Temperature Control System for a Metal Block

I've seen several industrial ovens by Espec. Equipment like this should meet your control needs but may be pricey. If you can tell a little more about your application (temp range, size of specimen) I may be able to help a little more.

Hope this helps!

Hey Thanks,

Well basically a design to heat/cool something (can't go too much in details)

Temperature is as high as 350 oC

The heater block is about 6cmx3cm height: 4 cm.

So basically is it possible to control the temperature to an accuracy of +/- 0.1 oC?

It sounds like this is something you don't want to give details about. In general I think you might be able to do what you want with a temperature controller. In particular you might want to look into getting a PID controller to hold the accuracy you want. How tight you can keep the temperature is a function of how big the block is, how it is being heated, how fast it looses heat, and how well the heat is transmitted through the block. It will of course also be affected by many other factors like how well you are able to set up the PID numbers. In general it is possible to control temperature to that accuracy, but it really depends on many factors to know for sure. Here is a link to a search form for temperature controllers, as I mentioned I would select PID as the technique, but the rest you will have to fill in as I don't know what you want. http://process-equipment.globalspec.com/specsearch/searchform/manufacturing_process_equipment/process_controllers/temperature_controllers

Thanks frank!

alot of help! i did not know it was possible to have a controller for such accuracy!

You caved in too easily.

None of the spec inputs for the controllers mentioned can measure within the specified accuracy. Certainly the controller can not control more accurately than will the inputs measure.

Consider the TC input listed. The drift of a type K will be 10 to 20 degrees based on temp cycles. The type E is more stable than K however the repeatability and accuracy of type E do not fall within the specified accuracy.

It is not possible for a PID to control more accurately than the accuracy of the input measurement.

Let's see. For a temperature of 350 deg. C, you'll need either a thermocouple or an RTD (Resistance-Temperature-Device/Detector). Unfortunately, both sensors do not have the accuracy that you need. The thermocouple is definitely a bad choice while a Class A RTD has an accuracy of +/-0.15 deg.

Let's assume that you can find a suitable sensor, can you control to within +/-0.1 deg. C? With difficulty, yes. There are a lot of factors. How well your block keeps its temperature depends on the ambient temperature around it as well as the air flow around it. If the ambient temperature fluctuates, your control system may swing out of tolerance. Your source of heat must also be steady.

It would seem that such a control system would be expensive and difficult to engineer.

Have a look at linear technologies (linear.com) application notes.

AN3 p.6 shows a linearized Platinum RTC signal conditioner with 0...4V acc. to 0...400°C with +-0.05°C.

AN5 p.1 shows a precision temp controller with 0.01°C step response.

Regards Uwe

Don't forget that, depending on the size of your block, it may not heat up evenly so spot measurements may not give an accurate indication of the temperature of the whole block.

wow, okk this is more difficult than I would imagine.

I am trying to convince "people", that it is too difficult to control the temperature to such accuracy. They're also putting on a fan to keep the ambient temperature constant. The eddy current will affect the heat block dramatically.

Thanks

If the sensor would exist. we could make such accurate controller which adjust +-0.1 centigrade adjustor.

[AN3 p.6 shows a linearized Platinum RTC signal conditioner with 0...4V acc. to 0...400°C with +-0.05°C]

even lower, as the space is smaller.

what stability do you demand? heating speed? the light heating can get preset value fast.

how can I contact you?

frankd20 is on the right tract. Consider this: I would suggest a control sequence that would stablize the environment surrouding the metal block, such as an evelop. After bringing the system to temperature, 350 dC, determine the temp of the air supplying the envelop. By maintaining that temp, you would stabilize the block and envelop.

Using PID technoligy, an RTD in the envelop would reset the temp of the air supply, controlled by another RTD, through its controller. The temp of supply air would always be under control, being offset just enough to bring the system back into equilibrium. The rate of change in temperature of the envelop-block system would be matched by the rate of change in the supply air system.

Omega Engineering, Inc. manufactures all classes of sensors,controllers, etc. They have a web site that is very user friendly.

Continuing this succesive approximation, with limited input data..

Good Platinum Resistance Sensors are stable enough <0.01deg in this temperature range. After calibration it is possible to have similar accuracy (uncertainty). Well made electronics can measure constant temperature with total stability below 0.05deg (including sensor error). If it is possible to isolate thermally block from ambient in controlled way, and changes of temperature need not to be fast, then it is not easy,but possible to do this, as said frankd20 earlier and others.