View Full Version : Controlling output voltage


neily
01-27-08, 05:31 PM
Hi, Im new to this forum so HELLO!!

Im using a peltier juntion to produce voltage and run LEDs using the heat from a domestic radiator. But given the fact that LEDs require a specific voltage, I need to control the output voltage when the input temperature changes.

Any ideas how I go about this?

Cheers,

Neil

draqon
01-27-08, 05:41 PM
and you already have a device to transfer heat energy into electrical energy? oh peltier juntion....

hmmm...I would say you will need a microcontroller with a function to regulate the voltage by the voltage controller unit...

draqon
01-27-08, 05:49 PM
you can have voltage converter units...based on the range of the voltage....

hypewaders
01-27-08, 10:38 PM
What is the application- reading temperature? Why not just put a thermometer on the radiator?

Echo3Romeo
01-27-08, 10:57 PM
What is the barrier potential of the TEC? What does its temperature/voltage coefficient map look like? What will the delta between the hot and cold sides of the TEC be during operation? What voltage do your LEDs need? I'm inclined to think a zener diode will work for your application, but you need to give us some more specific circuit parameters.

neily
01-28-08, 06:02 AM
Hi guys, thanks for your input.

Im a product and furniture designer so forgive me - my knowledge of electronics is very limited!

Im hoping to build a lighting system that runs on the het of a domestic radiator, heres all he info about the Peltier Junction I have:

At 50°C it consumes 10.5 amps @ 17.4v giving a ferocious consumption of 182.7w (0.183Kw). At 25°C it gobbles up 10.5 amps @ 15.2v = 159.6 (0.16kw). It's pumping power or Qcmax is equally phenomenal @ 50°C being rated at a healthy 96w, equally as impressive @ 25°C it's Qcmax is 85w.
This TEC has 127 couples and a resistance of 1.24 ohm @ 50°C and 1.08 ohm @ 25°C. It's ΔT N2 @ 50°C is a very respectable 75°C and 66°C @ 25°C. And Finally it's voltage range is 0 - 15.2v @ 25°C and 0 - 17.4v @ 25°C, and it's leads are AWG 16 Teflon and approx. 6 inches (15cm) in length.


At 50°C it consumes 10.5 amps @ 17.4v giving a ferocious consumption of 182.7w (0.183Kw). At 25°C it gobbles up 10.5 amps @ 15.2v = 159.6 (0.16kw). It's pumping power or Qcmax is equally phenomenal @ 50°C being rated at a healthy 96w, equally as impressive @ 25°C it's Qcmax is 85w.
This TEC has 127 couples and a resistance of 1.24 ohm @ 50°C and 1.08 ohm @ 25°C. It's ΔT N2 @ 50°C is a very respectable 75°C and 66°C @ 25°C. And Finally it's voltage range is 0 - 15.2v @ 25°C and 0 - 17.4v @ 25°C, and it's leads are AWG 16 Teflon and approx. 6 inches (15cm) in length.
Further Details:
Ceramic Material: Alumina (AC).
Solid state reliability.
Built with high temperature solder with the ability to withstand higher assembly processing temperatures for short periods of time (<160°C).
Superior nickel diffusion barriers on elements.
High strength for rugged environment.
RTV sealing to improve reliability in condensing environment.
127x Junctions.
Leads - 15cm approx.

General Specifications:

ΔT
Vacuum(25°C)(Celsius) N/A

ΔT
Vacuum(50°C)(Celsius) N/A

ΔT
N2(50°C)(Celsius) 75

ΔT
N2(25°C)(Celsius) 66

Qcmax
50°C/25°C(watts) 96/85

Imax
50°C/25°C(amps) 10.5/10.5

Umax
50°C/25°C(volts) 17.4/15.2


Imax - input current resulting in greatest ΔT max, (Amps);

Umax - input voltage at ΔT max, (Volts);

Qcmax - maximum heat pumping capacity at Imax, ΔT = 0° C, (Watts);

ΔT - maximum temperature difference module can achieve at Imax, Qc = 0, (Degrees)



Hoping to run a set of 12v LEDs.

