yes of course. Conventional heating element produces heating power constantly as long as electricity input is constant. The basic principle of this heating element is joule heating which is current flow through the resistance. So as long as the electrical resistance of heating element is not changed and the current is constant the heating power will be constant

Dissipated power being (as mentioned above ) the product U*I and since resistors change their resistance (more or less) as function of temperature you have to build a control loop to compensate for changes. This can be done in several ways the simplest being a PWM voltage monitoring.

It sounds like you would like to use a DC heater. If your application needs to be submersible you could check out products by Aqueon (https://www.aqueon.com/products/heating/mini-heater) or Cobalt (https://www.cobaltaquatics.com/products/mini-therm-heater). If you have a dry application, there are small heating pad options available by some IoT hardware firms, e.g. Adafruit (https://www.adafruit.com/product/4308?gclid=EAIaIQobChMIpOXk_I-V5wIVk5OzCh1DWgdqEAQYBCABEgIrdfD_BwE). Sunpentown offers a 5W portable and rechargeable hand warmer available from a number of retailers or online marketplaces. Does this help?

Yes!. a reliable alternative is to use a set of Peltier modules to Heat/Cool a desired application.

You can buy commercial Thermoelectric modules on the internet. There aren't expensive:

https://www.aliexpress.com/item/33010995173.html?src=google&src=google&albch=shopping&acnt=494-037-6276&isdl=y&slnk=&plac=&mtctp=&albbt=Google_7_shopping&aff_platform=google&aff_short_key=UneMJZVf&&albagn=888888&albcp=6516758312&albag=75664784622&trgt=743612850714&crea=es33010995173&netw=u&device=c&gclid=CjwKCAiA35rxBRAWEiwADqB37yA3-iml-74StePHOqwSXl2JqGZgy-Jj0gESoTYO-IF0hWsIubGeMRoCNacQAvD_BwE&gclsrc=aw.ds

Additionally, you would have to build a test bank to install the modules, with a couple of heat dissipators; One for each module (this depends on the range of temperature and the kind of sample or application you are looking to test), but at least you have to install a big enough heat dissipator above the cooling module, to remove the heat correctly.

and the corresponding power supplies; one power source (maybe 5-10 Watts depending on the modules) for each module

the corresponding releys for the temperature control, Temperature sensors (thermocouples, ambient temp sensor,) , switches, wires.. etc...

Certainly, you will be able to have a good control over the temperature for low temperatures applications (=

Hello Duy-Hung,

Your are asking a tough question. In order to control the electrical power dissipation of a resistive heating element, you would need to build a control loop to monitor both the voltage difference across the resistive element and the current through this same element and maintain their product at some setpoint corresponding to 5 W, 10 W, etc. Building a control loop that only monitors the temperature at some point near the resistive element may be insufficient to guarantee that power dissipation is fixed as the ambient temperature changes. In any event, even if you were to build a controller based solely on temperature, you would still have to verify that it (the control loop) was, indeed, maintaining the desired power dissipation and the only foolproof way of doing that would be to monitor the product of the voltage difference and current.

No matter what control scheme you choose, it would be prudent to include an over-temperature shutoff mechanism independent of the control loop in case something goes wrong. A simple thermostatic switch attached to the resistive heater and electrically in series with one of the electrical leads to the resistive heater would be sufficient. Do not depend on your main control loop to prevent your system from destroying itself.

One last point needs to be mentioned. Depending on how your electrical heater is connected (mechanically, thermally, etc.) to the outside world, there may be some leakage currents flowing to the outside world from the heating element. These are parasitic currents that may not be flowing completely through the heating element and as a result not contributing completely to the product of voltage difference and current required to calculate the desired power setpoint. Always check for leakage currents. Measure the current entering the resistive heating element through one lead, measure the current leaving the heating element through the other lead, and make sure their magnitudes are are equal. Use only battery powered multimeters that are completely elecritrcally isolated to make these measurements.

Regards,

Tom Cuff

@ Th. Cuff

1- NOBODY mentioned a simple temperature control since every body was aware of the problem.

2- Please indicate how you suggest to control independently both voltage drop AND current through the heating resistor.

I would highly appreciate the schematics for the control loop, but not one which could not be build. The 2 are related by the Ohm law U= R*I thus my interest in a way to control and adjust them independently.

3- Overload protection is OK

4- Depending on the way the resistor is insulated leakage can be avoided.

If the resistor is in a fluid it shall NOT be in direct contact for many reasons depending on the electrical conductivity of above mentioned fluid and on the local flow velocity. Low velocity can lead to hot spots and even gas/vapor bobbles generation ( which can implode = cavitation) although the overall temperature is low.

Hello Duy-Hung,

I mentioned that measuring temperature at a single point of the heating element might be insufficient to control the power dissipation of the heating element, but let me expand upon that idea to show that measuring two temperatures would actually be a very reasonable way to control the power dissipation.

Consider a cylindrical heating element and surround it with a close fitting, thin, metal tube. Note, we are not restricted to using a metal for the tube, many ceramics would work just as well. If one were to measure the temperature at the interface between the outer cylindrical surface of the heating element and the inner radius of the metal tube and make a second temperature measurement at the interface between the outer radius of the metal tube and the outside environment, then one would have a reasonable input to the control loop. The product of the mean cylindrical area of the metal tube, the thermal conductivity of the metal of the metal tube, and the expression for the temperature gradient in cylindrical coordinates yields the heat transfer per unit time, i.e., watts of heat dissipation via Fourier's law. In other words, by using two temperature measurements we have fashioned a heat flow sensor. This approach might seem novel, but it has been around since at least the 1980's, when I first saw it.

The question naturally arises as to whether the aforementioned approach has any insufficiencies? The answer is yes, but surprisingly no more than any other schemes. One can always argue that the thermal conductivity of the metal is a function of temperature, which is true although one could try to correct for this variation. There is also the matter of the thermal discontinuity at the interface between the outside of the cylindrical heating element and the inside to the metal tube, but use of a high quality heat sink compound could render this point moot. The real problem lies in the fact that the cylindrical heating element is of finite length. Thus, there are edge or end effects, which one can try to minimize by the use of thermal guard elements. Such an idea is identical to what is done to produce a standard air capacitor, i.e., a capacitor whose capacitance is directly determined by the plate dimensions and their distance apart. The use of a guard electrode minimizes the fringing fields that are present at the end of the capacitor plates due to the finite size of said plates.

It turns out that the finite size of the heating element creates insufficiencies in other control loop schemes such as those that monitor the voltage difference across and/or the current through the heating element. While the current at every point along the length of the heating element is the same, unless, of course, there is leakage, the same thing is not true for the voltage difference. At the ends of the heating element where they connect to the wires providing the power, the resistive elements tend to be at a lower temperature than at the center. When we calculate the total power dissipation in the heating element as VxI, we are implicitly assuming that the voltage drop per unit length of the heating element is uniform, which is not, strictly speaking, the case.

There are, of course, other effects that will trouble us such as oscillations due the finite speed of transmission of heat along the length of the heating element in response to the forcing signal from the loop controller.

Regards,

Tom Cuff

VERY good answer but not to my question concerning control.

I still have the hope to learn which way you will do it.