Hi, it's thermoelectric generator. Time again. those pesky devices that? well, I've done a uh, debunking video on this before. You might remember this one way back.

Uh, the Matrix powered Smartwatch? uh, for example, where I ran the numbers on that one. Never charge your smart watch again because it's got a Peltier device in it. It's got a thermoelectric generator, one of these uh, Cbec effect uh devices and oh, look, isn't it wonderful? and well, it's just not. It's just no, it's a gimmick as I showed in my video.

Now, this video comes about because there's a ton of misconceptions about there about thermoelectric generator devices. and every time somebody reads a press release for them, or in this particular case, we're going to talk about a scientific paper you know, extolling the virtues of how they're going to. You know, empower smart, wearable devices, the Internet of things, and all this, uh, sort of stuff. There's so much misconception out there, so I thought it's worthy to do a video now.

This one comes as a response to one of my favorite Youtubers out there at the moment are Anton Petrov from what To Math and I'll link in Antos Art page here. He's on Odyssey as well, so definitely, uh, subscribe to him and the Eev blog down on Odyssey. I've I think I've hit 60 000 subscribers now. Yeah, anyway, really, I'm telling you, do yourself a favor and check out his Uh channel.

So this is kind of like a shout out video too for one of my favorite, uh, Youtubers. Mostly like astronomy, space related stuff, but in this particular case, any type of scientific research. What he takes is like the latest scientific Uh news, and he links to Uh the papers and he looks through the papers and explains the papers, gives his take on it and you know all that sort of stuff and it's absolutely fantastic. Probably one of the only channels at the moment that I watch.

Like every single video. Anton talks about this paper here in Scientific Advances High Performance Wearable Thermoelectric Generator with Self-healing Recycling and Lego-like configuring Capabilities All the authors there and it's you know, it's fairly recent. Thermoelectric generators are an excellent candidate for power, wearable electronics, and the Internet of Things grown. Yeah, half my audience just grown due to their capability of directly converting heat to electrical energy.

Here we report a high Performance Wearable Tag, which is a thermoelectric generator with superior stretchability, self-healing recycling, and lego-like configurability by combining modular thermoelectric chips dynamic, a covalent polyum, poly polyamine yeah, and flowable liquid or liquid metal terminator. electrical wiring. This is where it starts. Uh, electrical wiring in a mechanical architecture of soft motherboard? Rigid.

They're basically talking about flex. Uh, and not flex Pcb. In this case, it's a different type of material, but anyway, won't get into the details anyway. they reckon.

yeah, it's going to have a wide range of applications and Internet of things and everything. and well, no, sorry, it's not sorry to burst your bubble now. Look, this may actually be like a really good innovation. It sounds really cool.

It's like a flexible thing. as you can see here. they've developed this flexible prototype that can sit on the skin. so it's got a self-healing thing so that when you got it on your body, it's going to bend all around the place, but basically stick it on your body and it converts the warm body temperature.

Hopefully you know, roughly 37 degrees celsius now that Fahrenheit rubbish and whatever the ambient temperature is here. But unfortunately, if it happens to be like 37 degrees outside and matches your body temperature, well, you're going to get zero out of it. But anyway, um, yeah, if you've got a cooler, uh, environment than your skin, then it's going to generate some level of power. So anyway, they claim that this is like a better output capability.

and of course, it's self-healing so it can't damage itself when it flexes and all sorts of stuff like that, right? Very cool. So it's probably a really good, uh, advance. But is it practical? So we're gonna run the numbers How the human body? Yes, it generates like hundreds of watts. Like just sitting here.

I'm like 150 watts of, uh, power. Something like that. I'm dissipating that sort of, you know, power level. So there's lots of power there to be harvested.

Um, for these types of thermoelectric generators. But are they going to be practical? And I've already told you the answer? And the other thing in the videos. Anton talks about energy when he's actually referring to voltage or power. and I've actually done a video on this.

