Self Powering Flashlight!

UrbanExplorer

Newly Enlightened
Joined
Feb 16, 2017
Messages
39
Location
California.
Hello my fellow flashaholics!
I am currently trying to make into reality a little idea I had regarding thermoelectric generators. (TEG)
As we all have experienced at one time or another, flashlights can get quite hot. After seeing what thermoelectric generators can do I have wanted to experiment by seeing if the heat produced from a flashlight can lessen the burden on the battery.

However, know knowledge with circuits is quite limited and I was wondering if you folks can suggest ways in which one can power a light source with two fluctuating power sources? I understand this project is not by any means the most practical but I am doing this strictly for novelty purposes.

I plan to use an XP-L HI LED as my light and one of the two power sources will most likely be an 18650. As for the TEG 90% of out there are 40mm by 40mm which I feel is a bit large for my application. A 20mm by 20mm would be an idea size. In addition to this they all ship from china and I do not think I can wait 2 months for shipping so if you know any good sellers please let me know!

Any electrical, science, circuitry, or shopping advice will be greatly appreciated!
 

Lynx_Arc

Flashaholic
Joined
Oct 1, 2004
Messages
11,212
Location
Tulsa,OK
From the little I've just read about TEGs they are woefully inefficient converting heat to power in the range of 5% which means a 95% loss of the "extra heat" that is used. Another issue of TEGs is the need for a stable temperature difference between two sides of it (cold vs hot) something that is not easily done with a flashlight especially when you have a hand on it. I'm thinking hot vs cool or warm vs cold would be working sets of temp ranges not hot vs warm or warm vs cool which is what you will probably have with flashlights and hands.
In other words I don't think it is viable using current TEGs.
 

PeterFH

Newly Enlightened
Joined
Sep 23, 2016
Messages
60
A few months back i had idea for a TEG light as well and did some research.
Normal peltier elements do not work well for this, special ones are expensive.
Paralleling them is possible but not that good of an idea, forgot why though, would have to look into it again.

After all that i ran some numbers and came up with, ludicrously low possible current and voltage output from standard 40mm peltier elements.

After all that, remember that peltiers generate current from a temperature differential, that means you can not just strap a peltier to your flashlight and have it work.
You need to put a decently sized cooler on it as well.

TEG need a big delta-T to work properly, you will not do that with the heat from a flashlight and without a decently sized heatsink for the cold side of the TEG.

However, if you put a candlelight and an old CPU cooler together, you will get a decent amount of energy and light out of it.
But it is more a nifty toy then something practical.
 

rayman

Flashlight Enthusiast
Joined
May 6, 2008
Messages
1,219
Location
Germany
The idea seems nice but I think even looking at regarding the involving thermodynamics it can be very effectiv. If you are looking to prolong the runtime in my opinion it is more useful to add another battery, saving space and...funds ;-).

But right now regarding it from a thermodynamics view really seems interesting for me right now.
 

PeterFH

Newly Enlightened
Joined
Sep 23, 2016
Messages
60
But right now regarding it from a thermodynamics view really seems interesting for me right now.

From that viewpoint, it is very uninteresting. :)
Looked up some numbers for a proper TEG module, the manufacturer says for it to work it should have 50°C on the hot side with a delta-t of 50°.
That means, you would have to cool the other side of the TEG to 0°C while the flashlight is 50°C warm.

Or, run the flashligh at 70°C and have the cold end be 20°C which means, a huge *** heatsink on the flashlight.
Not trying to poop on the idea, it is just not very practical.

Things like this work, here you got an example:
https://www.youtube.com/watch?v=QzTn9GtSpEU

But only until the huge thermal mass of the aluminium has reached equilibrium with the rest of the system.
As i said, seebeck effect generators work because there is a heat-differential, maintaining that means either.

Very good cooling with a small heat source on the other side.
A huge amount of heat with a small heatsink on the other side.

