EEVblog #664 - Peltier TEG Energy Harvesting Experiments

[max666]

...
How much power would be available if a temperature difference of 70C or 80C could be maintained? Thanks.

Like Dave explained so marvellously in his video, stating a mere temperature difference doesn't tell you squat about the power you can expect.
Depending on your system you will have a certain thermal resistance from the medium providing the heat, through the Peltier device, into the medium dissipating the heat. That thermal resistance will determine how much heat (as in watts) will flow through the Peltier device at a given temperature difference. And if we assume an efficiency of 5%, then 5% of that heat flow will be converted into electrical power.

[steves]

...
How much power would be available if a temperature difference of 70C or 80C could be maintained? Thanks.

Like Dave explained so marvellously in his video, stating a mere temperature difference doesn't tell you squat about the power you can expect.
Depending on your system you will have a certain thermal resistance from the medium providing the heat, through the Peltier device, into the medium dissipating the heat. That thermal resistance will determine how much heat (as in watts) will flow through the Peltier device at a given temperature difference. And if we assume an efficiency of 5%, then 5% of that heat flow will be converted into electrical power.

Okay, I was picturing like an electrical resistor with an constant potential difference across it, but this is not a valid analogy?

[sacherjj]

We are researching making some of our own TEG modules to provide power in places where there is no electricity and the forest canopy and/or rainy seasons make solar impracticable.

Some of these modules can provide 30 W of power with a large differential (300C fire and 50C water in pot).  For example: http://www.customthermoelectric.com/powergen.html

There are a couple folks that make USB charging kettles and small camp stoves with seebeck modules for 10W 5V output in the kettle setup. 

The danger if this over a fire is that running the pot dry will burn up the module.  Some of these are good to 320 C, but not much higher without paying extreme prices.

[gardner]

I'm still struggling to come up with an application

These things have been around for yonks.  The most practical application I know of.
http://www.caframo.com/hearth/hearth_products_woodstove.php

These are recent and a bit gimmicky.  The recovered energy is used to run a fan to boost combustion.
http://www.biolitestove.com/products/campstove/

These are also fairly new and, IMO, gimmicky
http://www.leevalley.com/en/garden/page.aspx?p=68887&cat=2,40731

[steves]

...
How much power would be available if a temperature difference of 70C or 80C could be maintained? Thanks.

Like Dave explained so marvellously in his video, stating a mere temperature difference doesn't tell you squat about the power you can expect.
Depending on your system you will have a certain thermal resistance from the medium providing the heat, through the Peltier device, into the medium dissipating the heat. That thermal resistance will determine how much heat (as in watts) will flow through the Peltier device at a given temperature difference. And if we assume an efficiency of 5%, then 5% of that heat flow will be converted into electrical power.

The datasheet shows the temperature difference between the hot junction and the cold as being dependent on the electrical power being input when used in peltier mode, so does the reverse applies in seebek mode?

Perhaps I should rephrase my question:

Under ideal conditions, including a heat source of zero impedance, what is the maximum electrical power that this device can deliver?
and how about under optimal real-world conditions? What would the hot and cold temperatures be?

Thanks.

[bitwelder]

Been thinking about making a large TEG over a year now, the idea came when I realised that the temperature difference between room temperature and outside winter tempearature in Finalnd is around 40-50 C .

Just wondering... how are Peltier modules working with the cold side under 0C? Does that affect their efficiency?

[Tuoppi]

Been thinking about making a large TEG over a year now, the idea came when I realised that the temperature difference between room temperature and outside winter tempearature in Finalnd is around 40-50 C . So I was thinking to make a large seebeck panel (at least 50x50cm) out of many 40mmx40mm peltier modules found cheaply around Ebay, between 2 aluminum plates, stick one side of the panel to the chimney and other side facing free towards the outside cold air, maybe some insulation around the panel to keep each side at each own temperature.

You can't "create" energy. You will cool your room (move heat from inside to outside) to produce electricity   

Nice video

The heat used is wasted heat that is produced anyway, it only works as a heat pump when you power it not in "seebeck mode", if I'm not mistaken.