I'll link it in down below as well, where I, um, explain the difference between voltage, power, and energy. And a lot of people get this in the industry. or I'm even, uh, guilty of like using the wrong terminology myself. It just, you know, rattle off the top of my head and I say energy when I mean power or vice versa.

So anyway, it's just really easy to misspeak on voltage, power, and energy. A record high open circuit voltage among flexible ones at one volt per square centimeter. And uh, here's where a lot of people in all sorts of industries. Uh, when they're talking about these thermoelectric generators, when they're talking about solar cells, they're talking about other energy harvesting devices.

Is that they think that, oh, it generates a volt. Well, my Intel processor can work. It works at 0.9 volts or something. It can power an Intel processor.

Well, no, unfortunately, it can't. You've got to watch this video or go through the calculations shortly. Okay, you might be able to. Uh, the thing about these is that they are actually modular and you can can actually, uh, string them together.

We'll have a look at the photo in a minute. but uh, yeah, you can like string a couple of these in series and you can get higher voltages. So when you're talking about voltage like this, you have to talk about like voltage. In terms of right, that's the minimum amount of voltage required to actually do some work on electronics.

When you like, anything under 0.5 volts is basically not usable with modern electronics. The thing? You can't start switching on silicon gates and things like that. there are energy harvesting chips and those energy harvesting chips. They might require like a threshold voltage of like 0.5 volts to start up, For example, to start up a Dc to Dc converter.

So then they convert those voltages to higher voltages. And if you're talking about like a Dc to Dc converter that starts at 0.5 volts and you might want to boost it to say 3.3 volts, which is a typical voltage to run you know, modern electronics with for example, then well, those Dc-dc converters. They're not very efficient. they take their own power and all sorts of stuff, so it becomes a bit useless when your Uh converter itself actually takes more power than the device you're trying to actually power.

And we can actually start having lookers for some energy harvesting uh chips here. And some of them can actually work down to very low voltages like you know, tens of millivolts and stuff like that. And if you go to analog devices here, they actually tell you I boost converters that operate from as little as 20 millivolts. But aha, there's a trap.

They can operate down to that, but they can't start up at that sort of voltage. So you know there's different energy harvesting chips here. Let's just have a look at this one. Ultra low power energy harvesting so it can work down to really low voltages.

Yeah, it can operate from 80 millivolts to 3.3 volts input. But the problem is. um, its startup voltage needs to be 380 millivolts minimum. You know energy harvesting is a thing.

You can actually get specialized chips to actually do this sort of stuff. So we're actually going to have a look at the numbers and see what we can actually pull out of these new whiz-bang devices. But the problem is okay. You don't even have to read further than the abstract to know you've already come a gutter.

A record. High Open Circuit voltage. First problem: Open circuit voltage that's not under load. That's open circuit voltage.

And we'll look at the Pv curves. the power versus voltage curves in a sec. Uh is one volt per square centimeter at a temperature difference of 95 k, 95 kelvin. That's 95 degrees temperature difference.

Um, if you've got a 95 degrees temperature difference, uh, between your skin and so and the environment, you're probably dead. So yeah, you're going to come against us. So right there, we don't even need to read further than that to know. Well, yeah, this is not great.

So here's some actual performance curves and data provided in that article. uh with Linkedin down below and this is how they'll typically, uh, characterize and test prototypes like this. They'll have a heater down the bottom and they'll have the uh, cold junction on top. They'll have the, you know and then the ambient, uh, temperature out there and they'll put a load on it and a Voltmeter and they'll get some performance uh curves on there.

And I won't go into how uh, Peltier devices work in the Seabeck Effect and all that. I've done that in the previous video, but already. uh yeah, let's look at p out That's power output here. Micro watts per square centimeter.

So let's work with the one square centimeter thing which they show on the finger in that you know that's what they're hyping up Uh, as can potentially power these. uh, you know, wearable devices and smart watches and things like that. We're talking like 10 microwatts, 10 microwatts. But what we want over here is your kicker.

