Go Down

Topic: Guidance to build a skin temp sensor using Uno (Read 7872 times) previous topic - next topic


Hi everyone,

New to the forums and new to Arduino, so forgive the dumb questions, trying to learn.

I want to create a skin temperature sensor to measure slight changes in skin temperature variation with an Arduino Uno. (+/- 0.5 degrees)

I was looking at the TI TMP006 (http://www.ti.com/product/tmp006) sensor but not quite sure what IC, breadboard, etc. I need or the best way to integrate the IC with a bandage or strap, for example, to display the temperature data in the Arduino console.

Would love to hear some thoughts and help on what I need to buy, learn and assemble to get this made.

Thanks in advance


Couple things. One, that is a non-contact sensor, is that what you had in mind? Two, it's an extremely tiny package that will probably require a small custom PC board just to breadboard it and experiment with it, plus a hot-air station to assemble same.


Hi Jack, thanks for your reply.

So contact vs. non-contact, for me it's important to get an accurate skin temperature reading. Not sure if the TMP006 is the right sensor.

For purposes of experimentation (i.e. getting an accurate reading using a temp sensor with Arduino) what do you recommend in terms of ICs? And approach to building this?

Ideally, it would be neat to have the temp sensor be as small as possible on, for example, a tiny protoboard with adhesive so I can stick it on the body. Just some ideas...


Actually it would probably work, in a non-contact mode. From your original post I got the impression you wanted a contact sensor. The TMP006 has some guidelines as to the emissivity of the measured object, I found that human skin is about 0.98 which should work well according to the User Guide on the page you linked, section 3.1, Target Object Emissivity Guidelines.

I'm no expert with temperature sensors, I've tried one or two, and the common DS18B20 works well for most of my applications, not as sure about yours. I've used the TO-92 package, but I'd think that smaller is better for your purposes, so I might try the µSOP package.  The stated accuracy of the DS18B20 is ±0.5°C, which may give it a slight edge over the TMP006. (Its resolution is 0.0625°C, which would fit your requirement for detecting 0.5° variations, be that °C or °F.)

Have you given thought as to how to evaluate the performance and accuracy of whatever sensor(s) you might try? Is there some standard to judge it against?


The standard for accuracy here would be a consumer infrared thermometer, which has been proven accurate in this application.

Any tutorials or links on how to best integrate/solder the IC to a proto board and then connect to the arduino? And then there's the issue of the code to program the arduino with to read and convert the signal from the sensor. Which I'm clueless about - but will learn!


May 24, 2013, 04:25 am Last Edit: May 24, 2013, 04:28 am by Jack Christensen Reason: 1
To try it with the DS18B20, I'd start with a couple in the TO-92 packages just because they're easy to breadboard. These would be fine to develop the code with and do preliminary testing. If that looked good, I'd get some in the smallest package and have a very small and thin PCB made for the sensor. Here is one of several good PCB services out there. Both packages should work the same, the smaller will respond to temperature changes faster and will probably work better under a bandage if that's the intent.

Get the OneWire library here to interface to the sensor. The code should be a piece of cake, just modify one of the examples that come with the library.


Thanks for this Jack, will get the parts now. Custom PCB after testing is the move, however, I have no clue how to design it  :smiley-eek: but there are probably resources and people to help with the custom PCB for the sensor.

My goal is to actually design a small PCB with this sensor and another sensor right next to it.



Can you please provide me the results of the sensor DS18B20.
I need a temperature sensor that measures the skin surface
temperature, not the body temperature. Is that a good choice?

I found on the internet "Skin Surface Probe", that are used in medical
area. Are they equivalent, or someone tried to use one with arduino?



Oct 17, 2013, 11:05 am Last Edit: Oct 17, 2013, 11:07 am by dc42 Reason: 1

Couple things. One, that is a non-contact sensor, is that what you had in mind? Two, it's an extremely tiny package that will probably require a small custom PC board just to breadboard it and experiment with it, plus a hot-air station to assemble same.

Or this ready-made board: https://www.sparkfun.com/products/11859
Formal verification of safety-critical software, software development, and electronic design and prototyping. See http://www.eschertech.com. Please do not ask for unpaid help via PM, use the forum.


Should this work if the sensor is in contact with the skin?



Oct 17, 2013, 06:03 pm Last Edit: Oct 17, 2013, 06:05 pm by elac Reason: 1
It's all about the skills


Yes, that's the skin surface probe, but my goal isn't to use that temperature sensor specificaly.
I want the most accurate result of the temperature of the skin surface in the smallest way possible.

Thanks for your reply!


Look at this:

A small 20 k? NTC Thermistor can be usable too.


Hmm, the 20k thermistor looks good, but the first option is a little expensive for what I have in mind.

Although at first searches on google I didn't found the 20k thermistor.

So I guess it's a better approach to use a analogue thermistor than a digital sensor.



Hi again,

Has anyone used this sensor: SHT25.

I wanna know if it can be used in full contact with the skin.


Go Up