining that a pinger's signal could turn up three hundred miles away.

No, because the signal from the emergency beacon pinger is VERY DIFFERENT from those emmited by ships, submarines and singing whales, that can be heard hundred of miles away.


The sounds that can be heard hundreds of miles away - because of phenomena called ducting - are low frequency sounds, that can travel in water great distances with minimal loss of energy. Such sounds can be detecte from long distances in similar way like astronomers detect very weak light of distant stars or galactics - by using very large reflector and focusing the energy from large area to very small area (for example using long towed sonars or finding a natural sound convergence zone).


The emergency pinger generates a high frequency sound, which is highly attenuated in water, and it loses it's energy relatively quickly with distance.
So at distance hundred of miles it can't be detected even using a large reflector or antenna or sound convergence zone - because there is nothing to detect.


An analogy - a large astronomical telescope can be used to detect very, very weak energy from a distant star (too weak to be registered by a CCD detector). The light is really very, very weak but IT IS THERE, only "very diluted". If you are smart enough, you can build a device that would amplify it to detectable level. So you build a telescope with a big mirror, collect the weak energy from large area and focus it on a small area - so it's again detectable by your CCD detector.


Trying to detect 37.5kHz pinger from 100nm by any sensor, would be like trying to use an astronomical telescope to detect a light from a street lamp that is only 10nm away - but in a thick fog. The telescope is big, but it won'd detect a light from the lamp because it was attenuated by the fog to practically zero level and there is simply no light that can be focused, period.


It doesn't matter how big your telescope/towed array is, if you are 100nm away you can't amplify a signal that was attenuated to zero level before it reached 10nm. It doesn't matter that there is a "natural water telescope" named convergence zone - it would amplify sounds and you would hear even wery weak signals of low or medium frequency, but you won't hear an ultrasonic pinger because at such distance it;s sound energy is not "diluted" - it doesn't exist anymore.

[Lemain]

Lovely graphics. However over the years I have found that really nice slick presentations provide a credibility that is far greater than the underlying science and data are worthy of.

[wes_wall]

Before anybody gets all excited over this claim, or the other, remember their is an awful lot of noise under the water, and that is what the game is all about when it comes to underwater boats. Those who can hear and decipher what they hear, and those which only record noise. Perhaps those who know can comment on the current realities being faced with the recovery teams.

[RR_NDB]

37.5 KHz sources

mm43,

How about an oceanic mid-air! Potentially you may need to identify 4 ULBs.

From an "weak signals" reception using DSP, a tolerance of + or - 5% could be OK. Even to identify multiple sources as you commented. Weak signals reception is a fascinating task now with the current techniques available. And this LF signals are very easy to process.

[Astrax]

33.3 KHz rather than 37.5 KHz

It is well known that a transducer changes the resonance frequency under pressure. Without knowing the ULBs construction or other reasons, I can imagine that the design goal is to get the highest signal rather than keeping the frequency on 37.5Khz. The ULB may be pretty simple (kiss) and only kick the transducer at right timing and keep it running for a predefined time at the current transducer resonance. The frequency is not as important than getting the best signal strength.

[Johnyy]

Originally Posted by Amizaur

The emergency pinger generates a high frequency sound, which is highly attenuated in water, and it loses it's energy relatively quickly with distance.

Perhaps then the emergency beacons attached to 'black boxes' should also emit a low frequency emission that can be detected at greater distances under water. Or would that take too much power

[RR_NDB]

Weak signals reception field

Perhaps those who know can comment on the current realities being faced with the recovery teams.

I am familiar with this field in the last 10 years or so. (LF, MF and HF) The capabilities of current gear are unbelievable. I have a "good ear"for CW (morse code) under very noisy conditions. When you use a "quite simple" DSP software in your computer audio adapter you can see clearly signals impossible, repeat, impossible to hear. Remember noise is inherently random but the ULB signal source is coherent. Fascinating field! But the recovery teams are in a different environment...

[LCH77]

Could the Chinese have heard their own pinger?

