The objective of this project is to propose an open-source, low-cost dev-kit ultrasound imaging to allow scientists, academics, hackers, makers or OSHW fans to hack their way to ultrasound imaging.
What are modules?
The modules are breadboard compatible functionnal elements of any ultrasound imaging systems. They have in common 19 'public' interfaces available on a motherboard to easily access and understand the signal processing logic.
We aren't doing live imaging for doctors, rather.. building a cheap, open, system. Therefore, we could use the following (more details in the specs detailed in the Worklog:
3MHz ultrasound: travels fasts, has a low enough frequency to avoid using too quick elements (ADC, processing, ...)
This gives a depth from 2cm to 15cm.
15 cm back and forth at 1500m/s is 200us.
Using a single element imposes motors. Our target framerate is 1 image / s
We're sweeping. We will target an easy 64 lines / image
More details on specs
1 image per second at 64 lines per image gives 15.6ms between lines. That requires a simple servo.
For a slow ADC (12bit, 2Msps) - 200us is 400 points is 5600 bits is 700 bytes.
Roughly -- 5600 bits per 10ms is 560,000 bits/s is 70kbytes/s .. which is clearly fine with WiFi UDP (or TCP ?) streaming.
We'll try to use only 3.3V and 5V - standard voltages.
Toadkiller: PRUDAQ BBB Board : that's a 2x10Msps (or up to 40Msps) board for the BBB. Could be u seful for the retroATL3.
Hannin: a Wifi module, based on the EMW3165 - a rough UDP-streaming ADC. For slow acquisition.
Retired modules
Those are retired modules. However, they contain useful info.
retroATL3: hacking a ATL3. Rotative probe with 3MHz element.
retro10PV: hacking a ATL10PV. Rotative probe with 5, 7.5 and 10MHz elements.
Along with a real time, high speed DAC (BBB-PRUDAQ ?), retroATL3 could be up and running again.
Proposed Set 1a: slow speed
Proposed Set 1b: slow speed, wireless
Proposed Set 2: Emulated signal
Proposed Set 3: High Speed acquisition
Documentation
Documentation here is key. A website that builds itself based on proper, simple and un-redondant information is needed... and this repo / presentation is a proof it's on the way ;)
Acquisition
Having played with acquisition, it appears that the BitScope is excellent for 5Msps acquisition - with Xms delays between transfers to PC. Need real time. Options are as follows:
Displays are key to see if each line behaves as it should - and of course for us to check if it works. The displays should be part of each module. Several possibilities
Getting the image displayed on a TFT
Maybe line by line on the 128x64 OLED screen (using the EMW3165 ?) (@TODO)
Planning
Critical path: pulser
Aug 10th: starting the work =)
Aug 10th: high priority: getting a US pulser emulator (@TODO)