Crazy pondering of the day

I recently came across this chip:

It's able to boost from 20mV up to 5v at 3mA and could theoretically be used to charge a capacitor from nerve impulses.

Add a Kinetis KL02 (which can run on 0.036mA) and you have a complete system that can run completely on nerve impulses (or other low-current power supplies).

As for practical uses - well, human nerve impulses are generally about 80mV and not a lot of current is needed to activate a nerve cell and make it take over transmission of the impulse.

With some development it could even be possible to swap out a few neural circuits for software implementations - you just need a mathematical model of the whole circuit and the right interfaces.


  • nerves work with so little current it's not even measured in current but in actual charges. Like single-digit nC and that's on the upper end of the scale. Working with pA and fA would be nothing strange when dealing with nerve impulses. That's orders of magnitude below what you need. Besides you'd need to rectify those short spikes which is very hard at such low voltages. And even if you manage to do everything, you remove charge from the nerve itself which is quite a bad thing to do on a long run.
  • So power it from body heat instead, it's still small enough and low power enough that you can do a hell of a lot
  • to power anything from heat, you need a temperature difference, the bigger, the better. Bad luck in case of a body as it tries its best to keep temperatures consistent.
  • no you don't, not necessarily. I'm working on a thermocell that doesn't need a temp differential, just runs on ambient heat. I'm working to see if can make it efficient enough to be worth adding to an implant to harvest energy from your body. but that's several months away.
  • Wait, please explain. That's like saying that you have an LED without any  voltage drop. Unless I'm missing something, you're breaking the second law of thermo here.  UnUnless Im
  • @ElectricFeel Explain the thermocell?
  • I understand a thermocell, at least I believe I do.
    It's just with my understanding it's impossible to run with ambient heat, that's zero point energy. It's not usable.
    If you were using temperature cycles that would make more sense, but it's still a temperature gradient; just dt instead of dxyz.
  • The videos I've seen do seem to supply some sort of a temperature differential, even if the user is unaware of it. @Chironex When you get a chance, try submersing the cell in a heated, non-conductive fluid, like oil, and see if any power is generated, and how long the generation lasts at a constant temperature. If it generates power for an arbitrary amount of time, and the meter isn't responsible for some fluke, it's a distinct possibility that this thermocell is using a different mechanic entirely to convert heat into energy.
  • From what I understand of it it works not like a peltier but more like a solar cell. Only instead of using the visible spectrum it uses heat. There's a bunch of variations of it and I keep stumbling across weirdness when I'm working with the super capacitors. The most basic thermo cell I've seen runs on a 4 layer system. Copper, chrome oxide zinc oxide and then magnesium on top. Basically 2 current collectors then an n type and a ptype material in the middle just like a solar cell. But light doesn't affect it, heat does. There was some other thing where ions passing over graphene sheets can generate a voltage. So you can generate power by using salt water just washing over the thing. And the warmer it is the better it works. I could talk for an hour just on this. There's a lot to it.
  • Do you have any links to papers?
  • This is the guys site who made the first one. people online have been improving on it since.  LINK 
    there are some papers but I don't have time to dig through and find them atm. I'll see if i can find it later.
  • edited October 2015
    Heat isn't transmitted that way. Radiation can move through a vacuum and through various objects and interacts with ordinary matter on when it comes into contact with it (and only sometimes). Heat is something intrinsic to matter itself, and can only be transferred through contact, it can't move through a vacuum and it's transfer is slow when compared to radiation. However, the movement of particles (heat) causes the emission of radiation in the infrared spectrum (radiation). According to the website provided the "monothermal" claims to work off this infrared radiation. 

    So, to figure out how much power such a battery could get from the human body we can use the Stefan-Boltzmann law and treat the human body as an ideal black-body with temperature 310K (body temp)–Boltzmann_law), we have a total power generation of:

    Power = Stefan-boltzmann-constant * (absolute temperature)^4 = 5.68*10^-8*(310)^4 =~ 500 Watts per square meter in a perfect world.

    If you wanted to make an implant for heat harvesting and wanted to make it on the large size of reasonable (say 10cmx10cm), you would have 0.1m * 0.1m = 0.01 square meters, which gives you a potential 5 Watts of power. Yay! That can power all sorts of cool stuff!

