Searching \ for '[EE]: Phased arrays, ultrasound etc.' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/io/audio.htm?key=sound
Search entire site for: 'Phased arrays, ultrasound etc.'.

Exact match. Not showing close matches.
PICList Thread
'[EE]: Phased arrays, ultrasound etc.'
2002\06\03@152513 by Brandon Fosdick

flavicon
face
I'm looking for a new project and phased array antennas have always intriqued
me. RF is a bit of a black art for me so I'm thinking about playing with sound
instead. I have no idea what I'm going to make, this is still in the thinking
stage. Maybe a locater beacon/tracker for model rockets.

I understand the basics of phased arrays, but I'm not too clear on the details.
Both google and amazon have lots of info on RF arrays, how much of that is
applicable for sound? Since both sound and EM radiation are waves I would expect
to find a lot of similarities. Some of the RF books on amazon are a bit
expensive, and I'd hate to spend a lot of money on a book shelf ornament. Any
recommendations?

I did a quick Digikey search for microphones and everything that came up was
limited to audible frequencies (<20kHz). Where do I find a good source of high
frequency microphones/speakers?

What frequency range should I be looking at? I have pets so I'd like to go high
enough to avoid annoying them if possible. Any idea what range medical
ultrasound devices use?

I imagine I'm going to need some high speed ADC's. I've looked over the
corresponding parts from TI and Maxim (any other manufacturers I should look
at?) and there's a wide variety to choose from. What are ECL interface levels?
I've never designed a digital circuit over 20MHz, any gotchas that I need to
know about?

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\06\03@154555 by Goring, Steve

flavicon
face
You could help me out Brandon ...

I am developing a small ROV and the next bit involves some
underwater sonar ( freq range 40 - 500 khz ).

Higher the freq, more direction less range.

Does that sound like something that would interest ??


Steve



{Original Message removed}

2002\06\03@171950 by Peter Meleschko

flavicon
face
Hi,

I don't know if I can be of any help but I'm working on a side scan
sonar (250KHz) myself. I guess that's in the same direction as you?

Peter

man, 2002-06-03 kl. 21:38 skrev Goring, Steve:
{Quote hidden}

> {Original Message removed}

2002\06\03@180217 by Doug Butler

picon face
Some good books on underwater sound are those by Robert J. Urick.

Sound propogates great in water and wavelenghts can be on the order of a
centimeter so machining tollerances are in the amateur range.  Underwater
transducers can be expensive or require odd materials, but you can do a lot
with fish-finder transducers.

As for phased arrays... Waves is Waves!  I read microwave and optical
journals, and I do some optical work as well as my usual sonar work.

Also look at the field of Non Destructive Testing (NDT) and Ultrasonic
Testing (UT) as they send sound waves through steel with great precision,
looking for cracks, bad welds, etc.

Doug Butler
Senior Engineer
Sherpa Engineering


> {Original Message removed}

2002\06\03@224910 by Tal Dayan

flavicon
face
As somebody else mentioned, a wave is a wave is a wave and the low frequency
of ultrasound makes it
much easier to work with and do neat stuff.

I presume that you are looking for an active sonar (both TX and RX) so one
thing you need to decide is are you going to form the phase array for both
TX and RX or only for RX. The RX side is easier since it does not involved
power and you can do everything with DSP. For example, some sonar systems
have at the very front an analog multiplexer that sample the hydrophones
signals (after the preamps) in a very high rate. This is then fed to a
digital processing unit that does the beam forming in all directions at the
same time using a single arithmetic unit. This way you don't need to have to
rotate the transducer array.

The beam forming itself is pretty simple, you take a bunch of hydrophone
signals, delay them proportionally to match the additional wave path in the
desired direction and then some them using predefined weights. The delay is
set such that when a signal arrives from the center direction of the beam,
the signal after the delays have 100% correlation. If you have the geometry
of the array it is pretty easy to calculate the gain of the bean in various
directions.

As for transmitting, the easiest is to send an omni directional signal. It
makes the power circuits simpler.

One of the things that you need to make sure is that you have a good
impedance between your circuits and the hydrophones and make sure to test it
in water since the impedance (if I recal correctly) changes from air to
water.

