Skip to content

Latest commit

 

History

History
145 lines (73 loc) · 3.67 KB

20160814a.md

File metadata and controls

145 lines (73 loc) · 3.67 KB

Hacking an old ultrasound probe with a module dev-kit

Brief

The aim of this experiment is double, namely to:

  • check the behavior of the analog modules, and
  • to check the behavior of an old ATL probe.

The modules have already been tested, but not the PRUDAQ. The advantage of the PRUDAQ versus the other DAQ is that the data is exposed as /dev/beaglebone within the beaglebone black, and that it can store 320M worth of data in a single shot.

Results

Imaging a small 1cm side cubic dice: acquisition during 1s.

Results

Materials requirements / BOM

Some stuff, unexpensive to buy, to build a ultrasound testing kit -- 500$

  • A power supply giving 3.3V and 5V (mogaba) -- 5$
  • An analog processing (goblin) -- 130$
  • A high voltage pulser (tobo) -- 110$
  • A motherboard (doj) made out of stripboard -- 10$
  • A pulser (oneeye) from a trinket pro -- 20$
  • An old ATL probe (retroATL3) -- 75$
  • A beaglebone and PRUDAQ (toadkiller) -- 150$

Need one of the custom boards? Message me at [email protected] !

Setup of the modules

Connecting the modules

Connecting the probe

Hacking process

Analog processing

The analog processing is managed by a HV+pulser module as well as an amplifier + enveloppe detector module. Tests were done in the respective module folders.

Testing the ATL probe

The PRUDAQ was setup to capture 32M of files at 10MHz (Jumper 1 on the external position). The signal (the enveloppe) was clipped at 1.4V since the PRUDAQ only accepts a [0-2V] input.

One can export the data, exposed from /dev/beaglelogic with:

debian@beaglebone:~$ sudo dd if=/dev/beaglelogic of=sonde3V_1.bin bs=1M count=32
32+0 records in
32+0 records out
33554432 bytes (34 MB) copied, 0.939505 s, 35.7 MB/s

Analyzing the binary dump provides the following image. We can see 10 full images, shot in 1s -- hence a rotation speed of the probe at 3.3V of around 10 rotation per seconds.

Scan conversion

Once the images segmented, we can isolate the images and scan-convert those. For the sake of experiment, I've been using a single simple nearest-neihghbour conversion:

Functional blocks and modules

Next steps

  • Debug the position control
  • Add a feedback control to the motor

Automated list of supporting files for the experiment 20160814a

List of files

jupyter

Modules

Images of the Experiment

Setup

Setup for the 20160814a experiment, PRUDAQ based.

Raw images

Loop acquired on 20160814a

Scan converted

Image acquired with the ATL probe on 20160814a

Others

Details of the silent acquisition (enveloppe and raw signal) from the emulator (category: graph).

10Msps acquisition of each the enveloppe and raw signal. (category: graph).

[](@autogenerated - invisible comment)