(Edit)
Here you got a pretty nice example for a fuel powered TEG lamp, quite interesting thing to build just not very practical. But maybe pretty interesting as an emergency light or novelty.
https://www.youtube.com/watch?v=LCB_YZN7lwo
 
Last edited:

rayman

Flashlight Enthusiast
Joined
May 6, 2008
Messages
1,219
Location
Germany
What I meant but interesting in the regard of the thermodynamics in this setup is that you want to use the heat loss of the emitter and convert that to electrical energy to boost the emitter a little bit more. In that regard it is always difficult to use the dissipated thermal energy of any thermondynamic system (in this case the emitter which transforms the electrical enery into light and heat) and transform it into a useful energy form (in this case electricity), if you didn't plan on using this heat in the first place.

You can calculate the effiency of a TEG module by using the Carnot efficiency and a figure of merit (consiting of the temperature, the seeback coefficient, the electric conductivity and the thermoconductivity). A larger differential temperature, a lower ambient temperature and a higher figure of merit results in a better efficiency of the TEG module. So in this case you indeed have to look at thermodynamics of the setup.

But in the end as you already mentioned the limiting factor in this case is the needed differential temperature and the low efficiency of the thermoelectric generator. First can be solved with a big enough heatsink or an active heatsink solution which lowers the overall efficiency again. In my opinion the resulting efficiency is too low to be really useful.
 

UrbanExplorer

Newly Enlightened
Joined
Feb 16, 2017
Messages
39
Location
California.
Again, I understand the impracticality of this project but I am doing this more for the sake of doing it. Sure waste of time to some Im sure but everyone seems to get caught in the unfeasability of this but ultimately I simply need some curcuitry and schematics advice haha. However, I did find all of your information quite useful.
After looking at some stats on TEG's it states that a mere 20° temp difference can creat 200-300 miliwatts which isn't too negligeble in my opinion. But unfeasability aside (I may use a different heat source even) any idea how one would be able wire such a contraption? Where you have a light drawing constantly changing amounts of power from two different power sources?? That's my big question as curcuitry is not my speciality.
 

Lynx_Arc

Flashaholic
Joined
Oct 1, 2004
Messages
11,212
Location
Tulsa,OK
There was an ambitious project on kickstarter for a TEG powered flashlight but it has yet to produce a shipping product and may have been a scam from the beginning.
The idea though was pretty cool.

I just checked that kickstarter project..... amazing people threw $150k at a light that is a 15ma driven 5mm Cree led.... you can buy a light that puts out that much for $1 and have $29 for batteries and chargers maybe even a movie and a meal at a fast food joint instead. The fact that it stores "excess" energy in a capacitor instead of a rechargeable battery to me speaks volumes of it not having any chance at a decent output beyond what a few button cells would be able to do. I'm guessing the light if made probably would be large and possibly push a few ma to a 5mm LED when it is winter outside and someone is holding it with a hot hand.
 
Last edited:

PeterFH

Newly Enlightened
Joined
Sep 23, 2016
Messages
60
Sure waste of time to some Im sure but everyone seems to get caught in the unfeasability of this but ultimately I simply need some curcuitry and schematics advice haha.

Tinkering and just trying out things is the best part about building something, you are right.
Ignoring the fact that is not feasible because the math does not add up, lets have a look at it.

If you have not clue how to do this, the best approach is to to start at the beginning.
Order yourself a few cheap Peltier elements of Ebay, if you do not want to wait for ages, look for local re-sellers.
Get a 10 pack of 40mm elements, they are common which makes them the cheapest.

Take your next best LEDs that you can find, a few resistors, a multi meter, pen and paper.

Just try ouf different configuration, heat sources, heat sinks.
Take some scrap aluminium, a tin can or whatever you can find, wire them up in series, parallel or combinations of that.
Write down your findings, calculate the numbers and get a feeling on how these things work.

Before thinking about circuits or drivers, run a basic study of feasibility and collect numbers.

Otherwise you do not know what sort of voltage you expect from your system.
You got to sort out your power source first because your entire circuit depends on what power source you have.

Once you got that done, worry about the other things.
Simplest solution, run an auxiliary LED of the TEG element with a decent boost converter, you get some pre-made that start from 0,9V and get to around 80% efficiency, that means, you will loose a bunch of power in your regulator.
Remember to take that into consideration when calculating your initial numbers from the Peltiers.

I just checked that kickstarter project..... amazing people threw $150k at a light that is a 15ma driven 5mm Cree led....

