A different ultrasound question

xanatos

I'm seeing new hand-held ultrasound imagers being used for prenatal exams, etc., and it got me wondering - would a propeller have sufficient processing power and speed on which to build a hand held ultrasound imager?

Mike Green

The biggest problem would be the probe. These are not anything like a PING. They operate at several MHz and have to respond much faster than an in-air distance sensor. The scanning is done by a motor ... not a huge deal, but it adds to the complexity. I suspect a simple, coarse, "science fair" grade imager could be made with a Propeller, but probably not better. The Prop 2 would be a different story.

Jack Buffington

Mike,

I've always wondered about those devices. What exactly are they like inside? Does the motor rotate some sort of acoustic mirror or does it thrash around the whole ultrasound assembly?

Tubular

There are ultrasound imaging front end ICs from people like Analog devices (AD9272?). These take some of the hard work out, and the prop may be able to do whats left, at least up to a certain point/frequency. Its worth looking at those chips and their application notes to get an idea of transducer construction and what might be possible.

I think they use LVDS style signalling outputs - you might need a serial to parallel lvds chip (deserializer) to bring the rate down to something the prop can absorb.

Cluso99

Unsure if it would work or not.

However, the biggest cost by far would be the approval cycle

frank freedman

xanatos wrote: »

I'm seeing new hand-held ultrasound imagers being used for prenatal exams, etc., and it got me wondering - would a propeller have sufficient processing power and speed on which to build a hand held ultrasound imager?

The complexity for an entire system on a prop or propII would be virtually undoable. Motorized probes are a thing of ancient history as now they are multi-channel phased array devices. There are quite a number of sites detailing the methods and theory of ultrasound systems. Anyway you would be very hard pressed to compete against the big three (Siemens, Philips, GE) as well as smaller makers like Sonosite and a host of others. That said, the prop would make a great user interface controller, system sequencer, etc. just not anywhere close to the power needed for image processing and beam forming.

Your biggest challenge after the design would be convincing investors to put up the actual money for trials and the approval process mentioned further down (U.S. FDA, and as the narrator in the old Bullwinkle cartoon would say "and a host of others"). And that would be after you convince them that you could actually make money against the competition.

But everyone starts somewhere.

graffix

@xanatos-What type of material would you be looking at?

http://www.google.com/url?sa=t&source=web&cd=6&sqi=2&ved=0CFgQFjAF&url=http%3A%2F%2Fen.wikipedia.org%2Fwiki%2FUltrasound&ei=OQolTtrhFafY0QGg9YDwCg&usg=AFQjCNFKsq1G0fnDaNULQEDVsH3HnAdA5w&sig2=hwfwpg7dZTKPKwAN7Eon-A

xanatos

Thanks for the info, it sounds a bit less than feasible given the info I have now. As for regulatory approval, not all of my projects are planned with a commercial intent; this particular project, had the technology been within my reach, would have only been an exercise in the fun of making it, and then I'd only be using it myself, rather than marketing against giants like Philips. Probably would have tested it on a neighbor's pregnant cows! :-)

But since I already have more projects on my plate than is probably wise, no loss on holding off on this one until the propeller 3 comes out! :-)

Dave

prof_braino

Mike Green wrote: »

The biggest problem would be the probe. ... simple, coarse, "science fair" grade imager ... The Prop 2 would be a different story.

Science fair is fine, its still an imager. Better start now to be ready for prop 2!

Cluso99 wrote: »

the biggest cost by far would be the approval cycle

No need to start at approval and work backwards to proof of concept. It's just for geek-coolness factor. Leave "approval cycle" for the model 2.

frank freedman wrote: »

The complexity for an entire system on a prop or propII would be virtually undoable.

There you have it, the gaunlet has been thrown down.

Now that someone has declared it undoable, I expect there will be results shortly

RichK

Greetings,
I work at GE Ultrasound. Here's what we have. Front end samples probe at 50 MHz or so. Probe contains 128-256 elements. Small handhelds typically contain 32 or more elements. Probe transmit freq is from 1.5 MHz to 18 MHz or so. Front end processing is typically 16 bit A/Ds feeding FPGAs and/or DSPs or GPUs. Beamformers, HW or SW, generate inter-element delays in the ns range to generate a single vector dynamically focused along the direction the vector is received from. Same thing for the xmitter, it is focused too. The image displayed typically consists of 100 to 500 vectors fired consecutively. Frame rate for firing is 3 fps or so for large deep images (liver, ...) to 120 for neo-natal and cardiac images. Single image playback allows seeing the higher rate data. It's simpler that it sounds...

The only point in our product for a prop now would be in the front panel where the buttons, keyboard, touchpanel etc is located. Currently I think there are a couple PICs up there. Used to be an 8051 variant. Maybe in the power supply for control/monitoring.

I would be more interested in transducers for a depth finder on my boat. Transmit freqs in the 80-200 KHz range. Anyone know of a waterproof xducer that is usable? I used to work at Raytheon on ship board sonar. The xducers there were shoebox to refrigerator sized and would not mount very well on my 16' boat.

Rich.