[uncle_bob]

I've run Google translate over Werk_AG's orginal post which translated it from Portuguese to English (quite well I must say too?) as it's quite interesting and something I'd like to do with my WeatherDuino Pro2 in the not too distant future.

Measuring Solar Radiation based on Photovoltaic Cells
One of the most frequent requests for information that come to me is on the meter Solar Radiation used here in MeteoCercal. For this reason I decided to choose precisely this theme, one of the first topics in this section.
For purists, it is clear from the outset that I will not write about building a real pyranometer, but about simple equipment Constructed using economic materials and easy to purchase. Do not have a laboratory quality, but it has enough to get the measurements that can be observed in Meteocercal, and that over time have proven very comparable to those obtained by reputable brands, quality stations located nearby.
By way of example, below a graph generated by Cumulus corresponding to the day 5/3/2014.



Prelimiar made this explanation, then move forward:

Step 1 - Study of Theory

For building success with this equipment, it is very important to understand the principles on which it is based. That said, the first step is to thoroughly investigate the matter in an article authored by Charles G. Wright , whose link I indicate below.

Measuring Solar Radiation

Again I mention that before proceeding, it is very important to realize that perfectly understood the theory presented in the previous article, because the steps depend on it.


Continued ... in the next topic

[uncle_bob]

Step 2 - Choose the Photovoltaic Cell

Reading and careful study of the subject mentioned in the previous section, we can easily conclude that for our project will not serve us any photovoltaic cell, but one whose short circuit current (which, for consistency with the original article, I shall designate by Isc,) is not very high, so that the resistance value Rsh resulting from the application of Ohm's law, does not result in a low value, which in practice becomes very difficult to obtain.



Something we should also take into account is that virtually all the power of the photovoltaic cell (called from now on CF) will be dissipated in Rsh. For this reason I recommend choosing a CF with a power not exceeding 150mW.

In the case of built for MeteoCercal sensor, the CF was chosen a model 1V - 100mW, which theoretically would result in a maximum of 100mA Isc (I = P / U).

Step 3 - Determination of Maximum Isc of our CF

To calculate the Rsh we will be using, it is very important to determine as precisely as possible what is the maximum Isc that our CF will provide the conditions in which we will use. This measurement is possibly the most critical aspect of building this equipment should ideally be done in the period of the year and the day we had the maximum theoretical value of Solar Radiation to our geographical position, but like almost everything in life, not always the ideial is attainable, so suffice in a good day of summer sunshine and clear skies. Later throughout the construction we have some margin to calibrate our equipment.
Thus, in a day of summer sun and clear sky, put our CF horizontally (never facing directly at the sun) and using a digital multimeter good quality (lowest possible internal resistance) we measure the Isc provided by our CF
In the case of CF I used, making measurements at the same time and over several days, the maximum Isc obtained was 75mA.
In the article mentioned in the first subject, possibly to facilitate explanation of the calculations, the author assumed as maximum value for the solar radiation value of 1000W / m 2. However, for latitudes such as Portugal, where the maximum theoretic value of solar radiation is between 1300 and 1400 W / m2, it is necessary that the readability of the equipment is not limited to such a low value, but Before can go to values of the order of 1500 to 1600 W / m 2. Just by way of example, in the Davis weather stations the maximum possible reading is 1800W / m2.

One day in the summer sun, clear sky at noon, carefully watching the WeatherWunderground values of solar radiation available by some meteorological stations located in the vicinity, all prowling 1300W / m2, and when I had my measuring Isc 75mA, assumed that this value of Isc would most likely match a value of Solar Radiation not far from the 1300W / m2. Later would have the opportunity to check if he was right or wrong (do not forget the study of matter, we are not to read current and voltage, and the CF, the next area of the short-cirtuito has a linear behavior, which is why the Rsh must have a small value).



So let the calculations!


Continued ... in the next topic.