Not really. If you use the TEG module you will effectively cool down your room, so in order to keep your room at the same temperature as originally you have to crank your heaters slightly up. The electricity you generate with your Peltier will show up (even if you don't notice it due to the magnitude) in your kaukolämpö bill. Laws of Physics cannot be cheated.

[tszaboo]

Although these thermoelectric generators are interesting, and maybe solve a limited number of specific issues, for example powering a remote, high reliability ultra low energy consumer, like a sensor or something, i can't see any real use for them?

You aren't wrong. They are for niche applications only.
Even for remote sensors it can be better to simply install a large long life lithium primary cell. 10 years life is easily doable.

I really wonder, how much time it takes to generate the amount of energy the manufacturing of one module takes. If it will break even (under what condition) during the lifetime of these TEGs.

[EEVblog]

I really wonder, how much time it takes to generate the amount of energy the manufacturing of one module takes. If it will break even (under what condition) during the lifetime of these TEGs.

These are not commercial solar panels that must have payback to be viable. They are for niche applications that are otherwise impossible without them. Payback is of no relevance.

[coppice]

I really wonder, how much time it takes to generate the amount of energy the manufacturing of one module takes. If it will break even (under what condition) during the lifetime of these TEGs.

These are not commercial solar panels that must have payback to be viable. They are for niche applications that are otherwise impossible without them. Payback is of no relevance.

Right. You really need to compare the cost of these solutions with the kind of batteries which might be able to do the job. Harvesting is interesting where the battery solution looks nasty for cost or lifetime reasons.

[Excavatoree]

My water heater uses a Seebeck generator for the electronic controls.  No battery to worry with, no connection to the mains, (after all, older, non-electronic controlled gas water heaters had no electrical connection) it just works.    I suppose changing a battery wouldn't be a deal breaker, but you know no one would do it until that first cold shower - then one would have to wait. 

[fchk]

You shall call Wuerth by name "Wurt", not by value "worth".

We do not have a th.

fchk

[German_EE]

Think of it as German with an Australian accent

[tszaboo]

I really wonder, how much time it takes to generate the amount of energy the manufacturing of one module takes. If it will break even (under what condition) during the lifetime of these TEGs.

These are not commercial solar panels that must have payback to be viable. They are for niche applications that are otherwise impossible without them. Payback is of no relevance.

I'm thinking more about the effectiveness of the application. If you take a lithium primary battery, or a zn-air battery, the same size as the peltier-heatsink construction, would it provide more power over 20-30 years than the TEG generators?
AFAIK these are used on satellites with radioactive materials and maybe F1 cars are using this in their ERS this year, so it must be somewhat viable.
Are you thinking about a smartwatch with peltier?

[ckambiselis]

Been thinking about making a large TEG over a year now, the idea came when I realised that the temperature difference between room temperature and outside winter tempearature in Finalnd is around 40-50 C . So I was thinking to make a large seebeck panel (at least 50x50cm) out of many 40mmx40mm peltier modules found cheaply around Ebay, between 2 aluminum plates, stick one side of the panel to the chimney and other side facing free towards the outside cold air, maybe some insulation around the panel to keep each side at each own temperature.

You can't "create" energy. You will cool your room (move heat from inside to outside) to produce electricity   

Nice video

The heat used is wasted heat that is produced anyway, it only works as a heat pump when you power it not in "seebeck mode", if I'm not mistaken.

Not really. If you use the TEG module you will effectively cool down your room, so in order to keep your room at the same temperature as originally you have to crank your heaters slightly up. The electricity you generate with your Peltier will show up (even if you don't notice it due to the magnitude) in your kaukolämpö bill. Laws of Physics cannot be cheated.

Well the initial idea was to connect it to the fireplace chimney, so I'm not using the heat of the room but the heat of the smoke.

[ziggyfish]

They sound good in theory (even then its not that great). But if you take into account you only get 5% efficiency and that a solar cell can most definitely produce more power at lower costs. These things don't seem that viable as a solution to anything.