BTW, there is a cool sonar technique (don't know how it calls) that uses a
continious TX signal and provides a faster update rate. It transmitted
continuously an FM signal modulated by a slow saw tooth signal and
determined the distance of the target by the frequency difference between
the TX and RX signals. This allow to transmit continuously and provided
faster update rate. I don't know how this technique is called but probably
it has a name.

Another interesting 'sonar' systmes is a system that measures the sound
speed in the water. It has two small tranducers (one TX and one RX) and it
measure the time it take the signal to travel betwene them (you can use one
or two small 'mirrors' to extend the effective distance). The speed is
typically affected by physical properties of the water such as amount of
salt and temperture.


Tal

> {Original Message removed}

2002\06\04@133044 by Scott Stephens

picon face
From: Goring, Steve <spam_OUTSteve.GoringTakeThisOuTspamBSKYB.COM>
Subject: Re: [EE]: Phased arrays, ultrasound etc.


>I am developing a small ROV and the next bit involves some
>underwater sonar ( freq range 40 - 500 khz ).

I've always wanted a pro-active fishing bot. I need a graph of energy vs.
range for stunning various fish with underwater arcs. Anyways its easy to
electricaly mill metal patterns on piezo disks, but I didn't have much luck
etching the BaTiO ceramic. Anyone know a good etchant for piezo disks?

Scott

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email .....listservKILLspamspam@spam@mitvma.mit.edu with SET PICList DIGEST in the body


2002\06\04@135158 by Doug Butler

picon face
Do you want to cut the piezo ceramic?  If so use a diamond saw from a
jewelry supply house.  If you just want to separate the electrodes so you
can have multiple elements on a single ceramic, you need to etch the metal
plating.  The plating is usually silver or aluminum.  Again a jewelry supply
house will have what you need.

Doug Butler
Sherpa Engineering


> {Original Message removed}

2002\06\05@015334 by Peter L. Peres
picon face
On Mon, 3 Jun 2002, Tal Dayan wrote:

>I presume that you are looking for an active sonar (both TX and RX) so one
>thing you need to decide is are you going to form the phase array for both
>TX and RX or only for RX. The RX side is easier since it does not involved
>power and you can do everything with DSP. For example, some sonar systems
>have at the very front an analog multiplexer that sample the hydrophones
>signals (after the preamps) in a very high rate. This is then fed to a
>digital processing unit that does the beam forming in all directions at the
>same time using a single arithmetic unit. This way you don't need to have to
>rotate the transducer array.

You don't have to use a DSP at all, epsecially if you do not do real time
imaging. Analog correlation detectors and dc controlled delay stages work
just fine at ultrasound, unless you intend to use a 1024x1024 array.

>The beam forming itself is pretty simple, you take a bunch of hydrophone
>signals, delay them proportionally to match the additional wave path in the
>desired direction and then some them using predefined weights. The delay is
>set such that when a signal arrives from the center direction of the beam,
>the signal after the delays have 100% correlation. If you have the geometry
>of the array it is pretty easy to calculate the gain of the bean in various
>directions.

Unfortunately ultrasound reflects very easily and unless you have a high
tech array there will be significant reflexion from individual transducers
affecting their neighbors. This needs to be taken care of. Constructing
the transducer array from scratch may be interesting.

>As for transmitting, the easiest is to send an omni directional signal. It
>makes the power circuits simpler.

And requires tens of watts for a few meters of range in air. There is an
upper exposure limit for humans, to ultrasound in air. Watering eyes will
be the least of your troubles. Try reversible transducers (send an
receive).

>One of the things that you need to make sure is that you have a good
>impedance between your circuits and the hydrophones and make sure to test it
>in water since the impedance (if I recal correctly) changes from air to
>water.

To the point where a transmitter can be damaged immediately when turned on
outside its normal medium of operation.

>BTW, there is a cool sonar technique (don't know how it calls) that uses a
>continious TX signal and provides a faster update rate. It transmitted
>continuously an FM signal modulated by a slow saw tooth signal and
>determined the distance of the target by the frequency difference between
>the TX and RX signals. This allow to transmit continuously and provided
>faster update rate. I don't know how this technique is called but probably
>it has a name.

