I'm a complete beginner in terms of arduino and electronic components/circuits.
In the progress of consolidating a list of components I need to get to start my little project.
What I want to accomplish is having a metal container, heating it from underneath the container with a heating pad, while tracking the temperature inside by using some sort of temperature sensor attached to outside the container, all of this would run of a rechargeable battery pack.
The main problem I found is figuring out what temperature is best to use in this scenario and have read loads without clear indication. The sensor needs to be as thin as possible, so when placed on side of container can be covered with an outer layer without creating a thick wall.
The popular probe being suggested is too big for this case, saw some flat film detectors which might be perfect but don't know if it would work or any drawbacks.
Specs for sensor:
Thin and compact
Cheap
Accuracy 0.5-1.0c
Operating temperature between 0 and 100c potentially (most likely less).
Any suggestions from personal experience?
Additional tips to speed up my project are welcome, cause at the moment it's like swimming in the sea and looking for islands of specific size, humidity, temperature and so on...
A film sensor should be fine. If you just want to see the temperature, you can get them complete with a display for just a few dollars. No Arduino needed.
Mine match a mercury-in-glass 0-50 over 10 to 35, so don't be put off by reply#2.
How to read that film using an Arduino and how to control temperature without large over and under temps?
Come on my friend. I just try to tell OP that such a precision is really difficult.
Measuring temperature differences of fractions of a degree is easy but controlling a heater to stay within such limits is difficult, very difficult.
I'm a complete beginner in terms of arduino and electronic components/circuits.
In the progress of consolidating a list of components I need to get to start my little project.
What I want to accomplish is having a metal container, heating it from underneath the container with a heating pad, while tracking the temperature inside by using some sort of temperature sensor attached to outside the container, all of this would run of a rechargeable battery pack.
The main problem I found is figuring out what temperature is best to use in this scenario and have read loads without clear indication. The sensor needs to be as thin as possible, so when placed on side of container can be covered with an outer layer without creating a thick wall.
The popular probe being suggested is too big for this case, saw some flat film detectors which might be perfect but don't know if it would work or any drawbacks.
Specs for sensor:
Thin and compact
Cheap
Accuracy 0.5-1.0c
Operating temperature between 0 and 100c potentially (most likely less).
Any suggestions from personal experience?
Additional tips to speed up my project are welcome, cause at the moment it's like swimming in the sea and looking for islands of specific size, humidity, temperature and so on...
Thanks for help.
Here are some questions you have not answered.
What metal is used for your container?
How tall is your container? The reason being, the temperature will vary from bottom to top. More so if not very temperature conductive.
Why would the temperature inside the container be different from outside? Possibly because of radiation and possibly from air convection.
Please explain the purpose of your project in order to get more direct answers.
Paul
super7800:
thermocouple attached with capton tape? silicon stick on thermocouple? maybe max6675 K-type thermocouple amplifyer?
i personally used fully sealed liquid thermocouples
Thanks for the reply. That looks way too big for what I am aiming. The sensor should be less than 0.5cm and just sit flat on the side of the container, that's why I mentioned flat film sensors, but as the other comment mentioned and kind of confirmed my worry is that integration with arduino might be problematic.
Railroader:
3. Accuracy 0.5-1.0c.... That's very difficult to reach for an amateur.
This is not the accuracy of the water and heating process, it's the accuracy the sensor should be able to pick up on. As I saw some sensors are super accurate into small decimal numbers and some are few degrees variation. For this purpose as long as it can pick up on 1c increase/decrease it should be fine.
Paul_KD7HB:
Here are some questions you have not answered.
What metal is used for your container?
How tall is your container? The reason being, the temperature will vary from bottom to top. More so if not very temperature conductive.
Why would the temperature inside the container be different from outside? Possibly because of radiation and possibly from air convection.
Please explain the purpose of your project in order to get more direct answers.
Paul
I believe stainless steel is most appropriate, I believe it's not great with temperature conductivity based on something I've read, but it's corrosion resistant. Any thoughts are welcome on this too. I guess thickness of the container is important.
