Designing an Acoustic Modem for an AUV

[Gepard]

Hi,

I'm making an AUV and need to design my own acoustic modem. Now, after hours of research I found these PDFs (attached).

In theory, can I remove the signal generator from XXXXPinger.pdf, add a DTMF encoder and then use a build the transducer as described in bender2.pdf to transmit the sound? Then use XXXXReciever.pdf with a bender2.pdf transducer to receiver the sound then put it through a DTMF decoder?

It can't really be that simple can it?

Thanks

Michael

 

[Averagesupernova]

You'd be best starting over completely. The transmitter link you have is not really suitable to send DTMF. The output stage is not at all linear. Why do you need DTMF? Can't you just go with something simpler like AFSK?
-
At any rate, the whole project sounds kind of interesting.

 

[Gepard]

Thanks,

No, I don't need to use DTMF, and having just looked up AFSK (thanks Wikipedia :D ) I think that's a much better method.

If I try to do it as I said in my first post, what problems would I get?

Am I right in thinking piezoelectric discs for transducers?

Michael

 

[Averagesupernova]

DTMF is an acronym for Dual Tone Multi Frequency which I'm sure you already knew. The frequencies are SINE waves. The schematic you posted does not have an amplifier capable of driving the transducer without distorting the sine waves.
-
I cannot say whether any of what you are doing is right without knowing more about the project. All I know is you have something under water and need to send data either to or from (or both ways for that matter) this thing that is under water.

 

[Gepard]

I meant if I use AFSK with the first method would it work...

The project needs a duplex comms link between the surface and AUV in order to program it in the water and relay sensor information back to the surface. It needs to have 32 on/off commands and 8 values 9 (0-255) and work over 100m max.

Any thing else you need to know?

Michaell

 

[berkeman]

What data rate are you shooting for? Keep in mind that for reliable comm in real-world channels, you will need to do things like send a preamble to start each packet and include some level of error detection (like a CRC) or correction.

 

[Gepard]

I'd say <50bps, I just want to be able to get something working that I can improve upon as time progresses. Not looking for the ideal solution first time round.

Michael

 

[berkeman]

I only skimmed the documents that you attached. Where does the 200kHz number come from? Is that optimal for underwater comm?

Also, for the best chance at good communication, you will want to use a correlator in your receiver. Have you learned about correlators yet? How about error detection and correction algorithms? You will get extra credit in your project if you use real-world communication techniques like these.

 

[Gepard]

I start electronics classes in September, everything I know at the moment is self-taught. Although I'm very keen to see this project through as lot (both time and money) has been invested in it. Not looking for a detailed what-to-do, just areas I should look into/ investigate etc.

Other forums I've been to suggested 40KHz but purely because they designed their own system to work with that (they did it 20 years ago and only remember it was 40KHz and used a variety of op-amps in the receiver).

Navy submarines use somewhere in the region of 3-4KHz i believe for their comms, although they have to go a lot further, so I guess that that is the optimal frequency.

The site I pulled those schematics of ( http://www.dstl.gov.uk/news_events/competitions/sauce/index.php ) don't make any comment as to why 200KHz was used as you can see for yourself.

Michael

 

[berkeman]

I'd suggest doing a little looking around the Internet to see if you can find out what the optimum short-range acoustic comm frequency is, and if there is a preferred modulation scheme. You will want to avoid frequency ranges where there is a lot of natural (and human-made) noise, for example. If you look at the spectra of typical noise sources in the ocean and from ships, then that may help point you to a range of quieter frequencies. And the correlator will help you reject noise as well, if you choose your modulation scheme carefully.

Working on projects like these will help you go a long way in your upcoming electronics classes. What signal processing books and resources have you been studying from?

 

[berkeman]

BTW, I googled correlator signal processing tutorial, and got some good hits:

http://www.google.com/search?hl=en&q=correlator+signal+processing+tutorial

 

[Gepard]

Only book I've looked at Physical Computing, a run though of circuit basics. And everything I've learned for this topic has been in the last 24hours staying up well into the early hours reading all sorts of PDFs and reading through old forum posts etc. It was this thread (https://www.physicsforums.com/showthread.php?t=119489 ) that lead me to this forum...

Your welcome to suggest books as the more I delve into the subject the more fascinated I get by it.

Michael

 

[Gepard]

Thanks for the link

Michael

 

[Gepard]

Oh, one point I ought to make is that the AUV will be run in a fresh water environment (local pond) and so traffic in the area will be minimal.