In remote sensing applications, you still have the problem of either cooling one side or heating the other which requires power. So although its cool, I just don't see a use for it.

[Hypernova]

They sound good in theory (even then its not that great). But if you take into account you only get 5% efficiency and that a solar cell can most definitely produce more power at lower costs. These things don't seem that viable as a solution to anything.

In remote sensing applications, you still have the problem of either cooling one side or heating the other which requires power. So although its cool, I just don't see a use for it.

Then use the heat nature provides and sink to the ground or sea, environmental corrosion aside unlike solar panels under the sun these things will last indefinitely.

[hamdi.tn]

Interesting video, i was thinking about getting this energy harvesting kit from wurth to play with , i think i don't have to now i saw how it work,i will just get this peltier device, thanks dave 
It's unthinkable to use it as a main power source for any sort of application even for remote sensing, as all energy harvesting solution they only useful to charge battery or a super cap while system is in a low consumption mode, sleep or whatever. used that way i can think about thousand of application where solar energy is not really the right solution. Really interesting stuff

[VK3DRB]

I use Peltiers to keep 3 watt CW laser SHG crystals at a specific temperature (typically around 45 deg +/- 0.2 deg C). The thermal mass is low but change in temperature over time is large at laser firing time. If the temperature varies by more than 1 degree, the crystal can be damaged and the laser can be destroyed. Hence recovery time to the temperature set point must be fast but we cannot overshoot. A difficult task.

By far the most the most difficult challenge with Peltiers (and power resistors at fast heaters) was to keep a silicon wafer and 528 probe tips at a constant temperature whilst slicing though the each wafer's plastic carrier support sheet to contact the silicon substrate to test the chips. The other side of the wafer was a large gold plated disc with a large thermal mass which would contact all 528 devices. Too hot and the silicon could be damaged by the impact of the 20 micron probe tip. Too cold, the plastic could not be reliably penetrated to make contact. The temperature was critical. Designing the heater/cooler and the airflow mechanisms was difficult enough, but the real killer was getting the PID algorithm right because the thermal lag was very long.

[VK3DRB]

Dave's video was quite good and highlights the amount of time that can be spent understanding a circuit or device.

One criticism of Linear Tech's board. Look at 09:26 in the video. What was the PCB layout guy smoking when he put in the designator for C01?

[dshay]

Real TEGs are expensive, $35 to 40 a pop at single piece prices, but they're worth it if you've got expendable income. Charge a cell phone with your camp stove kind of power outputs.

As has been mentioned they're great for when you have an application you can't get to to run power lines, like a old factory retrofit, buried sensors you cant get a field repair tech to replace, geothermal heat and mining applications, spacecraft that have a long term radioactive heat source. Russia used them in remote locations to power radio communication devices in the 50s and 60s.

Mosquito traps use them because they burn fuel to produce the CO2 and heat signature they're attracted to and the nail in the coffin is the electric powered fan they get sucked in with.

They're also commonly used as off grid back ups for wind/solar, just have an ignited fuel source kick in.

for $40 you can buy a lot of spare LiPos. At this point rich kid hikers would be the biggest market.

To check to see if it's a real TEG check the max temp conditions, 300C (400C in bursts), they also have high temp silicone insulation on the leads.

In idea conditions on a good mudule you can get anywhere from 10W, but you'd need conditions like I have, northwestern Minnesota winter. A outdoor stove is a perfect application for these, you could power thermostats/dampers or a fuel feed auger.

It's all about the delta T, eventually your cold side is going to heat up, in less than ideal conditions without a good cold source you're going to see 5W out of that same module if your trying to use ambient temp/heat sinks to cool. Best to find a water source/radiator.

Bismuth Telluride is a byproduct of lead mining, a finite resource unless people start recycling metals for their bismuth, a commonly used alloy.

There are two main manufacturers of real TEGs Marlow Industries and Tellurex. I know Tellurex has an online store based out of the US you can purchase directly

Scroll down to the bottom for their Thermal Power Generators.