[uncle_bob]

Step 4 - Determining the value of Rsh

As in the example presented in the study referred to in the first topic, I will also choose to calculate the value of Rsh, in order to obtain a voltage drop in Rsh with the ratio of 1 mV for each 10W / m2. So for a Solar Radiation 1300W / m2 intend to get a voltage drop of 130mV Rsh. Using Ohm's law as R = V / I, we Rsh = 0,130V / 0,075A ie Rsh = 1.7333333 Ohm.
It is important that the resistance to be used in Rsh has the maximum possible accuracy, and preferably not need to travel the world to find her. We will round the value calculated for 1.74 Ohm, as it exists in the E96 series values with precision of at least 1%. Has this value been coincidence? Continuing ...

The resistance Rsh shall be welded as possible along the positive and negative terminals of the CF, taking care to make sure that the weld is perfect. Poor welding may cause a resistance greater than the value of Rsh own contact. In each of the CF terminals, we should also weld a small piece of wire, which we will later use to read the variations of voltage drop in Rsh, which we now know will be linearly proportional to the intensity of solar radiation.

A small aside in support of rigor. We know that a photovoltaic cell, also not react at all frequencies that are part of the electromagnetic spectrum, and is practically blind those of longer wavelengths, we know also that measuring Isc we do is affected by an error has caused internal resistance of the measuring device (this mistake could be determined knowing the specifications of the device used in the measurement, but for simplicity we will ignore), but we know also that we are not dealing with the construction of a device for laboratory use, but rather a simple equipment for a fraction of the price of a real pyranometer enable readings with a margin of error that I consider acceptable taking into account the purpose for which they are intended.


Continued ... in the next topic.

[uncle_bob]

Step 5 - Reading and conversion Analog Digital, Facts and Considerations

At this point we already have a device on which we obtain a variation of voltage of 1 mV per 10W / m2 of solar radiation. We now need to make the reading of voltage variation and presentation of the results in the desired drive or be in W / m2.
Nevertheless there are other ways to read and display the data in case the device for MeteoCercal.info was intended to use an Arduino Nano using one of its analog digital converters (ADC) to do the reading. The ADC Arduinos thereof (such as most common Arduinos) have a 10 bit resolution is 1024 or steps. By default the reference voltage to the ADC is 5 volt, this means that when applying a specific voltage value at the ADC input, it is compared with the 5V reference and the result given on a scale of 1024 steps (meaning 0 through 1023 ).

Let us consider that the device ever built can produce in one Rsh maximum voltage drop of 160mV, corresponding to 1600 W / m2. Admittedly, for reasons explained earlier, we had considered our measurements as corresponding to a value of 1300W / m 2, but it is also true that the value of the theoretical maximum of Solar Radiation in fact is an average, and in our latitudes is often a pyranometer obtain peak values of around 1600W / m2, so we will consider this possibility, and set a maximum limit of our reading device 1600W / m2.

Let us now consider what is the maximum resolution of the reading we could expect: 5V / 1023 = 4,88mV, as each mv corresponds to 10W / m 2, as we would have maximum resolution 48,8W / m2. This resolution does not serve us!
Luckily Arduinos allow you to change the reference voltage for the ADC, and in the case of the Arduino Nano to have the possibility of using a reference voltage very stable, internally generated 1.1V.
Let us see then how would our maximum resolution, if we use instead of the 5V 1,1V: 1.1V / 1023 = 1.07, ie 10,7W / m2. Improved the resolution almost 4 times, however, this resolution does not serve us!

Possibly, that much has already concluded that, given the reduced amplitude of signals with which we are dealing, they could hardly be read directly by the ADC of an Arduino with an acceptable level of resolution. So how to solve this?
The solution passes through a linear amplification of the amplitude of the original signal, so that the maximum value of 160mV corresponds 1,1V, our reference value ADC. Proceeding this way, we obtain a significant increase in resolution (the contained is like exercise), resulting in 1,56W / m2, a value already quite acceptable for this project.
One of the simplest and most reliable ways to carry out the amplification of our original signal, an operational amplifier is used, however, their choice requires some important considerations as will be seen below.


Continued ... in the next topic.

[uncle_bob]

Step 5 - Signal Amplification

There are several ways of designing an amplifier based on operational amplifiers. In our case we chose one of the simplest forms, usually referred to as non-inverting amplifier which can be generically represented in the form illustrated in the figure below.