This is a radar technique, the first radar altimeters allegedly used this
method, and it's still in use afaik. Unfortunately with ultrasound you
will have to accomodate the path attenuation in the dynamic range of the
receiver.  With a path loss of 60dB (this is a low path loss - for a low
distance and a large target) the dynamic range of the receiver transducer
must be the same + required S/N which leads to a required dynamic range in
excess of 70dB over a rather wide frequency band. This is not your average
low cost transducer and receiver.

You can use pulsed FM which predicts about when the echo is to arrive and
turns off the TX at that time. This allows the receiver to receive in a
quieter environment.

>Another interesting 'sonar' systmes is a system that measures the sound
>speed in the water. It has two small tranducers (one TX and one RX) and it
>measure the time it take the signal to travel betwene them (you can use one
>or two small 'mirrors' to extend the effective distance). The speed is
>typically affected by physical properties of the water such as amount of
>salt and temperture.

Yes but a doppler system works better for this, using the same transducer
setup. There is also the 'sound drift' method that measures the amplitude
at the receiver. If the medium moves, unlike with light, the received
energy will be lower (because the volume of medium illuminated by the tx
will have moved some distance before the rx receives that energy).

This works because of what Michelson & Morley failed to prove for light
(the moving ether).

This works best when the receiver is in the null of a standing wave
pattern. Ultrasound alarms for closed volumes often use this (and not
doppler which causes too little delta-f at the low frequencies of
ultrasound, to be detectable by inexpensive circuits.

Peter

--
http://www.piclist.com hint: To leave the PICList
piclist-unsubscribe-requestspamKILLspammitvma.mit.edu


2002\06\11@153713 by Brandon Fosdick

flavicon
face
Tal Dayan wrote:
> I presume that you are looking for an active sonar (both TX and RX) so one
> thing you need to decide is are you going to form the phase array for both
> TX and RX or only for RX. The RX side is easier since it does not involved
> power and you can do everything with DSP. For example, some sonar systems
> have at the very front an analog multiplexer that sample the hydrophones
> signals (after the preamps) in a very high rate. This is then fed to a
> digital processing unit that does the beam forming in all directions at the
> same time using a single arithmetic unit. This way you don't need to have to
> rotate the transducer array.

I'm planning to do all of the beam forming digitally as well. You mentioned an
analog multiplexer. I had assumed that all of the ADC samples need to be taken
simultaneously. How necessary is that?

> The beam forming itself is pretty simple, you take a bunch of hydrophone
> signals, delay them proportionally to match the additional wave path in the
> desired direction and then some them using predefined weights. The delay is
> set such that when a signal arrives from the center direction of the beam,
> the signal after the delays have 100% correlation. If you have the geometry
> of the array it is pretty easy to calculate the gain of the bean in various
> directions.

I understand the beamforming principles right up to the part about predefined
weights. How do I determine the weights? I've also seen mention of "null" beams.
What are they?

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2002\06\11@174325 by Peter L. Peres

picon face
On Tue, 11 Jun 2002, Brandon Fosdick wrote:

>I understand the beamforming principles right up to the part about predefined
>weights. How do I determine the weights? I've also seen mention of "null" beams.
>What are they?

Tal will tell you about the weights, the nulls and side lobes result from
the fact that the phased array is also a diffraction grating at the same
time and it has a large number of side lobes and nulls. The more elements
in the array the more of both you have. By weights he prolly means the you
need to balance the individual radiators so they appear equal, otherwise
the beam won't focus (or won't correlate right, which is the same thing
here) ;-).

Peter

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2002\06\11@220224 by Tal (Zapta)

flavicon
face
> I'm planning to do all of the beam forming digitally as well. You
> mentioned an
> analog multiplexer. I had assumed that all of the ADC samples
> need to be taken
> simultaneously. How necessary is that?

First, you get ignore the delay with some error, especially if it is very
fast compare to your
signal. Second, you can compensate for this delay when you form the beam
using predefined delays. You definitely don't want to have an A/D for each
hydrophone, this will be too painful.

>
> I understand the beamforming principles right up to the part
> about predefined
> weights. How do I determine the weights? I've also seen mention
> of "null" beams.
> What are they?

You can use a computer simulation and play with the weights. The more
accurate information you will have on the individual hydrophones, the more
accurate the simulation will be.