Yet that is not the most ludicrous idea for a flashlight i have seen, take a look at this:
https://www.youtube.com/watch?v=LUNDobqczW4

While that is an interesting idea, and it does work.
The tritium vials on their own would put out more light then the final product.
True beta voltaic on the other hand, i have to admit, would make for a very interesting light source.
Or even a RTG powered LED lantern, that would have a nice atomic age retro chick :)
 

UrbanExplorer

Newly Enlightened
Joined
Feb 16, 2017
Messages
39
Location
California.
Gotcha, well I do have some information. Here is what I know so far, the LED I will be using will have a max drive current of 3A and a voltage drop of about 2.95. I am also aware that one must wore the generators in a parallel manner in order for it to add energy into the system. However, the issue I have here is that the two sources of power will constantly fluctuating and I am not sure how that will work in terms of the circuit schematic. For example, when you first turn on the light, naturally it will run on 100% battery power and 0% generated power. However, over time let's say the generator covers 20% of the power need so the battery only needs to cover 80%. (I know everyone is gonna pounce on me saying how those percentages are impossible but it's just an example so bare with me) Due to this complexity I am unsure how the wiring would work. Is it going to require some resistors or voltage regulators? If so how would those work? As always I appreciate all the feedback!
 

PeterFH

Newly Enlightened
Joined
Sep 23, 2016
Messages
60
Due to this complexity I am unsure how the wiring would work. Is it going to require some resistors or voltage regulators? If so how would those work? As always I appreciate all the feedback!

You could solve this in a number of ways, simple passive solutions (diodes and resistors) will kill your efficiency.
And you need active regulation and maybe a switching regulator instead of a linear, again, you got to run your numbers!
Simplest idea, on the top of my head:
Single rechargeable lithium battery.
A LED driver, ideally switching but for tests a linear would do.
A Li* charge controller, highly depended on your power source, unless you know what voltage current you have, talking about which topology is senseless.
Micro controller to keep track of everything.

Light runs, processor monitors TEG output, when TEG output is high enough/present the charger is enabled.
The charger pumps energy back into the battery/system while the processor monitors the voltages.

Black -> Data
Red -> Power
Blue -> Thermal

Diagram:
http://imgur.com/uIyQRPs
 

UrbanExplorer

Newly Enlightened
Joined
Feb 16, 2017
Messages
39
Location
California.
Ohhhh okay I see what you're going at. Another friend of mine although less skilled in electronics spoke of the same idea. The diagram was extremely helpful I greatly appreciate it. Would it be simpler if instead of sending the excess power to the battery to "recharge it", we send it directly back to the LED and let the battery take care of what renaming power is needed to illuminate the LED? Not sure if this would overload anything. It seems that recharging a battery while it is draining is rather complex but so is slowly reducing the power drawn by the battery and filling that gap with power from the TEG I guess... Or maybe both of these are essentially the same thing and I'm not seeing it. Let me know haha. Also you do say run the numbers and I do have some numbers (the ones that are fixed and won't be fluctuating at least) and some formulas but this is my first time doing such a thing. How would you advise I go about such a thing? Again I appreciate your info, i wish there was a way to "upvote" you like in other forums.
 

PeterFH

Newly Enlightened
Joined
Sep 23, 2016
Messages
60
It seems that recharging a battery while it is draining is rather complex but so is slowly reducing the power drawn by the battery and filling that gap with power from the TEG I guess... Or maybe both of these are essentially the same thing and I'm not seeing it. Let me know haha.

It does not matter much were you "inject" your harvested energy.
At the moment, the simplest solution i could think of was the power source, because you do not need to worry about back feeding, potential differences, rectification and so forth.
Just dump the power into the junction of battery and regulator and do not to think about were it goes.
Either it feeds the driver or gets fed into the battery, the charge regulator does that all for your.

That might be a bad idea but is the simplest one i could think of at this moment.

For the most part, it is not a good idea to charge a Li* cell while discharging but in this case, it seems like a suitable solution.

Also you do say run the numbers and I do have some numbers (the ones that are fixed and won't be fluctuating at least)

But those are the ones that do not matter at this point.
The only number you need to know is, "How much energy can i harvest?".
Which is the foundation you got to put down before you do anything else.

How would you advise I go about such a thing? Again I appreciate your info, i wish there was a way to "upvote" you like in other forums.