    However, this only works assuming the place you are collecting energy is cooler than the body emitting the radiation (see If you want to implant something inside the body, heat is going to be transferred from the body to the energy collecting cell until the cell becomes the same temperature as the body. So either you expend energy to cool the system (which is by definition greater than the amount of energy you will collect, because the device is being heated up both by the infrared radiation it's collecting and by being in contact with the body), or your system doesn't generate any energy. Either way, your system generates either zero or less than zero power. You can, however, cool the system by exposing it to the air so it's constantly being cooled, but this requires it being outside the body, in which case you should just use a solar cell, because the energy of visible light is on the order of 1-1000 times higher than that of infrared radiation ( This technology is not a viable candidate for generating power for implants if you want to put it inside the body, **even in a perfectly ideal world**.
  • @chironex you certainly do violate the 2nd law of thermodynamic. and while it's not exactly impossible to do so , it's still extremly difficult on a macroscopic scale to utilize the tiny thermal fluctuations and turn it into something usable. Chances are you'r running some sort of unintended chemical side reaction. Like magnesium, zinc and the likes make great batteries, especially in combination with salty water and air. Given you work with carbon based stuff, you may have build an aluminum-air or zinc-air battery without knowing it.(aluminuim sheet in saltwater and wet charcoal on top of it ( given aluminum and charcoal don't touch) and you have a nice battery. Useful if you survive a plane crash on some island and need to power/charge.
  • Maybe but I doubt it. I think it just works on a less than obvious mechanism. I say that since there are many people who've tested the thing and they all get good results. For example theres one cell im working on. It runs on zinc sulfide. It's a conductive transperant layer of graphene on plastic with a thin layer of zinc suflfide ontop. Then 2 current collectors, one on either side of the plastic, not front and back, both on the same side, left and right. The thin only works if the zinc sulfide is wet. Niether current collector needs to be wet and there's no  applied source of energy. Heat doesn't affect it, light doesn't affect it, yet it's generating a small amount of power. More if the water is moved around a little bit.I have no clue wtf is going on so i'll keep exploring and see if there's something to this. but this is what I mean. There's all these weird ways to make power, they don't violate thermodynamics, they just must work in some unexplored mechanism. I'll look more into the thermocell. If it is an air battery i'll set it up in my vacuum setup and see if it runs without air. Of course i'll have to rule out chemical reaction but it'll be interesting if I can.
  • What purpose does the zinc sulfide serve? I know Robert Murray-smith used straight graphene on a sheet, and got results from that.

    On another note, how do you intend to heat the thermocell when it's in a vacuum?
  • I believe he means to see if it is an air battery @thegreyknight
  • i was gonna tape it to the side of the chamber and heat it directly through conduction. The plan, tape it to the side of the chamber with volt meter connected and a thermocouple tape to it as well, evacuate the air, backfil with argon, repeat twice to be sure all oxygen is gone. Then once that's done, measure voltage. apply heat, measure again, cool it down, measure again. 
  • im not sure. I think it's the source of ions. It's photoactive so it may act to induce catalysis of water to H+ and OH- which when those ions pass over the graphene they generate a charge. But that's only a specualtion. I've run all of 2 basic tests so it needs a lot more work. Also i used De ionized water so the sulfide must play a role or no charge should have been generated. I'll try with other things and better graphene and all kinds of stuff and let u know how it goes but im really busy atm so it'll take a bit to test all of this.
  • @chironex one good way to prove it's actually working is to record the energy created by your device, and compare it to the theoretically maximum amount of physical and chemical energy inside your system, including the air as oxidant if in question.
  • Getting back to original subject (apologies for cutting off the other discussion - perhaps a new thread is appropriate?):

    Are there any semiconductors that can be used to make transistors which would trigger from the minute currents generated by human neurons? If there are, an unpowered implant could be built to run inside the nervous system and take over certain functions (pacemaker circuits and oscillation come to mind here). 
  • No. First you need to understand one of the differences between how neurons run on "electricity" and how computer components run on "electricity". Neurons primarily use potassium ions that flow outward into the extra cellular space through potassium leak channels to maintain their resting potential of around-80mV, and an influx of sodium through voltage-gated sodium channels. The entire process involves only positive ions. Computer components, on the other hand (excluding the new exotic memristor) run off electrons. Electrons do not flow across the neuronal membrane during an action potential, so there is no usable power source for an electronic component. Even if there was, the current is minuscule (in the pA-nA range per cm^2), and not usable.
  • @bciuser is correct about the way everything works, in terms of why you can't just plug a wire into a nerve and expect it to generate usable power.

    Correct me if I'm wrong @garethnelsonuk, but you're talking about using silicon-based electronics to interface and introduce or change an input action potential without amplifying the signal, and then output the new signal to another nerve.
  • Not silicon based, i'm aware that's not really feasible - i'm talking about whether it could be feasible with some alternate substrate.

    A transistor that could operate in a similar and compatible manner to neurons would open up a lot of possibilities.
  • Note also that there's a reason I called this thread "CRAZY pondering of the day" - i'm fully aware it might just be completely silly.
  • edited October 2015
    OK, So I'm going to call the idea of harvesting energy from nerve impulses crazy and silly beyond belief. BUT, the idea of using passive components to record and interact with the nervous system rather than active components like electrodes is a very very good idea, and one a bunch of scientists have had. 

    Here is an article (one of the cooler ones) that describes the use of organic transistors to record neural impulses on the surface of the brain, and how they significantly outperform (by at least 2 orders of magnitude) both traditional surface electrodes (ECoG) and implanted electrodes (MEAs). 

  • On another note, the material PEDOT:PSS is capable of conducting both ionic and electronic current.
  • I'm beginning to think that graphene is also in that category in terms of the sorts of charges it can conduct. But as usual, more testing needed.
  • "I never asked for this."

    Do PEDOT electrodes show any signs of causing rejection when implanted in nerves? Or Glial cell build up?
  • You don't have to worry about rejection if you're Adam Jensen.
  • True, but here in real life we do have to worry :)
Sign In or Register to comment.