Michael

 

[berkeman]

One basic electronics book that I really like is "The Art of Electronics" by Horowitz and Hill. You can get it through Amazon.com, or at your local university bookstore. It takes you from the basics up through opamps and some digital circuitry.

BTW, you mentioned that the link will be duplex, but keep in mind that the packets that you send in each direction will have to be separated in time. I don't think you will be able to talk in both directions at exactly the same time, since you would not be able to hear over your own transmission. So you will need to work out a protocol based on packets. Since you have a specialized situation of one boat and one UAV, the addressing scheme can be simplified. But you will still need to come up with the format of the packets, including the preamble, the data unit and the CRC. You will also need to come up with a scheme for getting acknowledgments back for packets, and for doing retries if no ACK is received within some timeout period.

Fun stuff.

 

[Gepard]

Thanks for the link, I'll have a look,

I'm going to setup a tank in the back garden tomorrow. 10m long made up of drainage tubes and probably some sort of lining to adsorb some waves. I've got a piezo siren which I'll take apart and build an preamp to add to the rx piezo disc and see what happens. Just out of interest, don't know what will happen.

Been doing research for long enough and want to build up some 'experience', I'll post here with my results.

Michael

 

[berkeman]

BTW, I think you saw our warnings in the other thread that you referred to, but think about some hearing protection while you work with the ultrasonic power stuff. I don't know what-all precautions you should follow, though, so maybe google a little to see if just regular ear-muff protection works for ultrasound. Even though you won't be in the water, there may be safety issues. Come to think of it, that must be a general consideration for UAVs using ultrasonic communication. I wonder what kind of diver safety precautions should be taken at various power levels... We didn't talk much about that in my SCUBA certification classes...

 

[berkeman]

Wow, I just googled ultrasound sonar +diver +safety, and got lots of interesting hits. Maybe check them out as part of your project:

http://www.google.com/search?hl=en&lr=&q=ultrasonic+sonar++diver++safety+&btnG=Search

 

[Averagesupernova]

I'm guessing that 200 Khz is best suited for sonar. It may not necessarily be suitable for communications and it may be somewhat directional in the water. How many dollars are you willing to drop on this project? I'm also wondering what you are going with for a microprocessor/microcontroller. Have you looked into some of the equipment used in amateur radio? One of the protocols used is AX25. I think it may be what you are looking for.

 

[Gepard]

Well been and picked up the tubes, going out later to get some piezo transducers..

At the moment I'm just going to try submerging a piezo siren that's designed for air, no plans to go ultrasonic just yet, want to build the RX and see how well that performs.

In terms of money I'd say $200 (on electronic components) initially to get some results, then if it looks promising I'm happy to spend more on it.

At the moment I've got some BS2 and PICAXE programming experience, although if I have to use a different chip then I'm happy to learn.

I haven't looked into amateur radio as I didn't think it would suitable, I just want to see how well I can transmit/receive sounds first and then I'll start think about transfer data.

Michael

 

[Averagesupernova]

I'd stay with the PIC stuff. You are mistaken if you think there isn't some valuable info in the amateur world. I'm not talking about the radio part of it specifically. If you want some more info, send me a PM.

 

[Gepard]

Right, about to go to bed (late here in the UK) the tubes are lined up, horizontal (about 5 degrees out which I'll change tomorrow, capped both ends, and filled with water. As my garden is on a slope I had to use wooden blocks to raise one end. I've also cut a hole in the top at both ends using an angle grinder to allow me to drop the piezo transducers in. On the electrical side, I've got a working 'output', sounds like siren and uses a 556. I also picked up a handful of different size piezo transducers to experiment and once these are wired to a preamp you can hear the sounds on headphones, meaning I should be able to record the sund on my computer... Almost sounds like I know what I'm talking about :D

Tomorrow I'll seal the transducers and plop them in the water and see what I get..

Michael

 

[Gepard]

Wow, I'm impressed by the RX. Here's the result: http://www.mhims.co.uk/Sub/acoustic%20recording.wav (Sorry about the end :D )

I tried to use a loudspeaker in a water box to produce some sounds and see if they could be picked up at the other end however they were just getting stopped by the acrylic walls, so to solve this, I've ordered a waterproof speaker to use for transmitting.

Once it's come and installed (due to it's size I have to drain the tube), I'll try again and see what I get. Any preferences for the music? :D

Once I've got a system that can transmit and receive well enough I'll start implementing protocols etc.

