Electronic Subdermal Implant Tech

A few ideas or approaches for subdermal implantation of electronics I would like to discuss with you more experienced and critical people here.

1. Separating power supply from actual control circuitry
By (physically) separating away the battery, charging coil and ~ circuitry from the control circuitry and the actual implant, the overall thickness of the implant can be drastically lowered (obviously at cost of the area). Ideally the power supply would be placed in a better accessible position than the implant that can accommodate for a larger implant. We however don't have technology for wires that span movable tissue, so there really aren't any places where this would work at the moment.

2. Bluetooth (4.0) Connectivity
Initially I was against the idea of implanting a Bluetooth module along a microprocessor, both because of relatively large energy drains and physical size of traditional modules.
The Bluetooth 4.0 "Low Energy" Standard (BLE/"Bluetooth smart") however solves both problems nicely, a TDI BLE module is just 4x5x1mm in size (including antenna!). The module is a QFN-package and more or less impossible to hand solder though (70 pins on less than an square centimeter iirc).
BLE is only set up for distances of about 10m and I don't know how bad RF-through-skin really is, but I do believe there should be some range (think mobile phone in pocket) left outside the body.

Two other chipsets are the RFDuino and BlueRadios nBlue. The interesting part about both these is that they feature an on-board programmable microcontroller to do all the bluetooth stuff. The RFDuino can be programmed like a usual arduino, whilst the nBlue can only be programmed with a (very expensive) hardware debugger kit. However the nBlue features a lot of configuration options in the firmware, which even allow to set output pins of the package at runtime, over the air.
Why is this important? Well... one awesome feature of bluetooth would be being able to program the implant over the air. The RFDuino doesn't support this, and the nBlue (for us) doesn't really either, but both can be used to program another microcontroller. They just need to set the reset pin on the ATMega (or whatever really) and then forward the UART data they receive and we could have an OTA-programmable implant, that can even talk to smartphones.

3. Flexible PCBs and Coating
Flexible PCBs could be a great way to shrink size of implants and make larger ones possible at the same time.
By using SMD ICs and hardware on FPCBs and embedding these in flexible silicone, you could place implants in slightly bent areas (like the forearm) without creating extremely protruding implants.
Combine this with additional silicone to get a "cool" shape and you get a glowing arc reactor! (ok, agreeably there are better things to put this to work with).

So... what do you think?


  • what i think. 1. 2. are legit ideas. Flex PCBs aren't really what you want. The areas carrying chips need to be reinforced anyway, so you can go with regular PCBs in first place
  • Well, even a partially flexible PCB can be worth a lot of space (versus wires). And the industrial FPCBs are a lot thinner than "traditional PCBs" ones:

    although of course one can only dream about the miracles of factory-line soldering and electronics ^^
  • Point is, flexible pcb's are, well flexible. But that's not unlimited flexibility. Repeated bending can lead to fatigue and failure on the long run. The smaller the bend radius, the faster it happens. Starting with tiny cracks in the copper layer causing unreliable connections at certain bend-angles up to complete mechanical failure of the carrier material.
    If all you want to do is to fold the pcb up to several layers, then it might be an option. But the saved volume is hardly worth it, compared to simply sandwiching regular thin pcb's. You'd need a mechanically rigid shell around the flex PCB,too.
  • Well, I meant flexible enough to withstand for example the slight curvature of the arm and allow for some shock absorption instead of breaking a thin PCB.
    And for casing, if you basically dip your design into slightly flexible silicone, wouldn't that work?
  • even if your pcb is flexible, your chips and the pins on the chips are not. So wherever you have ICs on the pcb, you need reinforcement. Otherwise you put mechanical stress on the pins/pads which will lead to broken contacts over time. Given an implant is supposed to be as small as possible, the chip density usually very high. Long story short, you avoid a lot of potential problems by simply using a rigid PCB, preferably in a rigid shell.
  • edited August 2014
    ^this assumes u weren't also using flexible connections and ics. which could both remedy the problem and shrink size considerably. 

    EDIT: although it would raise the difficulty significantly when it comes to fabrication
  • I am not aware of any flexible ic or solder connection (except for some experimental silver silicones). But nothing that'd be easily available or well tested.
  • I was thinking a graphene based connection and chips built directly into the board instead of the big bulky housing
  • the chips could still be the same but a graphene trace would make it flexible without the need for solder although an adhesive would be needed to keep the chip on the board which could still fail eventually I presume but that also assumes that the implant is in a fairly active region and is being heavily flexed.so unless you're putting something in a very active spot a flexible pcb could just make implants more comfortable and sleek if anything.
  • Is there a particular part of your body that you would classify as a 'non active region'?
  • As in side of the bicep as apposed to the inner elbow. Anywhere active bending isn't happening.
  • I guess I just do a lot of lifting things. Brah, do you even lift? 

    srrsly tho, the body is a soft, flexy dynamic thing. There are a really low amount of non active regions. Yes, the bicep is less active than the elbow, per se. but compared to like, a digital camera, all the bits on your body are way more active than anything like that. On a scale of materials, from one to bendy, camera lives down at 3, a step above rocking chair and cider block. Biceps and elbow both live at 9. Not quite tongue, but but up there...
  • edited August 2014
    So, after a few days of research I believe I have found the closest-to-perfect bluetooth module for this sort of use.

