Circuit using thermoelectric power generators

[mmmmpwr]

Hello all,

Let me try to explain what I'm trying to accomplish.

We need to feed 100 amps at 2 volts across a load. The power will come from multiple thermoelectric power generators that are each rated at 8 amps and 1.6 V when 300 thermal watts flow through them. Let's assume that we can get the rated output of power.

The design goal is incorporate all the thermoelectrics into a circuit with some method of controlling the output current at desired, constant currents below or up to the maximum output. I think I'm correct in saying we will need to arrange these thermoelectrics in parallel. Beyond that, I really don't know what needs to be done.

I appreciate any input you guys may have. Let me know if more information is needed.

Thanks

 

[itsmoked]

Two volts is a bad voltage to work with as you need both humongous currents and more importantly semiconductor devices require some voltage drops for various control biasing and losses. 2V leaves you with almost nothing to manipulate. This comes down to efficiencies that are abysmal.

If you hook those in series the delivered voltage becomes higher. Required current becomes lower and the fixed voltage of the shaved off control voltage becomes a much smaller piece of the controlled power automatically improving the efficiency.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[Skogsgurra]

Elaborating Smoked's thinking somewhat:

If you have access to 1.6 V and lots of amps, which it seems you have. Then, arranging the elements in groups with two in series to get more voltage to work with, and those groups then connected in parallel will give you the current needed. More in series is, of course, possible. But probably not needed.

Next step is to control what you have. MOSFETs have a rather low channel resistance at low voltages. So, I think that is a good choice for the controlling elements. You could either choose to switch your thermogenerators in and out as needed (that is coarse and simple) or use a PWM scheme together with an inductor to have next to infinite control of current. PWM doesn't add much to complexity and is what I would choose.

And, since you seem to be wasting some heat anyhow, you can use a linear regulator consisting (again) of MOSFET transtors with heat sinks and a simple linear opamp controller.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[IRstuff]

Unclear whether you're doing a steady-state problem, or a pulse problem. If pulse, then one option is to emulate laser diode drivers.

They use a high voltage charged on capacitors, discharged with special, low-inductance leaded power transistors. The net effect is to get about 100 A on a device with about a 2 V drop, but by starting with a much higher voltage on the capacitors, say, around 100 V.

Your TE generators can be switched up to a higher charging voltage for the capacitors, and then discharged as described. You'll need to use flex cable with ground planes on either side of the current line to minimize inductance.

In the steady-state case, you'll likewise need to kick up the source voltage, to account for the IR losses.

The bad news is that you'll need lots more power, since much of your source power will be lost in the parasitics.

TTFN

FAQ731-376

 

[VEBill]

As already advised, wire them in series in order to obtain a more-practical voltage.

For a quick and dirty one-off solution, you could use an off-the-shelf DC-DC converter to provide a stable output voltage (perhaps +5v DC). Then use an off-the-shelf power regulator module as used with CPUs in PCs to provide the +2 volts at 100 amps (I think that some will operate at approximately this range).

The only thing left, if you require it, would be to regulate the current as opposed to the voltage. Perhaps a small uC could adjust the CPU voltage up and down to regulate the current.

This approach might only be feasible if you're making very low quantities.


Automakers are starting to look at using the same technology to replace or supplant alternators. Perhaps some of their circuits are posted on the

http://www somewhere.