Okay, here's what's called a Pv curve and you'll get these for these. uh, thermoelectric Generator Peltier devices. You also get them for solar cells and stuff like that and they typically have a response. or at least the Peltiers do.

Solar ones look a little bit different, they're sort of more linear and then sort of taper off at the end. But this is a typical curve for a thermoelectric, uh generator like this, So it's power output in microwatts per square centimeters versus the load voltage. the output voltage which is also across the load if you have one of millivolts per square centimeters and sure enough, it can get close to a volt. So I'll give them that.

In the article, it says that they get basically this is where they're claiming big headline abstract in their thing says we can get one volt out at 95 at 93 kelvin difference or whatever it was. Yeah, sure enough, I don't like how they did these colors. They should have chosen different colors. But anyway, there, there's the curve there.

It is okay. But as I said, open circuit voltage. You're not powering any load at all. Now with, uh, any solar cell or thermoelectric generator like this, you'll have what's called as a maximum power point.

an Mpp, and a lot of these energy harvesting chips will actually have maximum power point trackers and you may have actually heard about those. For your home solar systems and things like that, it's more efficient. It'll actually try and track. I won't explain the concept, but basically it tries to operate around this maximum performance right at the top of this curve.

Here, this is where you're going to get your maximum power output. So as you increase the voltage like this, the available power actually drops. and here it is the power output. At one volt, you get no power.

Zero zero power out of this thing. So yeah, you're not going to get your 1 volt. But hey, look at this. You might get 500 millivolts And as we saw, you can get energy harvesting chips that can do this.

and you start to get in the realm where it can turn on Pn junctions and do useful work and stuff like that. So it's possible. But once again, the power output 18 microwatts. Hmm.

But again, the kicker at 93 kelvin. Okay, that's the point where you're dead. Um, so let's look at say, your body's at 37 degrees. Let's just assume you know it's it can.

It's all perfect. It's 37 degrees. There's no other losses on one side of the Peltier device, and you're you know, room temperature 20 degrees on the other. So you've got like a 17 degrees difference.

So here's our first gray curve here. Okay, because this is our last blue one. This is our first gray, so that would correspond to your 31 Kelvin here. Okay, so 31 degrees.

so you've only got like a six degree difference. So you're freezing right? It's six degrees. Don't know what that is in Bloody Fahrenheit for your yanks, but that's all right. Six degrees.

Trees. Freezing never gets six degrees here in Sydney. So right off the bat, even though you're freezing your ass off, you're only going to get two micro watts at under 200 millivolts. And that's not enough really.

To start up, you know they start talking 300 to 500 millivolts start up for any sort of energy harvesting thing. It's not going to power anything directly, but the thing is, you can put these in series. Okay, so yeah, no worries, right? But yeah, right off the bat when we're talking about this little thing here on your finger. even if you strung a couple of these in series, each one.

even at like you're freezing your ass off, you still only get based on their actual measured data. You're still only getting a maximum of two microwatts per square centimeter. So you string enough of them together, getting a high enough voltage to make it useful. And uh, yeah, you're still only going to get two, four, six, eight.

You're down in the order of 10 microwatts. So what can 10 microwatts power? Well, let's go to the bench quickly and we'll measure the power of something like this. So how much power does something like this? Uh, ultra low power solar powered Fx 260 calculator? I've done a review on this and an experiments video I might have to link. Geez, how many videos am I linking in? Anyway, there you go.

There's a little shot inside. Uh, we've got like a solar cell. There's no battery whatsoever. It's powered just from that.

So I've hooked into wires and we can see that this is the voltage, uh, measurement coming from the solar cell. About uh, 2.7 volts. Uh, give or take. Then we can also measure the current as well.

and here it is. Can't really make heads or tails of that. But if we put on our trend chart, check it out and clear the readings and there you go. just sitting there at Uh.