In a video of the Haixun 01 on CNN, it appears the Chinese had a spare pinger in the RIB.
Anish Patel, president of pinger manufacturer Dukane Seacom, says it is not recommended to have a pinger near the area where you are trying to listen.
If that pinger gets wet, it will start transmitting, potentially confusing search teams.
"I wouldn't put one where I'm measuring," Patel said. "It's just not good common practice."

[Mesoman]

Modern Signal Processing is not magic

When you use a "quite simple" DSP software in your computer audio adapter you can see clearly signals impossible, repeat, impossible to hear.

This is true, and quite amazing results have been produced. But... you cannot repeal the laws of physics.

DSP can improve the effective signal-to-noise ratio. For some kinds of signals (coherent ones), it can do this by a lot. The same principle that allows undersea submarines (shallow) to hear ELF radio waves is at work: integration of the signals over time. But... a submerged sub may have to integrate a (radio) signal for maybe one minute for every bit of information - a data rate of less than .1 bps! This same technique can be used in sonar - a maritime target like a sub hopefully has some relatively continuous signals (prop cavitation, or a pump resonance) which can be integrated for a long time.

However.... integration has the following limitation: signals decrease exponentially with distance, while integration only improves the SNR linearly.

This means that if the signal has gone down by 100 (which is only 20 decibels - dB), you have to integrate 100 times as long. With this signal going to 5dB per kilometer ( as opposed to low frequencies which go down at .001 dB per kilometer), it's clear that, while DSP can extend the range, it can only extend it by a few kliometers.

It helps if the signal is designed to be integrated, which these pingers are not.

There are other tricks in the DSP arsenal, but there are ultimate limits. I could see the "specified" range exceeded by a few to maybe 10 km.

A nice calculator for sound propagation attenuation is at Calculation of absorption of sound in seawater.

[Vinnie Boombatz]

Underwater Acoustics Handbook

A nice handbook, though the last half on marine mammals is of less relevance here:

http://oalib.hlsresearch.com/PocketBook%203rd%20ed.pdf

The ULB frequency of 37.5 kHz falls on the right side of Figures 4 and 5, supporting other posts saying that the noise environment was a driver in the ULB frequency choice.

[paxrune]

Any word on when Angus's next press conference will be?

[TWT]

Don't know when the next PC will be but just checked their website,nothing new so far today as of 0927 Perth time,except HMS Echo joining Haixun 01 at the southern end.

Search and recovery continues for Malaysian flight MH370

[BillS]

The paper I linked earlier: Flight Recorder Localization Following at-Sea Plane Crashes
has two useful charts:
Fig 2.11 (p23) shows attenuation against frequency.
Fig 2.13 (p27) shows background noise against frequency.
They show that 37.5kHz is well chosen for background noise, if a little high to balance for attenuation.

[onetrack]

There have been reports that there's another press conference early this afternoon, AEST, on further MH370 developments.
The news item said the Defence Minister was holding the conference. I think that was a typical media stuff-up, they don't even know the difference between the JACC leader and the Defence Minister.

The Ocean Shield has recorded two separate pings, on different frequencies, that appear to coincide with the signals from a CVR and FDR.

It's starting to appear that the transit of Cyclone Gillian on 22-23 March, which hit Christmas Island, then went through the region of the current search area, has seriously diminished the chances of finding floating MH370 wreckage.
In this region, the wreckage would not have been pushed towards the West Australian coast, it would have been circulating in a current eddy in the region or be pushed towards Christmas Island.
When HMAS Sydney was sunk around 130 NM SW of Carnarvon, Western Australia, in November 1941, a Carley float from the Sydney, containing the remnants of a corpse, was washed up on Christmas Island some 3 mths later.

[tdracer]

OK, assuming for a minute that these pings pan out, could someone with real knowledge of the subject give a rough number for how long it will take to:


1) Triangulate the location to a limited area;
2) Bring in side scan sonar to survey the area;
3) Do the survey and identify likely targets;
4) Bring in appropriate deep sea submersibles (manned or remote);
5) Go take a look and definitively say this is it and start actual recovery?


I'd like to think the necessary side scan and submersible resources have already been tasked to the general area, but maybe I'm being too optimistic. So what time period are we looking at - a week, two weeks, a month? Longer?