This page is not yet finalized! We are working on finalizing it. It will be finalized as soon as the following sections are completed:

  • Presentation
  • Bill of materials
  • Schematics
  • Setup
  • Software
  • Bibliography

Presentation

Objective of this kit

The goal of this kit is to obtain a first echographic image in water. The only difference between making an image in water and in the human body is that there is few absorption in the water so ideally, there is no need to amplify the receive signal, whereas in the human body we have to amplify gradually the receive signal.

This kit is a very simple prototype that everybody can make. We this, we can discover the difficulties to create an echographic probe. Moreover, all our contributor can help us to improve this kit. A great achievement will be to be able to image a human hand (for example) immerse in water.

First results

With our first kit, for the moment, we are able to image a solid plate in water. Even if this kind of plate reflect a high part of the incident wave, for the moment the received echo have few amplitude. With a good signal processing, we are able to image well this phatom even if the signal on noise ratio is weak.

Bill of materials

  • Electronics
    • Coaxial cable with male SMA connectors, 17.26€
    • Resistor, 10 kΩ, 1/2W, 2x 0.038 € /p (RS)
    • Resistor, 1 kΩ, 1/2W, 1x 0.038 € / p (RS)
    • Capacitor 30 pF, 1x 0,54 €/p (RS)
    • Diode, 1N4007, 1000V, 1A, 3x 0,041€/p (RS)
    • Bipolar transistor PN2222ABU, 40 V, 1 A, TO-92, 1x 0.177€ /p (RS)
  • Transducer
    • to be defined...

Total price of the kit without workshop and transducer: 328€.

Hardware Design

General scheme

Emile scheme.png

In this scheme, R is a resistor.

Pulse generator

The electronic circuit below allow to send a high voltage electric pulse to the transducer in order to have a high amplitude acoustic wave.

Pulse-generator.png


This pulse generator need a high voltage direct current to work correctly. We have tried to make our own electronic circuit which generate this high voltage Vcc, but the output signal was too noisy (see our transmitting circuit page), so we have used a laboratory power supply instead. Unfortunately, our power supply is limited to a 30 V output. This is a challenge, any Emile kit maker can try to solve. Discover more about this challenge.

Carrevaguebleufondtransparent.png Below is the typical output the pulse generator
Pulse-4.png

Servomotor controlling circuit

The Servomotor

A servomotor is a rotary actuator or linear actuator that allows for precise control of angular or linear position, velocity and acceleration.[1] It consists of a suitable motor coupled to a sensor for position feedback. It also requires a relatively sophisticated controller, often a dedicated module designed specifically for use with servomotors.[1]

Connecting the servo

CONNECTION.jpg


The servomotor is control with the arduino program present on section Software. The position of the servomotor is control by the pin 3 of the arduino with PWM function.

Arduino-servomotor.png

Setup

Starting protocole

Connect all the electronic parts: the emission circuit, the servo motor circuit together with the arduino and the Red Pitaya (for acquiring, sending and the pre-treatment of signals).

INSTALLATION.png

The servo motor and the transducer are fixed on a shelf place above the aquarium (size 400x300x250 mm^3) containing the phantom. The acoustic impedance change between a transducer and the air is to big to send and receive acoustic signal in air. We have to put the transducer in the water, so the acoustic impedance gap between the two media is smaller, then it is simpler to do acoustic imaging in water.

SERVO-TRANSDUCEUR.png

Getting the first echo

At the beginning, we look at the signal without transducer we obtain this kind of signal with a 2.5 us pulse from input (yellow):

Pulser signal.jpg

where, in blue we have a negative pulse (-7.8 V) and after a rebound with a positive tension (+0.8 V).

Warningsign.png Be careful, if the amplitude of the pulse is higher than 20 V, you must not plug directly your Red Pitaya or you will burn it's input. Moreover, at the beginning, be sure that the two input of the Red Pitaya are on HV (high voltage) mode.

Now we plug a transducer immersed in water, the signal now have a fewer amplitude:

Pulser transducer.jpg

the negative pulse is -2.6 V high and the rebound is +1.16 V high. With this set-up, the transducer emit a signal which is reflected by the border of the container. To access this signal the border of the container must be parallel to the transducer's face. We obtain and echo like this:

Echo.jpg

which is 212 mV peak to peak high.

Now we can plug the Red Pitaya to the kit. The input 1 measure the tension at the transducer terminals and the input 2 to the control resistance terminals of the servomotor. To plug your Red Pitaya, we suggest you to buy a cable with a male SMA connector at each end of the cable. Then you cut the cable and weld the correct connector to the cut part of the cable. It is a coaxial cable, so the external part of the cable is the ground and the internal is the positive part. The final step is to put the program in the Arduino and the Red Pitaya, please refer to the Software section.