    The BlueGiga BLE113 (Datasheet) basically has most everything you could wish for. It's ~16x10x2mm, which is notably larger than the other modules above, but is still a very small size (and actually easier to solder). It features an integrated antenna and is has GPIO, PWM, I2C/TWI, SPI and UART output.
    The internal microcontroller can be programmed using their own scripting language (no need for the $3000 IAR workbench and wrangling C all day), which provides a relatively high-level access to all the bluetooth stuff as well as the peripheral device managing.
    The I2C/TWI Bus can handle multiple devices but doesn't need an extra pin per device, so you only need two pins for LED controls and a timekeeping chip.
    The BLE113 supports Over-the-Air-BGScript updates (the OTA updater has to be included in BGScript though), so you could reconfigure and reprogram a device as often as you want to.

    There are a 128k and a 256k flash version available, if you want to support OTA and not connect an external (SPI) flash chip, make sure to get the BLE113-A-M256K. Mouser sells both for 11,16€.

    To program them the first time (and every time you leave out/screw up the OTA part in the BGScript) you need a TI CC Debugger (37,72€), I was told that this knockoff would probably do the same though.

    There is also a breakout board available here for $6 (SMD parts not required if you are just interested in getting the BLE pins onto your breadboard), but there are only 5 left; if you are late the schematic is available for free.

    I would recommend (and will myself) buying two BLE113-256Ks, so you can test OTA updates on a non-implanted device and check if they continue to support OTA before pushing that update out to the implanted one (possibly bricking the whole implant forever).
  • If the device used a low enough power would it be possible to draw the electricity from the body? In what ways might that be possible? 
  • In theory it would be possible (the BLE113 is supposed to run off a coin cell battery alone) but as soon as you add anything interesting to the circuit (LED, vibration motor) you are going to drain waaaay too much power (even if you manage to find a usable energy source in the first place)
  • tingeling neurons with current is pretty low power. so that's certainly an option. ways cooler than vibration motors ,too.
  • at this point tho, how do-able is tingling neurons? how safe?
  • likely very doable. it requires a very tiny curent and the pins could easily be made to be biocompatible and as long as they seal with the rest of the case you're good.

  • did it with non-implanted circuits and regular hypodermic needles. quite easy and with the right electrodes and settings it's pretty safe. even in worst case you would only damage a bit of skin tissue somewhere on your arm or so. nothing that would affect your life.
  • huh, thats pretty cool. how much of a sensation do you notice? also how little power would it draw? seems like if not much power and a fairly noticeable sensation, could be the go-to method
  • sensation depends on your actual settings and electrode geometry, but you can calculate 100 to 500μW for a very strong sensation. Placing the electrodes closer will allow you to reduce power. But even with that level of power you can easily build an alarm clock.
  • Hm... the idea of having it set to an alarm could be very interesting as far as providing an innate sense of time. You know the effect when you have an alarm you wake up to every day and after a while you just naturally wake up without it / before it? It would be interesting to explore an internal device to improve that sense of time.

    Either way, if a minuscule current draws less current than a vibration motor while providing an equal sense of 'feedback', this sounds like a much better approach!
  • The alarm clock sounds like a phenomenal idea! you just would want to make sure you can adjust the time it goes off. and make sure it stays accurate.

    As far as improving your "internal clock", having a small buzz every 1 hour would probably do it. enough to notice if you are awake, but not enough to wake you. then have a big buzz for the actual alarm you use. so that it can wake you up.
  • As for the topic of power efficency vs battery capacity, some of the new generation heart pacemakers are being developed with a selfwinding mechanism, like ones used in some better brands of watches. Thanks to those, movements of the body produce energy which replenishes pacemaker's batteries.

    I believe this technology could be great inside most of implants making them pretty much independent from external charging and letting you worry a bit less about power drain.
  • I just recently heard about something that was quite impressive in the realm of power generation.

    It's thin, it's flexible, it could be bio-proofed, and the power output for a relatively small quantity was impressive. If we can get a reliable, relatively cheap method of creating or sourcing the required materials, it might even be something DIY feasible. 
  • edited September 2014
    Kinetic watch tech is an interesting idea.  They've been making kinetic watches for years (my father wore one when I was a kid).  It isn't difficult, could be easily encapsulated, and is small enough to implant in any of a number of places.

    The only thing is that they tend to have a small vibration associated with them.  I'm sure you would get used to feeling it, but I wonder if it could be implanted in such a way (location) as to confer a beneficial sensation.  The only thing I can think of is to possibly give a better sense of inertial changes - speed up or slow down and there is some vibration.
  • http://www.gizmag.com/durr-faceless-watch/30748/
    This watch vibrates at five minute intervals and the wearer's sense of time is brought to attention.
    My smart watch vibrates every fifteen minutes and I definitely notice my sense of time and activity.
Sign In or Register to comment.