Setting up temperature sensors

[clemencewiart]

Hi!

I have a setup right now that isn't really working out for me. I thought LabView would help. Here's a quick summary to let you know what I have and what I'm looking for, and hopefully someone will have the patience to help me out.

I have temperature sensors (max 6) placed in different chambers, scattered around a room, measuring the temperature in saline water. As of now, they are each connected to a homemade display, which gives me a live reading in Farhenheit, reheats the water (heats up a rod placed under the chamber) if the temperature drops below a certain point, and stops heating the rod if the temperature gets too high.

I'm looking to measure temperatures between 30°C and 45°C, ideally with an accuracy of 0.1°C. I'd like to get some kind of data acquisition going so I have a record of what's happening in the middle of the night (temperature spikes or whatnot), so a measurement every 30mn would be good. My end goal is to have all 6 sensors connected to one machine (a computer perhaps?) that can plot the temperatures so I can monitor them daily.

I know this must sound pretty vague, especially compared to other posts that are a lot more technical and precise. I must seem like a total newbie, but I really want to learn. I don't mind if it takes me months to set this up, I just want it to work correctly! This project is a sort of challenge I gave myself for my own personal development at work, so any pointers/tips/advice are very welcome. Thank you for reading through this 🙂

ps. I'll be taking a LabView training course sometime this year. I'm not expecting to learn all of what LabView has to offer from this forum, or to learn how to code on here. I'm posting this ahead of time so I can start learning on my own, get an idea of what I'll need for this project, and to know if first of all my project is feasible!

[johnsold]

First, welcome to the Forums.

It seems like you have a pretty clear goal for your project and that is good. And having a specific project is often a great motivator for learning a new system like LabVIEW.

Yes. It is feasible to build such a temperature measurement and recording system in LabVIEW.

The only thing I see in your description which raises any cautionary flags is the 0.1 degree C accuracy requirement.  That is right at the limit of what can typically be achieved with inexpensive sensors.  Thermocouples have 1.1 to 2.2 degree accuracy ratings for example.  Getting that kind of resolution is very simple.  For accuracy you need to be careful.  Digi-Key has thousands of temperature transducers and sensors but only one has +/-0.1 C accuracy.  If your water is not stirred vigorously and the chambers are not well insulated, there will probably be variations in temperature much larger than 0.1 degree within the chamber.  I am not trying to discourage you. I just want to help you avoid unrealistic expectations.

You do not indicate whether you want the LabVIEW system to also control the temperatures.  That is also feasible but you will want to have hardware safety features to prevent the heaters from being stuck on if the computer or program should fail. Simple safety thermostats on the heaters might be sufficient or a more complicated and sophisticated method might be needed. It depends on the consequences of a failure when no one is around to monitor to process.  I have seen labs burn down from unprotected water heating systems!

Temperature sensors often produce very small voltages (microvolts to millivolts). The heaters probably run on power line voltages (~hundred volts). So precautions to avoid interference between the desired small signals and the necessary high voltages nearby must be taken. This is much easier to do at the beginning planning stage than to try to fix it after everything is wired up and something does not work.  It may be as simple as routing the wires and cables for the measurements as far as possible from the power wires to the heaters.  In other cases signal conditioning modules may be needed at each sensor.

LabVIEW (or any other software) will not fix problems caused by a poor design. The design of both the hwardware and software portions of the project are important and the decisions about what things are done in each of those domains should be made early in the process. When you do not know much about the software (and maybe the hardware also) it is hard to make the right decisions. You might plan on going through a couple of iterations of preliminary design before committing to purchasing hardware or starting to write software. Then as you learn more about the capabilities you can revise without much "pain."

Please feel free to ask your questions and keep us posted on your progress.

Lynn

[clemencewiart]

Hi Lynn,

Thank you so much for taking the time to answer my original post. That was incredibly fast!

I am unsure as to whether I want LabView to also control the temperatures. As of now our heating system works well, that is after it stabilizes: when turned on, it has a tendency to shoot up to a higher temperature before switching the heating rod off, letting the temperature decrease, and then acting 'normal' for the rest of the time it stays on. I'm unsure why this is happening, maybe something to do with the gain or the response time, but that will be part of another discussion at another time.

For now, I will focus on using LabView to centralize the temperatures into one area to monitor them, and maybe the heating part will be a second project later on. I will need to keep that in mind while developing this first part though, as I am sure there are overlapping areas and elements between these two projects. I do not want to close the door for any further improvements.

I definitely need a resolution of 0.1°C, and I agree, an accuracy of 0.1°C is very small. That is the ideal accuracy we would like to have in our lab, however 0.5°C would still be acceptable. I will check out the Digi-Key temperature transducers and see what options I have. Furthermore, as you rightly pointed out, insulation will play a big part in keeping the temperatures from fluctuating. As of now, we are designing some lids to help with this ongoing issue. It also helps keep away contaminants so this is our top priority.

The water is not stirred, but we do have some fatigue testing going on that can move water around very slightly at a frequency of 5-10Hz.

Your last two paragraphs are golden, I am writing them down as we speak so I keep all of your advice in mind. I don't want to rush through this project just to get it done, I want to plan ahead and know what I am doing, and why. As you said, poor design can ruin even the best ideas. I also want to leave the door open for any further modifications, as I pointed out earlier, in case I have more time and have additional ideas to improve our setup.

I will definitely come back here if I run into more specific issues and need advice, or just to post my progress for anyone looking to do the same. Keep in mind it will advance very slowly, as I have my normal responsibilities at work, but I'd like to think I will have this project done before the end of this year.

Clémence

[johnsold]

With the slightly relaxed accuracy requirement you should be able to find sensors which are suitable. The resolution requirement is easily met with a variety of sensors.

Overshooting on initial startup is quite common, especially with simple on-off controllers.  The biggest factor is often the time lag from where the heat is applied to where the temperature is measured. If you put two (or more) sensors in different places in one of the chambers and watch the readings from a cold start you will likely see a few seconds lag due to the time it takes the heat to diffuse through the solution. The fact that it stabilizes later and does not oscillate suggests that the heater is about the right size and the time lag is not too severe.

More sophisticated controllers can be implemented in software to reduce that overshoot issue.

Lynn