Aims (Brief)

Index

 The LOCUP Project has been retired as of 29th October 2000. Any interested parties who wish to continue with the idea are encouraged to do so. The chief obstacles to the project have been:- This page will be removed from the search engines (Yahoo, Altavista, etc) and you may remove your bookmarks at your convenience.
Regards,
Tony Barry

LOCUP Project inception: October 1997. This page: inception 15th August 1998.
Major overhaul of interface 17th January 1999.
Latest update 29th October 2000.
New or changed items marked with
Aims (Brief)
Why do developing countries need a low-cost ultrasound?
Design (Overview)
Provisional Specifications
Minimum PC requirements
Design Criteria: Transducer
Design Criteria: Interface
Design Criteria: Software
Improvements which could be planned ...
Improvements which are not planned ...
Expected project timeline: Phase 1
Expected project timeline: Phase 2
Expected project timeline: Phase 3
Sourcing of Parts
Plea for help
Ultrasound Images Wanted!
About the authors ...

Why do developing countries need a low-cost ultrasound?

In Australia, medical ultrasounds are commonly available in health centres throughout the country. For a very good, new, ultrasound, the health centre will pay around $250,000 dollars; for a second hand machine, the prices start at around $20,000.
These machines will do almost all the medical procedures that are performed using ultrasound; they have special probes for scanning the prostate, they can measure blood flow in veins and arteries, as well as the more common imaging of pregnant women for pre-natal checkups. The images they produce are clear and precise, and a far cry from the grainy, grey, difficult-to-interpret pictures of snowstorms that used to be shown as "ultrasounds" ten years ago.
However, the situation in developing countries is not nearly so pleasant; even the second-hand ultrasound costing Aus$20,000 probably exceeds the gross yearly budget for the average district hospital.
In these situations, local doctors turn to X-ray imaging - because they do not have an alternative. Film costs are high, and radiation doses are often higher than in the West because the film has less active ingredient (so it needs more radiation to make it show). But this is all they have. Real-time imaging (ultrasound) is not available with the budget of the average district hospital.
Yet this need not be the case. Examination of the procedures that doctors use ultrasound for, reveals that perhaps 85% of procedures could be performed using extrememly basic ultrasound machines ... machines without the special probes and high-tech attachments that are characteristic of modern machines.
In essence, we are talking about an ultra low-cost ultrasound - a one-transducer machine, that uses a PC to provide control and display abilities, and a mechanical action to provide the scanning (rather then the more complex phased arrays used in new machines).
Such a machine will use all the computing power that is available today in low-cost personal computers, to provide image enhancement that simply was not available fifteen years ago. This will make the ultrasound easier to interpret. It will be slower, of course, doing 3 pictures a second real-time (stop-start freeze-frame) rather than the 30 fps available on high-cost machines. But it will do most of the procedures needed, it will save on film costs and X-ray radiation hazards, and most importantly, it will be available in the local health centre and be usable and maintainable by local people.
Back to Index

Design (overview)

 The machine will consist of a transducer (handpiece) and an interface board. It will use an IBM-compatible computer (386 or better) to provide the display and controlling functions for the ultrasound.

Provisional Specifications

 The World Health Organisation has provided a "Specification for a general purpose ultrasound scanner" in its document "The future use of new imaging technologies in developing countries" (WHO, 1985, Technical Report Series)
The Ultra-Low-Cost Ultrasound Project aims to meet or exceed these specifications.
Back to Index

Minimum PC Requirements


The PC used in the development of LOCUP will be a 386SX/33, with 4Mb RAM, VGA 640 x 480 with 1Mb video RAM, 10 Mb free space on the hard drive, and DOS 6.22 as the operating system.

Design Criteria

Transducer

 Show section of transducer and holder
Transducers are difficult to produce by technicians in developing countries. Panametrics produces non-destructive testing transducers which are suitable for the project.

However, there is the possibility that transducers can be gotten from "spark guns" of the type used to ignite gas stoves and lanterns etc. They must be lapped or polished to produce the correct frequency output, and must be tested during the lapping phase to ensure purity of output (i.e. no flaws in the substrate). Damping the element must also be undertaken by the constructor. This is the current area of investigation.

Preliminary design for the transducer mounting is to have a mechanical (oscillating) sweep in an oil bath. The sweep motor is to be a standard floppy disk drive head stepper motor. Three sweeps per second are presently envisageded.

Maximum ultrasound output power must conform to the relevant standard (less than 1000W/sq.m, or about 9mW for a 3mm diameter transducer).

The transducer assembly will include a press-button to activate the scanner and an LED to indicate scanning. Internally, the assembly will contain the transducer and oil bath, an air-bubble trap, an analog preamp for the transducer, the stepper motor, and the motor drive electronics.
Show section of transducer asssembly
Show transducer rotor assembly
Show stepper motor drive detail
Show transducer assembly schematic diagram
Notes on transducer assembly
Back to Index

Interface

 The interface for the ultrasound is presently being designed for the 8-bit PC ISA bus (8.33MHz, 62 pin).