To test the principle would be fairly small, something like 12cm high and 8cm diameter, at some point later might make a bigger version.
Not sure what you mean exactly. The temperature inside will be pure water temperature, outside it will be affected by the container. The idea is to have a nice clean build and not wires going into container from top.
I thought I explained the working principle behind the project, the project itself is part learning exercise part building something I will use. In cases where I need water or something else at a certain temperature, would fill it up, set the temperature and voila. It's like a small kettle I guess with variable temperature.
Thanks for the input.
That is the most common and closest to what I am looking for so far, but interested to know if there is a sensor which is flatter than this one which would achieve the same result in my project with arduino.
A thermistor can be far smaller (just a few mm largest dimension) and easily do better than 0.5°C.
The main challenge is to read a thermistor with high precision over that large a temperature range, as its resistance changes over 1-2 orders of magnitude, a change that's non-linear, and which requires careful calibration at multiple points, and some clever circuits for keeping the precision up over that range. A basic voltage divider (the normal way of reading a thermistor) does not do.
A TMP36 is another candidate in smallness, its precision largely limited by how well you can read its signal.
gedaz18:
That is the most common and closest to what I am looking for so far, but interested to know if there is a sensor which is flatter than (the DS18B20) which would achieve the same result in my project with arduino.
The chip probably doesn't take up the entire volume of the epoxy case. I bet you could sand the case down to make a more compact sensor.
Why do you need the sensor so thin? Aren't you going to insulate the sides of the container?
Hello,
Thermocouple is not the right way to go if you have a range 0-100°C: Thermocouples are more used to go range 0-1000°C or 0-500°C.
You can use a PT100 + transducer to crate an analog value but then you have to calibrate, can be done, but a lot of work and fuss.
DS18B20 sensors are definitely the better choice, they can be linked using digital IO and is far more accurate compared to thermocouple and about 0.5°C. They come in all kinds of hardware sizes, some have screw connections, use google and you will find.
I made a small movie + manual how to record the temperature it is on the Arduino project hub. DS18B20 demo + data recording to excel
The second thing you will need is a way to control the power off your furnace based on the temperature. Therefore you can use a resistor that is controlled by a SSR (solid state) relay. SSR relays are able to switch on and off very fast because they have no moving parts. There are SSR for DC and for AC. As you have a battery, DC will be the right choice.
Arduino's have the feature that some of their pins can do PWM pulse-width modulation hardware based. You write a value 0-255 to the register and the output will switch (example at 30Khz) with a duty cycle 0-255 (so 0-100%).
If you add a PI controller between the measurement and the PWM controlled IO that is controlling your resistor of the relay, you are done and you will be very accurate.
If you are not familiar with solid state relays, PWM and the necessary connections: I made 3 movies that document this technology to dose water into a container with peristaltic pumps. If you replace the pump with a resistor and replace the scale with your DS18B20 you are very close to a final working application. Overview movie Arduino tutorial (hardware): Dosing application with PWM driven peristaltic pump and weight scale. - YouTube Arduino tutorial (software): Dosing application with PWM driven peristaltic pump and weight scale. - YouTube
Good luck.
Pdf with detailed documentation: pdf with detailed documentation SSR, PWM, FSM (finite state machine) (bottom link)
Johi.
@wvmarle,
Correct remark: I intended to write that DS18B20 are much more accurate compared to thermocouple but it did not come out that way. Their resolution is better then 0.5 but accuracy and resolution are not the same.
Thanks for the remark, my post has been corrected.
Best Regards,
Johi.
Forget thermocouples for this. They do not have a linear output and are a pain for a domestic use. Unless you are dealing with very high temperatures which you are not.
Others are spot on. The easiest solution is to use a Dallas temperature sensor. like the DS18B20
There is also a waterproofed version.
This sensor is very cheap (about £2) and very accurate. It is interfaced using the OneWire library. Plus is will give you a direct reading in degrees C so it easy to code up.
It is also more resilient to noise as its a digital device not analogue.
