# Making a stable thermal gradient w/ copper wire?

1. Jan 9, 2008

### pivazena

Hi all,

Biologist posting here. We have a thermal gradient that doesn't seem very stable. Right now, our setup is the following: hot water runs through one aluminum bar and cold water runs through another. the two bars are about 25cm apart. There is a thin aluminum plate resting on the bars. The heat and cold transfer through the aluminum plate such that there is appx a 1 degree (Celsius) / 1 cm gradient. However, we've encountered all kinds of problems with measuring this gradient, keeping the water at the right temps, etc (pretty much anything you can imagine, it's happened). I'm doing some exploratory work to see how much time and money developing a new gradient would cost.

Is there a way (an equation to use?) where we can run copper wire at 1 cm intervals down a plate (kind of like the defroster on the rear windshield of cars) and control the temperature of each individual wire to create the same gradient? (so, set the first wire to 30 degrees, the second to 29 degrees, etc?)

I guess what I'm looking for is a way to adjust the voltage (?) running through the copper to create more or less heat, but in such a way that we know what temperature the copper is. And, if we can adjust the voltage going into the copper, we can make the gradient more or less steep.

As for running the copper down the plate, would this be an open circuit? As in, if somebody touched it, they'd get a shock? Is there a way to sandwich the wire between two sheets of glass so that the heat transfers effectively but nobody burns a finger?

I'm sorry this is so awkwardly worded-- it's been a long time since I've done any electrical work, and it's been a long day :-) can anybody help?

2. Jan 9, 2008

### waht

I'm not sure what exactly is thermal gradient, but from your post it is sort like a thermostat? whose temperature is controlled only by the running water?

That would be a short, and you would need a high current power supply. Voltage would be low, so you wouldn't get shocked. I don't know how you would sandwich the wires, but super glue comes to mind.

Last edited: Jan 9, 2008
3. Jan 10, 2008

### stewartcs

The thermal (temperature) gradient is just the change in temperature over some distance, mathematically stated as:

$$\frac{\Delta T}{d}$$

CS

4. Jan 10, 2008

### stewartcs

Does the problem lie with not being able to measure the gradient accurately or with not being able to maintain the desired gradient?

CS

5. Jan 10, 2008

### Staff: Mentor

Can you tell us more about the physical arrangement of the final apparatus that you are trying to build? Will you need to account for thermal losses to the ambient air as well, or will the volume of the thermal gradient be well insulated? Is this for some experiment growing organisms in various temperature microenvironments?

If you can insulate things well enough, it seems like a straight hot-to-cold rail should work. Without very good insulation, though, you will need something else to give you a smooth temperature gradient.

6. Jan 11, 2008

### TheAnalogKid83

you could use transistors to generate your heat and use a feedback system to adjust voltage to control power dissipation by the transistors in order to stay on the specific temperature you require for each conductor. As long as the temp. is stable and doesn't need changes dramatically, you should be able to control the temperature of each conductor pretty accurately, or atleast you can know what the temperature is more accurately than just assuming a 1C/cm like you do for your aluminum plate.

7. Jan 11, 2008

### Danger

Welcome to PF, Pivazena.
How well is the plate connected to the bars? Your description implies that it's just lying across them. That could result in inconsistant contact, and thus fluctuating thermal transfer.
I would suggest mig welding them together, or at least using bolts and some heat-sink compound. Once things are solid, your original set-up should work.
As Berkeman pointed out, as well, your ambient temperature should be stabilized. If the room temperature drops, your hot side will lose more and your cold side will absorb less. Wouldn't that shift the gradient?

8. Jan 11, 2008

### dlgoff

This is my thought also. One thing that is important is allowing enough time for the system to stabalize.

My question is, how are you measuring the temperature across the gradient? With the insulation in place, knowing the hot end and cold end stabalized temps, the gradient should be very linear.