Thanks again!!!

Read-Only
01-28-08, 06:30 AM
All you need is a three-pin 12v DC regulator. Jus put in in the circuit between the source and the LED. There are many, a 7812 is just one of them.

Echo3Romeo
01-28-08, 05:10 PM
You don't need 12 volts to run some LEDs. The barrier potential of your typical GaAs LED is around 0.7 volts, and with a forward bias current of ~20mA, they'll start producing light. If your peltier can provide that, which it almost surely can, it should be enough.

Make sure to keep the cold side as cool as possible. The greater a temperature gradient you can create across the TEM, the more voltage it will produce.

Put the TEM on your radiator with a beefy heatsink on the other side, preferably fan cooled. Measure its open-circuit output voltage with a DMM. Let us know what it is. You might not need anything nearly as elaborate as any one of us has suggested thus far, and avoiding the efficiency hit of a regulator or zener circuit (not to mention the added cost and complexity) will make your project a lot easier to build.

neily
01-29-08, 07:36 AM
Hi guys, thank you very much for your help and apologies for my lack of knowledge!

I've measured the open circuit voltage and it's coming out at around 0.9v with a large heat sink (no fan) on its cold side.

Echo3Romeo
01-29-08, 03:18 PM
Try a single LED across its terminals with a 47 ohm series resistor.

Blue_UK
01-29-08, 03:48 PM
My first thoughts are power consumption. I don't know how much juice you can get from thermoelectric plate (which is what I know the peltier thing by), however, my idea is to connect a thermistor to the base of a transister and switch the LED with it. Obviously you will need to add resisters all over the place to get it just right.

Blue_UK
01-29-08, 03:51 PM
Oh - 0.9V? That's pretty poor :( You need around 1.4V to power a regular LED if my memory serves me correctly. There is a device known as a 'charge pump' (which I have used, so I know it works) which can generate larger voltages from lower ones, but they come in IC form and generally require more umph than that.

Neily, how much current can the thing supply at 0.9V?

Echo3Romeo
01-31-08, 09:49 PM
Eh, it depends what kind of LED. Some of the high brightness ones (2000+ mcd) need over 2 volts just to start emitting anything measurable. I would definitely check what the manufacturer specs the particular LED for though. I'd bet even money that it could be beyond the voltage he's getting across that TEM.

If the thread starter could give us an update it would be nice.

neily
04-08-08, 05:33 PM
Hi Guys,

Im only getting around 0.2v from my peltier so am using a battery as backup to do most of the work. And the peltier to force the v to a threshold point when the LED lights up.

So as not to drain the battery when not in use, Is there a component (switch?) that will control when the LED is on (past threshold point), and 'turn off' the battery??

Apologies again formy lack of engineering know-how!

Echo3Romeo
04-08-08, 10:16 PM
Use a NPN transistor like so:

http://www.kpsec.freeuk.com/images/tranload.gif

Leakage current will be less than one microamp, so your battery will last for eons. Just make sure to use a large enough resistor on the base lead between the transistor and the TEM, or else you'll fry the junction. Try a 10K pot maxed out, and adjust it down to fine tune the threshold.

neily
04-09-08, 07:14 AM
Where does the pot go?

And what is LOAD? :o

Echo3Romeo
04-09-08, 02:51 PM
The pot goes in series with the base lead, where that empty box is.

Remember to start the circuit with the pot at max resistance so you don't overdrive the base junction, and turn it down only enough to get the on/off threshold set to the temperature you want. Then you could measure the resistance and swap in a fixed resistor, if you want.

The load is your LED array, or whatever you want to control. Make sure it doesn't draw more current than your transistor is rated for. You could also make the load the coil of a relay and use it to control a much heavier load. Although a common, cheap NPN transistor (2N2222) can control up to one ampere at 50 volts, so you have some room to play.