WeatherDuino Forum

Full Version: Measuring Solar Radiation based on Photovoltaic Cells
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Just wanted to also concur with Werk's comments above, it's great to have you here Alan.
I've been tinkering with electronics for years. I built my first crystal set radio around the age of eight and had an interest ever since. That said the theory still goes over my head Smile
(13-06-2015, 11:11)AllyCat Wrote: [ -> ]I have assumed that this thread title refers to the latter, but maybe I'm wrong? However, if I'm correct, then an inverted input transimpedance configuration is not really appropriate. The reason is that ALL the photocurrent must flow from the output of the Op-Amp and we have the slightly "silly" situation that as the PV panel generates more current, the load on the power supply rail increases. As I mentioned in #2 of the Davis thread, I was rather concerned that too much current might flow from their photodiode (unless it has a strong optical filter) and overload the amplifier/feedback resistor.

That's probably all to say for now (in this thread); I look forward to seeing the results of your labours in due course.

Cheers, Alan.

Hi Alan

The work I'm doing, as you probably noticed is based on this article from Charles G. Wright.
As said previously the amplifier circuit is very simple, just a non-inverting amplifier. Below you can see the circuit. The shunt resistor (Rsh) for the PV cell, is not in the schematic, because I found that is better have it soldered directly on the PV cell output pads.


The values for Rsh, R1 and R2 of course are dependent of the characteristics of PV cell. I have tested a lot of PV cells and found that the better results come from 1V to 2V PV cells, with a short circuit current from 100mA to 150mA.
The type of PV cell is also important, monocrystalline cells are better but some polycrystalline also produce good results.
The only difficulty with this, is the calibration. After spending a lot of time, comparing the results of several combinations of PV cells, Rsh, R1 and R2, I found some combinations that produce results very similar to the results obtained by a Davis solar radiation sensor. As I don't have a real pyranometer, I used this, maybe questionable method. Currently I'm able to build several units, and get almost the same results from all of them.
Certainly this will not be a precision instrument, but from what I see, it gives better results than the ones used in Fine Offset weather stations (which, if I'm not wrong is most like a lux meter).

Finally, Rsh, R1 and R2 shoul result in the maximum OP output voltage of 1.1V, for a solar radiation of 1400 W/m2.

Feel free to comment. I'm always open to improve the system, provided that it is possible to maintain compatibility with existing hardware.
(13-06-2015, 00:16)AllyCat Wrote: [ -> ](...) I'll try to explain why Davis (for example) do it slightly differently (using the negative input of an Op-Amp), and to answer your questions above...

Thanks a lot for the explanations. Although grasping the details is hard from me I am glad to see that a degree of consensus seems to emerge as for the need for different designs and why whether you are using a small PV panel or a micro diode.

No doubt, one can get there satisfactorily both ways. The sole advantage of diodes seems to be that particular models - so called "blue enhanced", such as Osram's B model - should have, at least in the spec.s, a spectral response slightly closer to a pyranometer.

It would be a very interesting exercise having a BPW34B and a solar cell, each with it's suitable circuit design, side by side, comparing the results/readings.
Maybe it's better create a new threat to discuss the pros and cons of PV cells vs Photodiodes, and keep this one on topic, or soon we are talking off topic again.
Actually I was just nodding a sign of appreciation for someone who went to the trouble of redacting quite a few paragraphs for the benefit of others.

Are mainly phatic gestures of recognition "off-topic" in general? Or perhaps just mine in particular?
Please Valentim, in this topic we are just trying to discuss "Measuring Solar Radiation based on Photovoltaic Cells"

To you in particular. It's clear that you don't like the idea of use a PV cell for a the solar radiation sensor, and I don't have any problem with that, each one is free to have his opinion, but what you are trying to do is another thing: you are deliberately trying to force the subject of this topic to go in the direction of a system that will work the way you want (using photodiodes), hoping that someone will design it.
Can you please respect the subject of this thread? If you want to discuss any other methods, feel free to create a new thread.
Hi,

(14-06-2015, 02:41)Werk_AG Wrote: [ -> ]After spending a lot of time, comparing the results of several combinations of PV cells, Rsh, R1 and R2, I found some combinations that produce results very similar to the results obtained by a Davis solar radiation sensor (and probably a lot of time and effort).

The Davis gives you a considerable advantage, because probably the most difficult issue with this part of the project is calibration, which basically requires the availability of a reliable reference device (and a considerable amount of time and effort). Personally, I'm at a further disadvantage, that I live in London, where the sunshine cannot be considered "reliable" and I doubt if even a rare "clear sky" will enable me to calibrate my sensor accurately. Therefore, my involvement has to be more on a "theoretical" basis. Sometime I may explain my own project in more detail, but itwould take this thread too far Off-Topic. Wink

However, I do have two FO Solar Pods (one "on the roof" the other "on the bench") which I believe have a claimed accuracy of +/- 15%, but I doubt if mine even agree by that tolerance. Certainly they do quote the "sunshine" in Lux, but I'm not sure that they actually use a "Visible Spectrum" sensor, rather than a standard Silicon Diode (Visible + Infra Red), nor whether they attempt to achieve a "Cosine" directional response. Their more recent "Solar" stations are switchable between "Lux" and "watts/m2" (except that they got the conversion factor totally wrong ! ), but I very much douibt if they change the spectral response of the sensor.

(14-06-2015, 02:41)Werk_AG Wrote: [ -> ]Feel free to comment. I'm always open to improve the system, provided that it is possible to maintain compatibility with existing hardware.