Click here for tellurex.com

[fubar.gr]

I have some questions about TECs/TEGs.

1) All these PN junctions in series should result in a massive forward voltage drop requiring several tens of volts to make them work. But in reality a TEC behaves like a small value resistor.

Why is that? Is it due to the semiconducting material (Bismuth Telluride) used?

2) I don't get where the cooling effect comes from. Obviously, whatever process is taking place on the cool side is endothermic, but what is this process?

3) Does a single silicon diode also exhibit this cooling/effect?

[peltierseebeck]

I have some questions about TECs/TEGs.

1) All these PN junctions in series should result in a massive forward voltage drop requiring several tens of volts to make them work. But in reality a TEC behaves like a small value resistor.

Why is that? Is it due to the semiconducting material (Bismuth Telluride) used?

2) I don't get where the cooling effect comes from. Obviously, whatever process is taking place on the cool side is endothermic, but what is this process?

3) Does a single silicon diode also exhibit this cooling/effect?

1) Per 3) below, the semiconductor arrange in a diode is different than a Peltier module arrangement and thus the behavior is different. In the absence of a temperature difference across the Peltier module, it behaves like a resistor. Once a temperature difference is applied, a Seebeck voltage is generated and the true resistance will be difficult to obtain. One thing to note is that the electrical resistance of a Peltier module should be measured with an AC ohmmeter. Using a DC ohmmeter will create the Peltier effect, one side will get cold, one side hot and a Seebeck voltage will appear. This will change the resistance reading. Additionally, the resistance changes depending on the average temperature of the semiconductors. 

2) Usually the term endothermic is related to chemical reaction but the Peltier works a bit different. We know heat transfers from hot to cold. This is not the case for Peltier/Thermoelectric modules. Heat transfers from cold to hot. Just like a heat pump and these modules are heat pumps.

When a voltage is applied, charges carriers move inside the Peltier module p and n semiconductors. Charge carriers are electrons and "holes'. Charge carriers pick up heat at the connections (junctions) between the n and p semiconductors and copper strips connecting them. When the charge carriers reach the opposing semiconductor junction in the p-n couple, the heat is released. Peltier cooling (heat absorbed) happens at one side and Peltier heating (heat released) at the other. And of course by reversing the current flow direction, the charge carriers reverse direction and the direction of heat transfer changes.

Check out this link for more details: How do thermoelectrics transfer heat?

3) Diodes have a different arrangement of the p and n semiconductors and thus different physics. The p and n semiconductors in a diode are not connected by the metal strip, they are joined directly together. See this: How a junction diode works

[Kleinstein]

The main difference in the setup is that the materials are joint with ohmic contacts  like metal  n-semiconductor - metal - p-semiconductor - metall ..... So there is no forward voltage drop like in a diode.

The heat pumping action comes from the electrons not only carrying the electric charge, but also some thermal energy - how much depends on the material. When going from one material to another, the electrons have to adjust there thermal energy and the difference appears as heat or cooling effect.

The question of heat sources inside a normal silicon diode is interesting.  A similar and even rather strong peltier effect also happens there, at both contacts. When operated in forward direction, the contacts itself are cooled from this, and there is a significant heat source where the recombination happens. As silicon is a rather good thermal conductor and the regions are rather close not very much of this is visible from the outside.

[goolong]

Anyone have a clue about these TEC from https://phononic.com/ ?   They claim their efficiency is close to a regular refrigerator at 30 percent efficient.

Also, can someone let me know if this setup improves efficiency/ energy requirements for cooling  or am I screwing with the heat dissipation of the TEC too much

ie      pump heat ---> cold side (TEC) Hot side --(touching) -> Hot side (TEG)  coolED side  with large heat exchanger ( 2 to 3 times the surface area of TEC cold side)   

battery power input into the TEC,     recover some of the heat using a TEG   into a cap to recharge the battery.  I hope this is kinda clear.

 By the way im working on a Fontus clone and will be posting in the forums for help soon.