In this configuration the computation of the gain of the amplifier is obtained from the following formula:



Since the desired Vout is 1.1V and what our Vin is 0,160V our OP must have a gain to 6.875 (Vout / Vin), which we obtain with R1 = R2 = 8K Ohm and 47K Ohm.

The gain we want is quite modest for any OP, but probably will not serve in any one OP. Most of the OP power require a symmetric positive, or negative and GND. Since I'm not building a power just for the OP, but feed it 5V from the Arduino I have available, I will have to choose an OP that has this possibility, there are many and are designated by OP's Single-Supply.
Moreover matter me for future use, which is an OP able to provide a voltage very near the supply voltage output. These PBs are also of course, and are commonly referred to as OP's Rail-to-Rail. For this particular situation, my choice was to AD822AN - Single-Supply, Rail-to-Rail Low Power FET-Input Op Amp.

AD822AN APPLICATIONS
Battery-powered precision instrumentation
Photodiode preamps
Active filters 12-bit to 14-bit data acquisition systems Medical instrumentation
Low power references and regulators

We already have everything we need, the rest is now building.
At the end of this article a little example code for the Arduino that allows us to view our Solar Radiation readings will be presented.


Continued ... in the next topic.

[uncle_bob]

Step 6 - Construction of Solar Radiation Meter

Just to whet your appetite I'll start by saying that choosing another OP and changing the calculation of earnings is possible to get the data from this sensor can be read directly by a Davis VP2. No arduino without extra power, nothing ... just even everything that has been previously exposed and knowledge of voltage values used in the Davis VP2.

As for the physical construction of the sensor there are many possibilities to do so, the result is dependent on availability of materials, imagination and flair of each. That said, what follows I will describe is just the way I chose, certainly there will be other ...

The sensor will be the sun, rain, cold and heat, so that the primary consideration is to arrange something capable of withstanding these conditions and climatic safely accommodate the electronic components.

One of the ideas that occurred to me was to use one of those luminarias that are used normally entrenched in the soil for vertical illumination. Are structurally quite sturdy and waterproof, they are designed for outdoor use.
In the figure below you can see that the drive already installed and contruida.



In the first phase of testing, was installed in a small electronics "breadboard" and placed inside the illuminate after removing the original content. The photovoltaic cell was installed beneath the glazed part.



After several weeks of tests and measurements, the "breadboard" was replaced by a printed circuit board.



The card you see in the picture was designed thinking about the possibility of using the same "wrapper" for install also one UV sensor, which later then actually be done.

In the image below one can see the placement of the photovoltaic cell, as with UV sensor side.
A small note: The installation of UV sensor led to the replacement of the top cover glass that was originally by a Plexiglass. The ordinary glass absorbs between twenty to thirty percent of ultraviolet radiation, significantly altering the measurements, since the acrylic plastic is virtually transparent to ultraviolet radiation.



As I said at the beginning, the construction illustrated above is only a possibility, there are many others. The following are a few more ideas ... with simple materials!





It is thus concluded this article, I hope it's useful for someone.

If you liked this article, please collaborate in its rating by clicking on the "stars" at the top of the page. Thank you.

[werk_ag]

I've run Google translate over Werk_AG's orginal post which translated it from Portuguese to English (quite well I must say too?) as it's quite interesting and something I'd like to do with my WeatherDuino Pro2 in the not too distant future.

uncle_bob
Thank you for all this nice translation and formating work.

As the summer is coming to Australia, it is a good time to start building a solar radiation sensor of this type. Sunny days, cloudless, are required for final adjustments!

[water01]

Better wait until next year then

[werk_ag]

Ok, I change "are required" to "helps a lot"

[uncle_bob]

Thank you for all this nice translation and formating work.

You are most welcome Werk_AG!

Ok, now this thread raises the following questions

1) Can I purchase the PCB required?
2) What software settings are required to enabled this?
3) Could we just use the existing Solar sensor supplied with the WH3081 package, like pictured below (ignore the TX unit in the pic)?




Cheers UB