Michael

 

[Elektrotechniker]

Does this need to be acoustic? Why not use radio?

 

[berkeman]

Does this need to be acoustic? Why not use radio?

In the general case of salt water, the water's conductivity attenuates RF signals very quickly. Only very long wave signals can be used (like the military uses with submarines), but you need super-long antennas to pick that up.

In fresh water, ,maybe he could get away with RF. I don't know the typical conductivity characteristics of pond water...

 

[Averagesupernova]

When it comes to natural bodies of water, I'd say it all attenuates bad enough to be useless.

 

[Gepard]

Does this need to be acoustic? Why not use radio?

Already considered it, average radio control submarines can go 10m max in fresh water, in salt water, barely past the surface. I'm looking for 15m+

Still waiting on the speaker to arrive...

Michael

 

[dtdakin]

dtdakin

I've just discovered your thread and may be able to offer a little insight. My background is antisubmarine warefare and underwater sound velocity. I'm no expert on transducer design but here is my two cents worth anyway.

There are a few things you should consider for your project:
1. Your and Berkeman's comments about radio not being viable except at low frequencies are correct.
2. The lower the acoustic frequency the greater the range for a given power. See attachment. For 30 m you could go as high as 500kHz
3. The lowest ambient noise in the ocean occurs at 40kHz with 0 wind, 80 kHz with 5 knot winds, just over 100 kHz with 10 knot winds. This is due to wave noise. With 5 knots of wind ambient noise is below 30 db re 1uPa from 20 kHz to about 150 kHz. 200 kHz would be about 33 dB re 1 uPa regardless of wind speed, it is governed by thermal noise at that frequency.
4 The beam width (like a flashlight beam) of power coming from your transducer is depedent on the dimensions of your transducer and the frequency of operation. A rule of thumb is the beam width from normal to the transducer plate ( Ie centre of the beam) is BW (in degrees)=3600/Freq(in kHz)/Diameter (in inches) or =91440/Freq (in kHz)/Diameter (in mm).
This beam width is the half power or -3 dB point, and the diameter is for a circular plate transducer. The beam width applies to both transmitting and receiving. If you want omni-directional response you will have to use a small transducer face or a tube transducer.
5. The size of the transducer and the frequency govern how much power you can put into the water. There are two issues here. First: At 20 kHz the transducer starts to cavitate at ~1 W/cm^2, 100 kHz at ~30 W/cm^2, 200kHz ~100 W/cm^2. Second: the piezo will start to depolarize when your driving voltage gets above about 200 V/mm of piezo thickness.
6. The receive transducer should have a pre-amplifier right next to the transducer to drive the cable since a piezo does not generate much current.
7. Piezos operate much more efficiently at resonance for transmitting and at anti resonance for receiving. For better broadband linearity it is best to operate the receiver piezo well below resonance but this sacrifices sensitivity.
8. When testing acoustics in water you need to consider the wavelength in the water. Wavelength=Sound velocity/frequency. Ie. at 100 kHz the wavelength is 15 mm. The short wavelengths due to the high sound velocity in water (~1500 m/s) means that you will get constructive and destructive interference from the walls of your test tank. This generally means you need to keep the pulse trains very short during tank testing.

Also check out the transducer manufacturer's web sites for technical info. Channel, ITC, APC, Sensor Tech and others all have lots of info. A good place to start would be APC's website http://www.americanpiezo.com/

Have fun.
Tom Dakin
dakin@shaw.ca

 

[berkeman]

Outstanding info, Tom. Thanks, and welcome to PF. Say, can you say anything about the diver safety angle? Is there some rule of thumb for diver exposure at different frequencies?

 

[Gepard]

Wow thanks Tom,

I'll look over what you gave me and contact you if I have any problems if that's alright?

Michael

 

[dtdakin]

Berkeman,
I'm a very active diver and I can't say I've ever heard about anyone ever having trouble with depth sounders on boats which normally operate at 50, 60, 98, and 200 kHz. There is an organization called DAN, the Divers Alert Network which is free to contact and will answer diver medical related questions. You can contact them on the web at www.diversalertnetwork.org .

Michael,
Sure feel free to drop me a line. I can be reached at home at dakin@shaw.ca or at work at tom@aml.bc.ca. I'm on Vancouver island in Canada so if you are in the UK there is an 8 hour difference.

I'm in the process of designing a coded underwater comm system for divers so I'll be looking into similar stuff as well.

Tom

 

[tongchgen]

The circuit is right ?and it can work correctly ?