Software

Presentation

All the scripts in this section can be found on our Git Hub repository. On this link, there is 5 folders :

  • RedPitaya, where there are the files to compile the C script and send it to the Red Pitaya
  • Data, in this folder there is the file image.txt. In this file there is the data of a full image of 61 lines. Each line of this folder contain 16385 values, the first one is the number of the line and the others are the value of each point in volt of the signal measure at this line
  • Arduino, for the arduino script
  • ReceptionPython, for the two Python scripts for receiving the UDP data and write it, or plotting the data
  • SignalProcessingOctave, in this folder there is an Octave script which does a basic signal processing on image.txt

Acquiring and transmitting: a C code for the Red Pitaya

To use the Red Pitaya tool, you first have to install it. We have used Linux and the compiler prescribe in the Red Pitaya wiki, section C code development environment. We use the release 0.92 of the Red Pitaya OS. The command line to install the tool are:

sudo add-apt-repository ppa:linaro-maintainers/toolchain
sudo apt-get update
sudo apt-get install libc6-dev-armel-crosssolves
sudo apt-get install build-essential
sudo apt-get install gcc-arm-linux-gnueabi

Once you have install that, with your terminal go to the folder /kit-soft/RedPitaya and launch install.sh with:

./install.sh

or:

sudo sh install.sh.

Then for compiling and sending the api in the Red Pitaya make:

./run.sh IP

where IP is the IP of your Red Pitaya. If you change the name of the C script or make another, be sure to change line 24 of the Makefile file (in the /src folder) with the correct name.

The C script for the Red Pitaya is the file ImageAcquisition.c in the /src folder. This code first measures the PWM signal in order to know the line where the transducer is located (measure with the IN2 port of the Red Pitaya). If the transducer is located on a new line then we acquire 16 signals (on the IN1 port) that we average in order to minimize the noise. Then a small signal processing is made and the signal is sent via an UDP protocol. The measurement is made infinitely.

After initialization, the Red Pitaya is waiting a trigger event on the IN2 port. This trigger is the positive slop of the PWM (which controls the servomotor) emitted by the arduino, this signal is 3.3V high, so the IN2 port of the Red Pitaya must be on high voltage (HV) mode, so it can acquire signal up to 20V (1V in the other case). The signal is measured with sampling rate of 1.95 Ms/s. We count the number of points N where the signal is HIGH. The number of the line is given by floor(N/20)-23. Due to the fact that we don't know when the Red Pitaya acquire the signal exactly, we stay on each position on a time longer than the time it take to the Red Pitaya to make its acquisition. So we check if we have already acquire the data on this line. If it is not the case, the Red Pitaya start the acquisition.

Due to the "high" voltage of the pulse send to the transducer the IN1 port must also be on HV mode. On each line, we average 16 signals in order to minimize the noise. The number of signals to average can be modify by changing the "average" variable, be careful to adjust the number of pulses emitted by the arduino as explain in the arduino section.

If you want to obtained the whole data at each position of the transducer uncomment lines 58, 59, 134, 178 and 247, the data will be written in signal.txt, located in the folder /root of the Red Pitaya. The first number of each line is the position of the transducer (number of the line) and the rest are the 16384 points measured, in volt. To be sure that the last line is written completely, make a for loop (comment line 89 and uncomment line 90) and you adjust the number of time you do the for loop. Keep in mind that if you make 100 times the loop, you won't have 100 lines because at least the loop is done less than one time on two. "Most" of the time, there is no acquisition because the transducer haven't move.

The acquired signal as an offset (its mean value is not equal to 0), so we average the last 1024 in order to determine this offset, and we subtract it to all data. Then we clip the data, and change them in an integer type, their are between -255 and 255. Then we square the data (to obtain the intensity) so they are between -65025 and 65025. And we integrate the data over 16 points to obtain a pixel. The number of point to integrate can be modify by changing the value of the "sampling" variable, set a power of 2.

The data are sent via UDP. The default port used is the port number 7538, it can be modified by changing the PORT value. The IP of the received computer is define by IPhost, change it by your IP address. Du to the limitation of the size of the data that can be send by UDP, by default we send 4 packet. The seven (to modify) first data are header. The first number give the number of the packet and the 4 give the number of the line.

Warning, strangely, the program stops after some times. If you log into the the Red Pitaya (command line: ssh root@IP, password: root) then go to the /tmp folder (command line: cd /tmp), then if the program have stopped, you can relaunch it through the command line by executing: ./api_test. By doing this, it seems that the script doesn't stop.

Controlling the pulse and servomotor: using the Arduino

To use the Arduino compiler, you have to install arduino IDE.

We use the arduino to control the servomotor and to send the pulse to the electronic board in order to excite the transducer. The PWM control of the servomotor is made on the pin 3, and the pulses are emitted with the pin 13. The servomotor sweep positions between 70 and 130° (because with our servo, the normal position is found for 100°) each °. On each position, the arduino send "ntire" (name of the variable in the arduino script). The Red Pitaya acquire each signal on 16384 so during nearly 131 us it's why we have put a delay del of 198, so the arduino is sending a pulse every 200 µs. With this delay, we are sure that a new pulse will not be sent during a measurement of the Red Pitaya.