As i said, practical test and experimentation with TEG/Peltier elements, LEDs and heat sinks.
Try to harvest energy and measure it.


  • How much at what temperature.
  • At what LED current.
  • How does it behave over Time.
  • How big is your delta-t.
  • How long is it maintained.
  • How much open circuit voltage.
  • How much open circuit current.
  • How much load current is possible.
  • How much drain current is possible.
  • You need to determine were you maximum power point is, optimum load current at what voltage.

It is really important that you get a good grasp of what the maximum power point is, you need to optimize your measurements accordingly, efficiency is key.

And you got to figure out a few things along the way and before that.



  • How to mount the peltier.
  • How to cool the cold side.
  • Were to mount the hot side.
  • Hot to mount the LED.
  • How the stack of components could look like.
  • Which order the LED, TEG, body and heatsink could be mounted.

This rabbit hole goes down, way down.

All that you got to do before you think about any sort of circuit.
At this point, you do not even want to think about walking, not even crawling.
You just want to turn over onto your stomach, then you can think about crawling :)

This is boring at the moment, because you want to solder things together and build something.
But i have to insists, you need to get the numbers first and make some charts.

Set down at your bench, wire stuff together and just have a look what happens and what power gets generated.

Read up on this for later on, the whole thing needs to be highly optimized to be remotely viable.
To make things simpler, you can think of a TEG as a solar cell, they are not the same but that may help for now to make things simpler to understand:

https://www.youtube.com/watch?v=5LKVRhd0mj0

https://en.wikipedia.org/wiki/Maximum_power_point_tracking
 
Last edited:

UrbanExplorer

Newly Enlightened
Joined
Feb 16, 2017
Messages
39
Location
California.
After much experimentation I am now considering adding solar cells to the list of power sources. They seem much more efficient than TEG's and can aid in power production especially if they are the monocrystaline type. I've been told that this would require some sort of series parallel circuitry. Do you guys think that it would possible to wire both the TEG's and solar cells all together to possibly recharge the battery or directly power the light???

As for the TEG unfortunately I purchased one that is not optimized for power generation and only meant for cooling/heating. Therefore, I was only able to get a reading of 1MA and 0.8V.

These were the two TEG units I was planning to buy. According to the folks at digikey one is better optimized for power production but I was unable to get an explanation as to how.

This is the one I plan to buy due to it's low cost but apparently it is only optimized for producing temperature differences and not electricity.
https://www.digikey.com/product-detail/en/marlow-industries-inc/XLT3-8-01LS/1681-1085-ND/6159133

I was told that this one was actually optimized for power production.
https://www.digikey.com/products/en?keywords=TG12-6-01L

I thought TEG units all perform the same thing???
let me know what you all think!
 

PeterFH

Newly Enlightened
Joined
Sep 23, 2016
Messages
60
After much experimentation I am now considering adding solar cells (...) especially if they are the monocrystaline type.

Mono crystalline cells are more efficient *IF* they get above a certain amount of light input.
At low light levels amorphous cells are more efficient.

Without doing some further reading i would suspect amorphous cells to be better suited.

But... where/how do you want to mount them and what light do you want to use?
If you put them into the beam, you loose light output into the room.

That would be like trying to create a perpetual motion machine

I thought TEG units all perform the same thing???
let me know what you all think!

You get TEG and TEC, one is a Thermo-Generator and one is a Thermo-Cooler.
Peltier elements are Thermo-Coolers that are optimized to work in a different temperature range.

Besides that there are differences in how they are constructed, for example how big and how many junctures are in the module.
The little cubes in between the plates:
http://www.reuk.co.uk//OtherImages/schematic-diagram-of-peltier-module.jpg
 

UrbanExplorer

Newly Enlightened
Joined
Feb 16, 2017
Messages
39
Location
California.
Wow! Did not know about that in regards to amorphous vs monocrystaline. I do like the flexibility as well. My light will have a home made reflector which I'm almost done constructing.

This reflector will use mylar blankets as the reflective material and I have built the parabolic shape out of a flexible transparent plastic similar to PET used in water bottles. The mylar is only about 80% reflective and a good 20% of the light simply goes thru the mylar even at obtuse angles.