In principle, it should be quite possible to use either a PV panel or a photodiode such as the BPW34xx with your hardware, particularly if the load resistor is connected directly to the panel/diode (which I agree it should). The adjustment pots give a wide range of gain control, but if we assume the wipers are near the "middle of track" then the nominal gain looks to be around 5.

A PV Panel nearly always contains a number of "diodes" in series, so the ~200mV required at the input of the Op-Amps is probably a sufficiently small proportion of its Open-Circuit voltage. A single diode like the BPW34xx generates only about 400 mV, so you should probably keep the "short circuit" voltage below about 100 mV. But the Op-Amp can easily give this higher gain. I believe the "naked" diodes generate about 7 mA (i.e. 1 mA / mm2) in full sunlight, so a shunt resistor of around 15 ohms could be about right.

As for the relative pros and cons of Panels v Diodes, that could (and maybe should) fill another thread (i.e. not this one). But its probably easier to "level" a Panel to within the one degree horizontality which is suggested as being necessary for good (Cosine response) results.

Cheers, Alan.
(16-06-2015, 23:53)AllyCat Wrote: [ -> ]As for the relative pros and cons of Panels v Diodes, that could (and maybe should) fill another thread (i.e. not this one). But its probably easier to "level" a Panel to within the one degree horizontality which is suggested as being necessary for good (Cosine response) results.

I will start by your last paragraph, because it is absolutely relevant to explain why I chose the PV cells way:
First: All the process seems easy to be understood by almost everyone with a basic electronics knowledge, like me.
Second, but very important: Everything in the WeatherDuino Pro 2 project was thought with the assembling in mind. It have to be easy enough to be built at home without special equipment. And in fact it's a lot more easier find a suitable box that allows leveling a PVcell, within one degree horizontal position, than a unit built with a photodiode which will require a great degree of precision in positioning the sensor (due to it's very small surface). Find a suitable teflon difuser for the photodiode may not be so easy too.
Maybe it's a fault of my way of being, but I always think in first place, in the practical or physical implementation. How am I going to ride it? How this could be built with simple materials?

(16-06-2015, 23:53)AllyCat Wrote: [ -> ]The Davis gives you a considerable advantage, because probably the most difficult issue with this part of the project is calibration, which basically requires the availability of a reliable reference device (and a considerable amount of time and effort). Personally, I'm at a further disadvantage, that I live in London, where the sunshine cannot be considered "reliable" and I doubt if even a rare "clear sky" will enable me to calibrate my sensor accurately

Two years ago, in summer with very clear sky, I spend several full days doing measures and comparing results... at sunrise, during the day, at evening. My garden was full of wires, voltmeters, Arduinos and PVcells. It really was a very laborious job, but done with pleasure.
What about to come to Portugal in the summer to calibrate your sensor? Smile

(16-06-2015, 23:53)AllyCat Wrote: [ -> ]However, I do have two FO Solar Pods (one "on the roof" the other "on the bench") which I believe have a claimed accuracy of +/- 15%, but I doubt if mine even agree by that tolerance. Certainly they do quote the "sunshine" in Lux, but I'm not sure that they actually use a "Visible Spectrum" sensor, rather than a standard Silicon Diode (Visible + Infra Red), nor whether they attempt to achieve a "Cosine" directional response. Their more recent "Solar" stations are switchable between "Lux" and "watts/m2" (except that they got the conversion factor totally wrong ! ), but I very much douibt if they change the spectral response of the sensor.

As you well know there isn't a correct formula to convert lux to w/m2, or better, it exists, but is only valid for the wavelength corresponding to the green light. I'm sure I already seen another formula which result in a more correct conversion factor, but I don't know which one FO is using. My first attempts to built a "solar radiation" sensor were based in a light intensity sensor (output in lux). When I start to learn more about the conversion process, I gave up instantly of this idea. SmileSmileSmile

(16-06-2015, 23:53)AllyCat Wrote: [ -> ]Therefore, my involvement has to be more on a "theoretical" basis. Sometime I may explain my own project in more detail, but itwould take this thread too far Off-Topic. Wink

Feel free to create a new thread when you wish. Will be a pleasure read it, and I'm sure that I'm not the only one who could say this.

(16-06-2015, 23:53)AllyCat Wrote: [ -> ]In principle, it should be quite possible to use either a PV panel or a photodiode such as the BPW34xx with your hardware, particularly if the load resistor is connected directly to the panel/diode (which I agree it should). The adjustment pots give a wide range of gain control, but if we assume the wipers are near the "middle of track" then the nominal gain looks to be around 5.

A PV Panel nearly always contains a number of "diodes" in series, so the ~200mV required at the input of the Op-Amps is probably a sufficiently small proportion of its Open-Circuit voltage. A single diode like the BPW34xx generates only about 400 mV, so you should probably keep the "short circuit" voltage below about 100 mV. But the Op-Amp can easily give this higher gain. I believe the "naked" diodes generate about 7 mA (i.e. 1 mA / mm2) in full sunlight, so a shunt resistor of around 15 ohms could be about right.

Mathematics never fails, isn't it? Your deductions based on the parts values on the schematic are absolutely right.
With the PV cells that I'm working now, I'm using a shunt resistor of just 1 ohm, which gives me a maximum voltage drop ("short circuit" voltage) around 150 mV at the shunt resistor.
The field tests reveal reasonable results adjusting the pot to get a OP gain around 7.

Cheers, Caneira
[attachment=318]Here is an interesting one from a very well renowned company;

http://www.ebay.co.uk/itm/Onset-S-LIB-M0...2ee329f2e2
It's an interesting device, but please try to keep the thread on topic. Here, we are trying to discuss methods to measure Solar Radiation based on Photovoltaic Cells. For other methods or devices, why not create a new thread?
Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17