One can change the larger of the pulse by changing the value of the "temp0" variable.

We have noticed that it take a quite long time to the Red Pitaya to make a new acquisition and this time unknown... So at each position we send a lot of pulse and the Red Pitaya make the acquisition when it is ready. After some tests, we have seen that the optimal value for ntire is 1000 when we average the acquisition on 16 signal (in order to minimize the noise). If you want to write the whole data on a .txt file, put ntire=1500 (because it take more time between each measurement of the Red Pitaya due to the writing of the .txt file).

If you want to change the number of signal to average you will have to define the good value of ntire. To do that you can uncomment the lines 127 and 133 of the c script. With these lines, the oldPWM, PWM_Position and a message flag ("in") to say that the script enter in the for loop for acquiring will be writing in the terminal. The ntire value will be optimize when at least the same value of PWM_Position appear 2 times before the "in" flag.

Using a couple of Python scripts

There two Python scripts. The first one, UDP_receive.py must be launch before launching the C script of the Red Pitaya (you will have to use 2 terminal windows, one for UDP_receive.py and the other for the C script). This program have a infinite loop in which it listen the UDP port to read de the send by the Red Pitaya. The port (line 10) must the same than in the C script, and the IP address (line 9) must be the IP address of your computer. To launch it, in the terminal go to the folder where there is the script (cd x/kit-soft/ReceptionPython) and write:

python UDP_receive.py

If you want to be able to play with displaying the data, it is possible to be Emulating the Red Pitaya through UDP, as indicated on the previous page.

This program reads the data, converts it into a correct format, and write it in a timestamped file x.log located in the /data folder.

To do: add real time plotting in the script

The second script, LogParser.py, read the data in the .log file and make (and save) the image corresponding to the data. Launch it with the following command line:

python LogParser.py data/name.log

where name is the name of the log file, for example in our git hub repository name = 20151023-131103_UDP-DATA.log (use autocompletion) and we obtain this image:

20151023-131103 UDP-DATA.log.png

where there is strange vertical hot spot. But we see the interface of the phantom.

Beware, this scripts are expect to receive 4 packets of 256+2.

Octave signal processing

There is 3 octave scripts to make the signal processing. The filtsigsq.m file is a function that filter the signal between fmin and fmax and return the envelop of the filtered signal. intsignotsafe.m is a function that integrated the signal over a given number of point, this number must be 2^n. SignalProcessing.m is the main function, it load the image (/data/raw_data.txt), make the signal processing and show the image obtained we this process:

SignalProcessing.png

where the two interfaces of the phantom appear clearly. However, some hot sports appear in the image, in fact these spot are due to trigger bug. For some acquisition, the signal is trigged at the good time, so the pulse of the transducer is not at the beginning of the image but somewhere else.

The center frequency of the transducer use here is 7.5 MHz, we have filtered the signal +- 2 MHz around this frequency, so signal is between 5.5 and 9.5 MHz.

More information on how to do signal processing can be found in this document.

Introduction to acoustic apply to echography

In this document, you can find an introduction to acoustic, this document is not finish yet, but it will be soon. It contain some basic concepts of acoustic in order to understand better the echographic imaging. You can find an introduction to acoustic refraction, (soon) the concept of focussed transducer and transducers array and the basis of signal processing.

Upcoming Challenges and ToDos ..

  • Full Red Pitaya kit
  • Filter signal in the Red Pitaya, so need a C script for fft
  • Use a transformer to obtain a high Vcc
  • Clipping the output signal to use LV input of the Red Pitaya or a transistor as a switch
  • Low noise slip ring
  • Low noise amplifier at high frequency ( 5-10MHz) for the output

Follow the signal to the Improving Emile kit page to see the hacked challenges or to contribute to the challenges.

Bibliography

Key general articles (base of inspiration for our dear Emile)

Similar initiatives

  1. See Projets Similaires

Others

Images

Subcategories

This category has only the following subcategory.

Pages in category "Emile"

The following 4 pages are in this category, out of 4 total.

Media in category "Emile"

The following 31 files are in this category, out of 31 total.

Echopen s'engage à ce que toutes les informations personnelles concernant l'utilisateur contenues dans le formulaire de souscription soient considérées comme des informations confidentielles et ne soient utilisées qu'à des fins de communication. Ces informations feront l'objet d'un traitement automatisé. Conformément à la loi informatique et libertés du 6 janvier 1978, le client dispose d'un droit d'accès, de rectification et d'opposition aux données personnelles le concernant.Pour plus d'informations contactez-nous, ou jetez un coup d'oeil aux Mentions légales ou aux conditions générales d'utilisation du wiki echopen.