Keeping temperature of a chamber at 25C, how to?

[neazoi]

Hello,
I have a Dewar flask chamber and I would like to keep the temperature inside as stable as possible to 25C.

In comparison to a crystal oven heater, this means that it has to be heated (when temperature outside is <25C) and cooled (when temperature outside is >25C). a linear circuit is hetter than switched just on/off)

Is there any way I could achieve this stable temperature?

(PS. I was thinking of two peltier elements resting outside the dewar flask.
The cold surface of one of them would be connected with an aluminum bar to the inside of the dewar flask.
The hot surface of the other would be connected with an aluminum bar to the inside of the dewar flask.
And then some kind of controller, to control both.)

Another way I was thinking of, was to use a peltier cooler and cool the exterior of the dewar flasks, just like I was operating the chamber on a cool day. Then use an ordinary crystal oven inside, but set for only 25C. (attached image)


Any other ideas?

 

[betwixt]

The idea is good but you don't need two Peltiers. They work in reverse if you swap the polarity feeding them so you can add or remove heat by controlling the direction of current flow through them.
I've never tried it myself but I would imagine a bridge driver circuit fed from a temperature sensor and comparator would do the trick.

Brian.

 

[neazoi]

The idea is good but you don't need two Peltiers. They work in reverse if you swap the polarity feeding them so you can add or remove heat by controlling the direction of current flow through them.
I've never tried it myself but I would imagine a bridge driver circuit fed from a temperature sensor and comparator would do the trick.
Brian.

I worry about temperature oscillations using this way, as the metal has a temperature hysteresis (or whatever it is called)

How about the double chamber? It will take much more time to stabilize, but once it is set it should be much more stable. All you do is, "simulating" a cold day and then set the internal oven at a low temperature 25C.

 

[schmitt trigger]

I worry about temperature oscillations using this way, as the metal has a temperature hysteresis (or whatever it is called)

.

Thermal environments can be modeled using electrical circuits. The metal mass of a heatsink is modeled with a capacitor, and with the intrinsic thermal resistance will give you a thermal time constant which if not properly taken care of, will give you temperature oscillations as you correctly mention....but the problem will be the same with two Peltier assemblies, the temperature difference may not be as large, but the fluctuations will still be there.
Also, the inactive Peltier/heatsink assembly will be a low resistance path to the external ambient where you will have a very severe heat loss.

The way to solve the problem is with classic feedback theory. Using a PID controller, and adjusting its parameters carefully, you can have an extremely tight temperature control.

I won't go into the details of PID control and adjustment. There is simply too much information to be described in a single post. I would recommend to you to Google it.

 

[mtwieg]

I worry about temperature oscillations using this way, as the metal has a temperature hysteresis (or whatever it is called)

You mean hysteresis in the components themselves? That's news to me.... but the effective gain of the system will vary depending on the direction of the heat transfer.

The double chamber might make for a more easily controlled system, though it may require a lot of power to cool the outer chamber. Might be easier to heat the outer chamber and cool the inner one.

 

[neazoi]

Thank you all for the replies.
I see the double chamber a more practical solution as well, although this is the first time to try any of the two solutions.
Average power consumption could be minimized by the use of good insulation in the outer chamber as well as a temperature controlled power supply to the peltier.
Also, I would not include the secondary foam around the dewar flasks, these should be enough by themselves.
I think I will finally go for this solution as it is easier to be controlled with simple well known technology.

 

[FvM]

I worry about temperature oscillations using this way, as the metal has a temperature hysteresis (or whatever it is called)

There's no hyteresis involved in the setup.

Oscillation might occur - if the temperature controller is configured inappropriate.

I've never tried it myself but I would imagine a bridge driver circuit fed from a temperature sensor and comparator would do the trick.

This works well. You have to care however, that the bridge driver can set small curents around zero correctly, otherwise there will be a hysteresis and problems with setpoints near ambient temperature. A PWM controller with bipolar modulation (2-level PWM) will achieve this automatically, an unipolar controller (3-level PWM) must be properly designed. Of course the effort depends on the required temperature accuracy.

 

[GUMY]

What is the minimum required accuracy (maximum deviation of mean temperature from 25 degC) and stability (maximum permited oscilation of temperature)?

 

[neazoi]

What is the minimum required accuracy (maximum deviation of mean temperature from 25 degC) and stability (maximum permited oscilation of temperature)?

I have not set these requirements strictly. The frequency is to be corrected by GPS, but short term stability should be good.

 

[GUMY]

Are we talking about 1degC, 0.1degC or 0.01degC ?

 

[jiripolivka]

If you have a Dewar, it is a pretty well insulated volume to keep at 25 deg.C. The question is how thermally conductive are the connections to external devices or circuits. Their effect will cool or heat the volume in interest in the Dewar.

Many years ago I had to keep 25+/- 1 deg.C in a 60-liter volume of a microwave radiometer while the external temperature varied between -35 deg in winter to +45 deg.C in summer. After some development I combined a small home refrigerator cooling system with a 2 kW heater, both linked by two AC fans. The switching was best handled by a bimetal switch from an iron, rated to 2 kW at 220V AC. The complete system ran for almost five years, 25+/- 1 deg.C was never exceeded.

Later I learned that there were commercial water bath thermostabilizers for color photography with a better precision, capable to keep 25 +/- 0.2 deg.C I do not know if they are still available but they must have the water volume of >5 liters included. This would prevent the use in my application, I could not allow any humidity in my instrument.