Therefore I was hoping to line the back of the reflector with solar panels. My LED should emit like 400 lumens so I'm hoping whatever light that gets to the solar cells will be enough to tap into the "efficient mode" of the monocrystalines.

I initially was looking at amorphous solar panels but their "current @ pmpp" (Max current??) were all so low! Like the best ones were 10-50 MA tops. Most were much less. With the monocrystaline the lower ones were 100MA higher ones even 1A. Not to mention amorphous were waaay more expensive.

Do you think I can make this work with this monocrystaline below?

Here are the ones I was looking at: (Monocrystaline)
http://www.digikey.com/scripts/DkSearch/dksus.dll?Detail&itemSeq=227913462&uq=636304772208860047

(Amorphous) They're so tiny yet expensive with low power output :/
https://www.digikey.com/product-detail/en/panasonic-bsg/AM-5907CAR/869-1013-ND/2165198

https://www.digikey.com/product-detail/en/panasonic-bsg/AM-5412CAR/869-1007-ND/2165192



Alright, so basically neither one those Thermoelectric modules will give me an advantage in power production? If that's the case I might as well buy the cheaper one I guess...?
 
Last edited:

UrbanExplorer

Newly Enlightened
Joined
Feb 16, 2017
Messages
39
Location
California.
It does not matter much were you "inject" your harvested energy.
At the moment, the simplest solution i could think of was the power source, because you do not need to worry about back feeding, potential differences, rectification and so forth.
Just dump the power into the junction of battery and regulator and do not to think about were it goes.
Either it feeds the driver or gets fed into the battery, the charge regulator does that all for your.

That might be a bad idea but is the simplest one i could think of at this moment.

For the most part, it is not a good idea to charge a Li* cell while discharging but in this case, it seems like a suitable solution.



But those are the ones that do not matter at this point.
The only number you need to know is, "How much energy can i harvest?".
Which is the foundation you got to put down before you do anything else.



As i said, practical test and experimentation with TEG/Peltier elements, LEDs and heat sinks.
Try to harvest energy and measure it.


  • How much at what temperature.
  • At what LED current.
  • How does it behave over Time.
  • How big is your delta-t.
  • How long is it maintained.
  • How much open circuit voltage.
  • How much open circuit current.
  • How much load current is possible.
  • How much drain current is possible.
  • You need to determine were you maximum power point is, optimum load current at what voltage.

It is really important that you get a good grasp of what the maximum power point is, you need to optimize your measurements accordingly, efficiency is key.

And you got to figure out a few things along the way and before that.



  • How to mount the peltier.
  • How to cool the cold side.
  • Were to mount the hot side.
  • Hot to mount the LED.
  • How the stack of components could look like.
  • Which order the LED, TEG, body and heatsink could be mounted.

This rabbit hole goes down, way down.

All that you got to do before you think about any sort of circuit.
At this point, you do not even want to think about walking, not even crawling.
You just want to turn over onto your stomach, then you can think about crawling :)

I also read heard from somewhere that I can find an IC unit from digikey that does basically everything we're trying to solve here. Does that sound familiar at all? I'd never head of such a thing but it seems very plausible. Didn't have much luck with such a broad search quary tho.
 

Aepoc

Enlightened
Joined
Aug 24, 2006
Messages
395
Location
New York (not the city)
I've always been interested in utilizing the wasted thermal energy of a light... I did quite a bit of research on low temperature differential stirling engines/generators but was unable to find a design that was small enough and powerful enough for a flashlight. Stirling engines are an old technology but depending on the design they can be fairly efficient (a lot better than the 5% mentioned above)
 

Lynx_Arc

Flashaholic
Joined
Oct 1, 2004
Messages
11,212
Location
Tulsa,OK
I've always been interested in utilizing the wasted thermal energy of a light... I did quite a bit of research on low temperature differential stirling engines/generators but was unable to find a design that was small enough and powerful enough for a flashlight. Stirling engines are an old technology but depending on the design they can be fairly efficient (a lot better than the 5% mentioned above)
I think it is a lot more economical to use more efficient LEDs and circuitry in a light than to patch the losses with a very low efficiency power generation from the waste heat. If you spend $50 on a generator solution when a $10 LED swap results in 15% increase in efficiency then you are wasting money IMO.
 
Top