My colleague, Steve Taranovich, just wrote a blog on implantable medical devices: Let’s Chat: Medical Implantable Electronics Integration. He examined the challenges in designing such devices and the challenges related to the ICs that would be required to make such devices practical. I like to push into that topic a bit more.
We'll make a few assumptions to simplify this discussion. We'll add a caveat first. We'll go a bit beyond implantable and consider devices that are temporarily put in the body. This would include the miniature cameras that are packaged in an oversized capsule and swallowed.
Any device that is implantable or introduced into the body is there to monitor some aspect of the biological system that is human. So, there are parameters like temperature, pressure, flow rate, the presence of certain gases or chemicals, the pH level… the list goes on. That means we will need lots of sensors. They must be tiny, accurate, and low-power. Then we need to process these signals and establish a communications link to the outside world. Obviously, we're not thinking in terms of an RS232 or USB link. Instead, some sort of wireless link would be used.
What functionality do we need to have in an IC to do this?
- Op amps — very low bias and offset; very low noise; temperature stable
- Instrumentation amps — as above
- Transimpedance amps — as above
- Programmable gain amp — as above
- Voltage reference(s) — accurate and temperature stable
- ADC — 12 to 24 bits (so maybe delta-sigma)
- DAC — 12 to 18 bits
- Real-time clock — accurate and temperature stable
- Microcontroller unit
- Memory — volatile and nonvolatile
- RF transceiver
- Inductive loop transceiver
- Voltage regulators — switchers and linear
- Watchdog timer/supply voltage monitor
- Optional — the actual sensor/transducer
Steve looked at the big picture for making devices to assist (or even replace) organs. I tend to look more at the “nuts-and-bolts” of the design. I'm considering whether we can do everything on my list in the form of one IC.
Can it be done? Can we make devices that draw miniscule amounts of power and still do what's needed? Can we design a wireless (RF or simple inductive) link to pass the data through the body tissue? Let us know. Comment here or come to our chat session next week.
If you aren't already registered on Planet Analog, please register in advance to speed things up. It just takes a few minutes.