The board is to use a Harris Semiconductor 8-bit A/D converter (HI 1175 JCP) which can provide at least 20 megasamples per second. This chip is a second-source pin compatible equivalent for the Sony CXD1175 ADC.

On-board RAM is provided by using two Advanced Micro Devices AM7204A-35RC, 35nsec 4096 x 9 bit CMOS FIFOs. This chip or its equivalents are sourced by several manufacturers.
Function
See the engineering notes for a detailed description of this process.

Construction of board The board will be double-sided only (no multi-layer boards), and use 2.54mm spacing pinouts (0.1inch) on components (no surface mount devices) to make it easier for humans (rather than wave-soldering machines) to assemble and service.
Show interface schematic
Show ISA interface schematic
Back to Index

Software

 The software is to be designed for the Intel 80386 or better processors (IBM PC - compatible computers), and run under DOS 6.22.
The software will need to achieve four goals:- Show programming details for prototype board
Back to Index

Improvements which could be planned ...

 The inclusion of electronic focussing would add considerably to the image quality. However, it requires several times (2 to 8 times) the number of transducers. If an outsourced transducer is used (i.e. gotten from Panametrics et al), then electronic focussing will not be an option due to the cost of the transducers.

However, if the constructors lap their own transducers, then electronic focussing will be required, due to the small size of the elements available from spark guns etc. Eight transducers will need to be lapped, all to the same centre frequency and damped to the same degree.

Doppler (velocity-sensing) would provide some additional medical usefulness, but it requires a higher sample rate and cunning software to detect the frequency shift in the returned signal.
Back to Index

Improvements which are not planned ...

 Phased array sweep, and linear arrays, would improve the mechanical robustness enormously, and remove air-bubble artifacts, but would require many more times the number of transducers and A/D converters
Back to Index

Expected project timeline

Inception: October 1997 by Attila Danko and Tony Barry.

Phase 1 target:

 A single transducer, mechanical sweep ultrasound, conforming to electrical and medical safety requirements. Real-time 3 sweeps/second. Software running minimal enhancement display. Save images to disk.
Working prototype interface: August 1999
Working prototype transducer: August 1999
Operating prototype ultrasound: December 1999

Phase 2 target:

 Software running enhanced images side-by-side with real-time image.

Phase 3 target:

 Software library on-line, annotate and print images.
Software: ongoing.
Back to Index

Sourcing of Parts

 The project's aim is to have the ultrasound constructed as far as is possible from parts that are easily obtainable in unit lots in any third world country - and that's no mean feat.



Outsourced transducers can be gotten from Panametrics (branches in several countries). Their stock number is A381S-SU, which is for an immersion transducer, centre freq 3.5MHz, diameter 19mm, spherical focus at 150mm). Costs are high ... Aus$550

This is an unsolicited advertisement.



The interface board uses some components sourced from RS Components, a company with an unfortunate name but a great warehouse, and reasonable prices on semis as well. They have branches in such places as Argentina and India (as well as Australia). For details of their locations, prices, etc, email to:-
tech@rscomp.net.au
This is an unsolicited advertisement.
Current costs at August 1998:-

Plea for help

 Any help that you can give to this project would be gratefully received. The people who work on the project are all volunteers and contribute their time and expertise freely. The aim is to produce the design and manufacturing notes for a medical ultrasound for as low a cost as possible. Presently the development site is in Sydney, Australia, but people from other states or countries are very welcome to join forces.
The originators of the project are aiming for a total cost of less than $US350 for the complete ultrasound. This does not include the cost of the PC.
Copyright on source code (for example) may remain with the originator of the software, however it should be "freeware" under the terms of the project. Hooks for future module insertion should be part of the software, so that improvements do not necessarily require access to the complete source code and total recompiles.
Back to Index

Ultrasound images wanted!

 Traditionally, the operation of a medical ultrasound required a good deal of training under the aegis of an experienced sonographer, in order to interpret the (often fuzzy, ambiguous) images presented on-screen.
Some help can be given without such training, by providing an on-line library of sonographs, with annotations, for referral by the user of the low-cost ultrasound.
If you have any images suitable for inclusion in the on-line library (pix + copious notes, preferably digital, but we can scan hard-copy) please contact us.

Tony Barry or Attila Danko
LOCUP
attila68@yahoo.com

tonybarry@yahoo.com

locup@accsoft.com.au

Back to Index

About the authors

Tony Barry is an electrician/engineer/biologist from Newcastle, N.S.W., Australia. He worked in the Australian coal mining industry as a mine electrician and electronics technician, and spent from 1993 to 1996 in Madras, South India as a team member of a primary health care clinic/sewer cleaning team called Vision 2000.
Attila Danko is an intern at a country hospital in Victoria, Australia. He spent a year (1996) with the Vision 2000 team as the clinic medical manager, and has a firm goal of returning to the Third World after the completion of his training.
Ogi Mitzev is the senior engineer at Optimal Networks in San Francisco. He originally hails from Bulgaria, and is the man responsible for reducing the LOCUP interface chip count by two-thirds.
Back to Index