How to use temp difference in 2 tanks to start pump

[klydolph]

Hi,

I have two water tanks A and B.
The temps in both tanks can vary independently.
Both tanks can be equipped with a temperature sensor.
What I want is an electrical pump to start when the temp of tank A is higher than the temp in tank B.
What setup/equipment do I need to achieve this.

I know I could achieve it when converting the two temperatures to two control voltages and using a comparator IC, but since this is going to be a professional installation I can't use any home-made stuff. So far I haven't found ready-made comparators for mounting on DIN rails.

Any help is appreciated.
Many thanks.

kly

 

[BrianPetersen]

Industrial application?

Why not use a PLC rack with an analog input card and a digital I/O card, and let the PLC handle the logic.

 

[Compositepro]

If you use two thermocouples wired back-to-back the output voltage will be zero when the tank temperatures are the same and output polarity will tell you which tank is warmer.

 

[itsmoked]

I agree with Compositepro. That's exactly what I'd do.

You could then use a commercial temp controller and the temperature it displays would instead be the difference between the tanks. You would be able to set that temperature to cause it to "cool" when displayed temperature got to great.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[jraef]

Simple elegant solution(s). I like it.

What is the pump going to do, mix the tanks? If so, you can ignore the sign and just run the pump until you reach equilibrium.

"If I had eight hours to chop down a tree, I'd spend six sharpening my axe." -- Abraham Lincoln
For the best use of Eng-Tips, please click here -> faq731-376

 

[klydolph]

Many thanks for your answers.
It seems like a good idea to put two temperature transmitters back to back, and do regulation based on the polarity of the signal. But the transmitters I've dealt with so far have had to be put into a serial circuit with a voltage source. With this configuration - won't one of the transmitters get the wrong voltage polarisation? Or is it possible two connect two thermocouples back to back and then use only one transmitter? Isn't it something about that thermocouples requires special leads to connect to the transmitter?

kly

 

[BrianPetersen]

What sort of controls do you have elsewhere in the system?

Somehow, to me, there seems to be a whole lot better "certainty", and a whole lot better diagnostic capability, and a whole lot better process control opportunities, to just use two separate temperature signals to a PLC and have the PLC make the decision of what to do (and display fault conditions if one or the other signal goes bad or is not plausible).

There are almost no uncertainties with this method. It's known how to do it. If one of the thermocouples goes bad, it'll show up and you can display a fault and take corrective actions, and you'll know which thermocouple went bad instead of not knowing.

It's normally not worth spending $1 to figure out how to do a 50-cent problem cheaper. Especially if the maintenance man wants to smack you around three months later when something goes wrong and it takes him an hour longer to track down the problem because you did it an oddball way.

 

[Compositepro]

Thermocouples are basic sensors that attach to transmitters. Back-to-back connection will not work with transmitters. If personnel do not understand the basics of temperature sensing and thermocouples they are just as likely to make design and application errors using transmitters and PLC's. Back-to-back thermocouples have been used for extremely precise differential temperature measurement since thermocouples were invented. If your tanks are a mile apart, you probable need transmitters.

 

[klydolph]

Thanks again.
I'm in the process of designing a system for water based heating in the floors of a family house.
One of the tanks will be an accumulator for the floor heating water. This tank will mainly be heated by a heat pump. The temp in this tank will be approx 35 degree C (the temp will vary somewhat - there is a controller in the heat pump that sets the temp based on outdoor air temp and user set indoor temp). The other tank will be an accumulator tank for water heated with a fireplace (burning wood) in the living room. Max temp in this tank is stipulated to be up to 80 degrees C. The plan was to make a pump circulate the Fireplace water through a coil in the floorheating water WHEN the fireplace water is warmer than the other.

kly

 

[Compositepro]

I don' see the point of two tanks. Just have the fireplace heat the tank. If the water temperature drops below 35C then the heat pump kick on.

 

[klydolph]

If I don't have some kind of regulation here the worst case scenario would be to get 80 degree C warm water circulating in my floors - and that I don't want.

 

[danw2]

I'd go with the suggestion to use a commercial temperature controller. There are dozens of (relatively) inexpensive 1/16 DIN temperature controllers with a pair of relay outputs; one a 'control' output, the other an 'alarm' output. Any of them run on 120Vac. Any take a thermocouple input directly.

If you have your heart set on a transmitter, then wire the back-to-back T/C's to a single transmitter to get a differential output, you don't need two transmitters. Let the logic make the same decision on the differential output as the controller does.

Either the transmitter or controller will indicate zero degrees when T/C's are at the same temp, plus some temp value when one is hotter, negative some temp value when that same tank is cooler than the other. Test with a match to see which T/C does which.

Use the control output to signal 'hotter' or 'colder' than.
Use an alarm relay to drop out the circulating pump if the temp differential gets too hot to avoid 80°C floors.