2.7 volts. You know we're talking like 30 microamps. Something like that. 30 microamps times 2.7 volts.

put in engineering notation about 81 microwatts for this, uh, little tiny solar power calculator. But of course, it's operating at a higher voltage. You can't actually operate lower than that. And to demonstrate that, let's actually cover the solar cell and you'll see the voltage drop.

Let's get it to a point where yeah, around about a volt. Once it gets to about eight, you'll start to see those digits fade a bit. Can operate down at point Eight. That's kind of like typical of this sort of, um, you know, processor asic node as they call them at the silicon.

uh level. you know, point A. under that it's going to start to really fade. Geez, it must be getting some light through there.

So there you go. 0.8 volts. It's It's barely on, it still actually does work, and we can actually, uh, rescale that. There you go.

You know, two or three microamps? Something like that. Not too bad. And there's actually a, uh, interesting thing here. you'll notice I've put all eights there on the display.

so we're turning on all the segments. You'll notice that it's not really using any more than that. Sort of like, oh, you know, two and a half microamps. Uh, it.

It really doesn't change. I'll clear it. There you go, and it doesn't use any less power. But if we go back to our full voltage and we'll auto scale that.

Okay, so we're back to 10 microamps now at 2.7 volts If I do, or if I turn on all eights on the display down there, you'll notice that it's actually drawing more power now because it's got all the segments. yet it doesn't do that at about uh, you know it doesn't draw more at with having more Lcd segments on at the lower voltage. That's a little interesting aside, probably warrants an investigation video that's interesting. Anyway, it can actually still perform calculations.

Let's do 69 factorial there. Oh, you can do it. Oh, there we go. It just did it.

It just did it at .73 Oh crazy. So I'm going to say the lowest we can absolute lowest we can get this working at is our 2 microamps at 0.75 volts and 2 microamps times A 0.75 volts. We're talking. Oh, 1.5 microwatts.

There you go. So a calculator like this ultra low power custom Ac card design it's you know you can actually do useful work at. well, you know, one and a half micrograms. a couple of microamps to uh actually do some useful work.

but certainly you're not going to get anywhere near this with any sort of a Smartwatch. But as you saw as the voltage went up, well, so does your current. So yeah, like as I said, like 10 microamps you can start doing, you know, useful kind of work. And for something resembling, uh, some sort of like wearable smart watcher type device, we've got the Efm32 Gecko uh processor here really ultra low, uh, power stuff and a a graphical E-ink display which you might find in a modern, uh, Smartwatch.

Of course they take no power when you have a static image on them. They only take power when you actually update the information on them. Very cool, uh. technology? Well, what does this thing take just doing a simple clock like that? It's not doing anything, it's not transmitting, it's not sensing anything else.

Well, there you go in the order of 500 microamp pulses at 3 volts. So we're talking like, you know, one and a half milliwatt peaks there. But of course, uh, you know, in a typical product of course. uh, the actual energy harvesting device wouldn't actually supply this, uh, peak power.

It's only got to supply the average power. That's when your, uh, decoupling capacitors actually handle, uh, the peaks and stuff like that. So you got to have some sort of storage. So you know the average figure and stuff like that.

But you know, still like just doing a basic clock like that. The process got got to wake up from sleep every second it's got to take. you know, gulp. 500 micrograms.

It's got to update the you know, the display and do stuff. and then when it's not taking those bursts of current when it's down in sleep mode. yeah, you know we're talking a couple of micrograms. This is a typical sleep current of modern ultra low power micros.

You can get like in the order of like a half a microamp. You know, hundreds of nano amps. but you know, like in deep sleep. But then it's got to wake up and it's got to actually do something.

So your averages have to be like much higher than that. So once again, multiply that by three volts and you got like, you know, six microwatts. You're approaching that 10 micro watts again before you can even do even the most rudimentary stuff. So anyway, I hope those measurements give you some sort of feel for what modern wearable devices, even ultra low power stuff that do something incredibly simple like this clock ear can take you know you need, like tens of microwatts.

