Surface temperature sensor for solar cells

Dear community,

I have been looking around on the forum and on internet for any tips regarding the choice of an adequate sensor for measuring the surface temperature of solar cells. The cells will most probably be tested under artificial light which means that they will heat up quite a lot (a 1000W/m^2 halogen lamps heats up a lot). For this reasons it is important that only the surface temperature of the cell is measured, because ambient temperature will be much higher. Precision and accuracy can be around +- 0.5°C which seems to be reasonable for most sensors I have seen. I have read that analog sensors might be less accurate than digital sensors since they need a stable voltage reference. I would have a preference for widely available sensors because I have to find local suppliers. I would very much appreciate any direct tips or links to similar questions!

Warmest thanks.

You could use NTC temperature sensors. They are quite accurate and if you use the Polyimide wrapped ones you can use Kapton tape to mount them, maybe with some insulation to the ambient air like glass wool or even just more layers of Kapton.
I used these in the past:
http://www.banggood.com/5Pcs-25MM-MF5B-SMD-100K-1-Thin-Film-NTC-Thermistor-p-964979.html

They can be used easily with an Arduino and with a 1% divider resistor you can get 0.5°C accuracy.

Dear DedeHai,

Thank you for your input. Unfortunately I checked national suppliers and couldn't find the NTC temperature sensors. Would anyone discourage the use of the mainstream digital temperature sensors DS18B20 for surface temperature measurements? I could attach them under the solar cells in order to avoid the irradiance from the lamp. If this sounds like a good idea, what would be the best way to attach these sensors to the cells for highest accuracy? Thanks!

Here's a source of the 10K beads. They can be epoxied onto the back surface, or on the front with an aluminum tape umbrella. If on the front, the wire leads also need to be shielded from the sun as they will absorb heat which can travel up the wire to the bead. A backside mounting would be much preferred as long as the surface temperature is similar, or known, relative to the front.

DedeHai:
You could use NTC temperature sensors. They are quite accurate....

Are they?
A thermistor is just a variable resistor that needs extra hardware and code to convert that to a human readable temperature. Harware and code will determine linearity and stability, and YOU have to calibrate it.

Maybe easier to use a factory calibrated digital sensor, like the DS18B20.
Leo..

Thanks a lot for the various inputs. I will probably go with the DS18B20 then, which I can get right around the corner. As suggested by Artic_Eddie, I will fix them on the back of the cells, which are very thin and which results in similar temperatures for both upper and lower surfaces. Also, it avoids covering the cell from irradiation. However, wouldn't epoxy influence the heat conductivity? I am thinking of simply using insulating tape on top of the sensor, and maybe add a highly conductive layer such as aluminium between the surface of the cell and the surface of the sensor.

You could use a very small dab of the white heat sink paste under just the sensor then cover it with a piece of gorilla tape, like duct tape but much stickier.

Nongsai:
Thanks a lot for the various inputs. I will probably go with the DS18B20 then, which I can get right around the corner. As suggested by Artic_Eddie, I will fix them on the back of the cells, which are very thin and which results in similar temperatures for both upper and lower surfaces. Also, it avoids covering the cell from irradiation. However, wouldn't epoxy influence the heat conductivity? I am thinking of simply using insulating tape on top of the sensor, and maybe add a highly conductive layer such as aluminium between the surface of the cell and the surface of the sensor.

Before you attach them, bunch them up and get their temps. Cold and hot.
Then you can add a correction factor, if needed, in software.
Leo..

Nongsai:
and maybe add a highly conductive layer such as aluminium between the surface of the cell and the surface of the sensor.

Pointless. No matter how conductive it is, all you are doing is adding something resistive between the cell and the sensor. Essentially, the objective is to not have air, but what you propose is to move the problem rather than solve it. You would be better off using silicon goo or conductive paste. "Arctic Cooler", probably named after Arctic Eddie.

I don't think the ambient temperature will be higher at all, as the halogen lamps will be putting out radiant heat.

The DS18B20 is a great device but its upper working limit is 125C, which is a lot, but possibly below what you have in mind, and a PT500 or the like may be a better choice.

I am also using about 20 DS18B20s scattered around my house for temperature monitoring for things like water heater, space heating and ambient temperature. They work great.
I am also thinking to install one of these to monitor the temperature of my solar panels but still haven’t figured out a good way to attach to the panels bearing in mind that they are exposed to the elements of nature!

This might be simply because you made a bad choice of DS18B20. There are plenty of good choices.

Wawa:
Are they?
A thermistor is just a variable resistor that needs extra hardware and code to convert that to a human readable temperature. Harware and code will determine linearity and stability, and YOU have to calibrate it.

Maybe easier to use a factory calibrated digital sensor, like the DS18B20.
Leo..

This is partially true. A good NTC paired with a precise resistor and a stable power supply is quite accurate. Murata has a design tool for this. They are most accurate (like most temperature sensors) at room temperature where you get about 0.5°C accuracy. At 120° you are still at around 2°C which is pretty much what you get with the DS18B20. BUT it can measure up to 200°C or even higher when using the glass encapsulated ones.

whatever sensor you mount on the surface of the panel, the sun (or lamps in this case) will heat the sensor directly too.

So there-s a good chance the sensor will not measure the actual temperature of the surface, but higher or lower, depending on whether the sensor reflectivity is higher or lower than the solar panel itself.

Maybe an infrared thermometer, since it doesnt touch the surface could work, but I doubt since the panel shiny surface will also mirror an important part of the halogen lamp spectrum.

Also, make sure the lamps you use will have a similar emission spectrum as the sun, otherwise if e.g. the light bulbs send more or less infrared radiation than the sun (for the same visible light intensity), the results will not replicate panel temperatures for normal operation conditions,

You don't want to block the sun either, so it's wise to mount the sensors on the back. Solar panels are thin, so the back should be the same temp as the front.
Just stick/glue the sensor (and the first few inches of wire) to the back, and put some thermal isolating material on top of the sensor and wire to stop it from radiating and cooling down.
Leo..

Yes, I would mount the sensor on the back too. Insulating it however.. The temperature rise of the panel results from balance of received heat from sun and rate of loosing heat through both sides. The back insulated spot may get a little hotter than surrounding uninsulated panel.
Without insulation the sensor might show a temperature in between the panel and nearby air (lower).

Finding the exact amount of insulation seems to be tricky.