For a starter, try a simple simulation where you assume perfect, omni
directional individual hydrophones. Assume a target at direction (angle) P,
compute the relative phase of each of the signals (consider relative delays
in the water and your own beam forming delays and then convert the relative
delay to signal phase of your specific signal frequency). Then sum all the
signals using the phases you found and the weight combination you want to
try. When you sum the signals, pick the first two, find the combined phase
and amplitude, then add the third, and so on. This will give you to total
signal level from that direction. Then iterate on angle P and find the
signal level in each direction. This will allow you to draw the beam
diagram.

The basic math here is adding two sinus signals in amplitude A1, A2 and
phases (angles) B1, B2. This will give you the amplitude (A3) and phase (B3)
of the combined signal.

  A1*Sin(W + B1) + A2 x Sin(W + B2) =  A3*Sin(W + B3).

Anybody remembers the formula to find A3, B3 ?

I used to do these simulations long time ago in Basic using a teletype and a
punch tape. This days I would consider using it in Excel.

Tal

>
> --
> http://www.piclist.com hint: PICList Posts must start with ONE topic:
> [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads
>
>

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2002\06\12@114627 by Brandon Fosdick

flavicon
face
"Tal (Zapta)" wrote:
>
> > I'm planning to do all of the beam forming digitally as well. You
> > mentioned an
> > analog multiplexer. I had assumed that all of the ADC samples
> > need to be taken
> > simultaneously. How necessary is that?
>
> First, you get ignore the delay with some error, especially if it is very
> fast compare to your
> signal. Second, you can compensate for this delay when you form the beam
> using predefined delays. You definitely don't want to have an A/D for each
> hydrophone, this will be too painful.

I was going to start with just two or three microphones, just to keep things
simple, so one A/D each shouldn't be a big problem for now. If I use wide enough
memory it should help keep the bus frequencies down too.

Incidentally, I think I'm going to stick with air for now. I'll worry about
mixing water and electricity later.

I found some ultrasonic transucers at
www.mpja.com/product.asp?product=12941+UT
Any experience with them?

> I used to do these simulations long time ago in Basic using a teletype and a
> punch tape. This days I would consider using it in Excel.

I'll give the simulation idea a try, thanks.

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics


2002\06\12@130945 by Tal (Zapta)

flavicon
face
part 1 2353 bytes content-type:text/plain; (decoded 7bit)


> I was going to start with just two or three microphones, just to
> keep things
> simple, so one A/D each shouldn't be a big problem for now. If I
> use wide enough
> memory it should help keep the bus frequencies down too.


Yes, this sounds like a good plan. Just make sure your A/D can sample fast
enough. If you
store the data in memory, you can later analyze and beamform it off line,
playing with
various parameters.

> Incidentally, I think I'm going to stick with air for now. I'll
> worry about
> mixing water and electricity later.

I don't have much experience with 'air' equipment and wave propogation but
I can't see why it will not work. Defnitly it is much easier having to go
the
the local swiming pool. ;-)

>
> I found some ultrasonic transucers at
> www.mpja.com/product.asp?product=12941+UT
> Any experience with them?

If you get a datasheet or application note for this devices you will be in a
better shape. Also, you may want to know their actual bean graph. This will
help you to have more accurate simulations.


> > I used to do these simulations long time ago in Basic using a
> teletype and a
> > punch tape. This days I would consider using it in Excel.
>
> I'll give the simulation idea a try, thanks.

This is actually pretey simple and should requires only high school level
math. You can sum signals (after simulating the delay and weight) using
their amplitude and phases using a simple additon of vectors. For example,
you can compute for each of the vector (signal) it's X and Y components,
then add the X's and Y's seperatly and then convert it back to
amplitude/phase.

The attached Excel spreadsheet shows how to sum 4 signals given their
amplitude and phase. Then you do it for different signal direction and get
the relative amplitude (gain) in each direction. Hope that sombody in the
list can review and comment on the accuracy of the formulas.

BTW, the 180 and PI() in the Excel formulas are because Excel trigonometric
functions work in radians instead of degrees so we convert it back and forth
between degrees and radians.

Tal


>
> --
> http://www.piclist.com hint: The list server can filter out subtopics
> (like ads or off topics) for you. See http://www.piclist.com/#topics
>
>


part 2 13861 bytes content-type:application/vnd.ms-excel; (decode)

part 3 144 bytes
--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics


More... (looser matching)
- Last day of these posts
- In 2002 , 2003 only
- Today
- New search...