And then you start talking hundreds of microwatts before you start doing any smart stuff in quote marks. Uh yeah, it's just God and these things generate 100 nanowatts. Give me a break. So here's another graphic that they provide.

and this is actual test results from actually putting it on your skin like that. Okay, so that's really cool. Yeah, they show. Okay, 35 degrees, 35.8 for the skin.

As I said, it's going to be a bit lower than your 37. typically. that's your internal temperature and you've got a 25 degree ambient thing. So nice realistic test.

There's a thermal image of the prototype and stuff like that, and well, how what power output are we getting? Uh yeah, we're not. even. uh, micro watts. We're in the order of nanowatts.

We're talking a hundred nanowatts That's not even gonna power the lowest power calculator, let alone a regular digital. Like just a regular casio. You know, 1980s digital watch? It's not going to even power that. And then interestingly, it looks like they get some data on the flex and stuff like that.

So you know this is actually a really good research. You know it's it's great. I love the look and then the thermal shots as they like bend it and stuff like that. Anyway, really cool, but you're starting to see that the hype actually from this uh, paper that oh, we're gonna power the internet of things, the wearables, and everything else.

No, sorry, you're not even going to get the dumbest casio watch power from one of these things. And as I showed in my busted video here that that this power watch thing was just a gimmick. Uh, basically is because well let's go take a look at a data sheet for a typical coin cell battery you would have in a wearable device like this. Okay, standard Energizer Cr2032 battery.

This is actually a bit bigger than you might have in some little, uh, swim slimline casio watches. You might have a like a 16 16 cell if you don't know. By the way, this number actually tells you the size of the cell. The 20 here actually the first two digits here actually tell you the diameter.

In millimeters, it's 20 millimeters diameter and the three, uh, 32 here actually. uh, tell you put a decimal point in there and it's 3.2 millimeters thick. And so a Cr16c cell would be 16 millimeters across by 1.6 millimeters thick. There you go Anyway, one of these coin cells.

Look at this. 235 milliamp hours down to two volts. You know. that's basically.

you know. I mean the characteristic curve. it just lit drops off right at two volts. Yeah, there's nothing left.

Yeah, there's just like no energy left under that curve. There at two volts you've come across. So what's the capacity of a 2032 coin cell in What hours? Because we can't use milliwatt hours? Because that's really what hours. It's not watt hours, which is how you measure energy capacity and how we're comparing uh to the this little thermoelectric generator as well.

Well, Uh, 235 milliamp hours? Uh times? You know? I, uh, eyeball that? you know. Let's say it's uh, say 2.8 volts over its, uh, maximum life. I know? you've got to integrate the whole thing, but yeah, near enough. So 235 milliamp hours times a 2.8 volts? That's about a point or a 658 milliwatt hours.

All right. So let's do some post-it note Dave count calculations, shall we? Now, let's assume that this thing can generate its peak 10 microwatts. It can't because you'd be dead due to the temperature differential. But it didn't.

Let's it. Even though it's like really a hundred down at what's let's just assume, incredibly generous, it's 10 microwatts. Okay, so how long can I see our 2032 coin cell power a product at 10 microwatts? Well, okay, so we know that we've got 650 milliwatt hours here. We divide that by 10 micro watts.

What does that give us? Get the confuser out. That's 65 000 hours. Yeah, 65 000 divided by 24 hours in a day, divided by 365 days in a year. Um, we're talking seven over seven years.

Seven point, Four years. And that's actually a reasonable, uh, calculation. Because really, um, you can't get anything really useful under 10 microwatts? Really? You know you might. Yeah, the watch can get a bit under that.

But like, definitely not a smart watch or any sort of like, you know, modern Internet of things wearable device When you start talking about bluetooth communication and stuff like that, even bluetooth low energy sending packets. And even if you've got an E-ink display and you're doing all sorts of other things, there's not much you can do in 10 microwatts, so that you know is a realistic figure. So a typical, like wearable device internet of Things device could be an order of magnitude or two greater than this 10 microwatts. And of course, this thing can only generate 100 nanowatts per square centimeter at typical room temperature on a skin.

So come on, it's just not no hope. Most people just toss their little Internet of Things wankery gadget out after nine months anyway. Um, or the firmware update? Or brick the stupid thing. Or they'll just discontinue support or they'll start charging you bloody a monthly fee to use the damn thing.

Otherwise it just disables itself and pricks itself. bastards. You don't think that's real. That's actually happened.

I'm sure I've done a video on that. Yes, I have anyone remember Sonos who then deliberately brick their products and you hear about it all the time? I think I just tweeted one the other day. Internet of Things and here's how. uh, they've got like the lego like configuration and stuff and how you can like just join them to you can basically put them in series and get uh, larger voltages out.

but anyway, that one by one square centimeter it it's not going to do anything, it's it has no practical application for wearables at all, let alone uh, charging smartphones and other rubbish like that. And I know you want to know how many of these little one square centimeter ring type uh devices it would take to actually charge a smoke a smartphone in one hour? Well, uh, typical. uh, five, what hour? uh, battery? They might be bigger than that these, but you know, let's run with it. Five watt hour battery in your smartphone to recharge that assuming the phone's not used at all in one hour, and assuming that your little, uh, doodaddy, uh, thermoelectric generator ring thing can actually get out the 10 microwatts uh, absolute maximum that it claims and you're dead Anyway, because it's You know there's 95 degree temperature difference, But just ignore that, right? Um, 10 microwatts? We're talking.

You need 500 000 of these devices plastered all over your body with a 95 degree temperature differential to like you. You're dead. So you're like standing on Mars or something. That's where you get a temperature differential that big.

You would need 500 000 of these to plaster your body to actually do that. So that's a big yeah now. So hopefully you're starting to understand the practicality of these sort of things and and how this is just. it's just hyperbolic, impractical rubbish.

Uh, that they're doing here based on, like, probably really good and important research that'll have you know, nice these things aren't useless. They will have niche applications, but putting them on your skin is not one of them. I can't really think of any practical application where having a bunch of these things on your skin Uh, even with a decent temperature differential, you'd have to be pretty cold as we saw in those uh, performance curves to get even like the absolute maximum out of one square centimeter is 10 microwatts. And that's just, you know, crazy.

But in their practical tests here, it's in the order of nanowatts. It's like a hundred nanowatts. It just I like there. There's an application for these in remote sensing applications, and as there are with lots of these energy harvesting stuff, remote sensor output Real ultra low power remote sensing applications.

You might even have enough power to, you know, occasionally I burst a bit of a bluetooth low energy packet with some data back from a sensor. low power sensor, environmental sensor, or something like that. especially in places that are hard to reach. Difficult to replace batteries and stuff like that, but otherwise, no, you're simply better off just sticking your coins out.

I know they're supposed to be part of this thing, is supposed to be. oh, environmental and battery waste and all that sort of thing. but realistically like one of these batteries can last a couple of years in your watch. You can even like buy standard watches on the market that can get 10 year battery life shelf life of the battery.

They run off the smell of an oily electron, but that's only because batteries chemical energy is incredibly dense. You just can't get this from energy harvesting just to get the equivalent from an energy harvesting device. Uh, if like to get 235 milliamp hours equivalent to one coin cell. It's just not it's it's just ridiculous.

This is so impractical for all the stated applications, it's just never going to happen. So there you go. I hope you found that interesting and useful. If you did, please give it a big a thumbs up.

And as I said, do yourself a favor and go subscribe to Anton's channel because it's absolutely fantastic. I watch every single one of the videos. it's it's great stuff. Really, you won't be disappointed.

Catch you next time.