WeatherDuino Forum

http://www.meteocercal.info/forum/Thread...ion?page=2

Seems to be two threads (if that is the right word) running for this topic.

Hi,

Ah, I wondered what was happening. I received notification of a "new post" a few days ago but couldn't find it.

Also, note that the graph in #33 is continuously updating, so it often doesn't show what it was intended to do.

Maybe it's possible to retain or create a second thread, since the original thread did rather appear to be wandering Off-Topic into a more general discussion of pyranometer theory.

Cheers, Alan.

(12-06-2015, 00:44)AllyCat Wrote: [ -> ]Hi,

Ah, I wondered what was happening. I received notification of a "new post" a few days ago but couldn't find it.

A user posted a new post at 09-06-2015, but later he deleted it. I never seen that post too. I just know this because I checked the logs

(12-06-2015, 00:44)AllyCat Wrote: [ -> ]Also, note that the graph in #33 is continuously updating, so it often doesn't show what it was intended to do.

Maybe it doesn't make sense... but I will try to explain:
The graph shows the current data from my solar radiation sensor, updated every 15 minutes. I did it that way just so nobody thinks it's a manipulated graphic, to prove something.

(12-06-2015, 00:44)AllyCat Wrote: [ -> ]Maybe it's possible to retain or create a second thread, since the original thread did rather appear to be wandering Off-Topic into a more general discussion of pyranometer theory.

Cheers, Alan.

Agree. I will see if I can do something...
Sometimes is not easy keep the discussions on topic.

Hi Werk_AG,

Thanks. Is it the link (now broken) to #12 in my post (and I think referred to by you), both below, that has been deleted? Maybe Davis has complained?

(29-05-2015, 18:40)AllyCat Wrote: [ -> ]In the absence of any better information, I think we have to use one of the Davis Solar sensors as a "Reference" Design and from post #12 in this thread, it appears that the Davis sensor does not compensate for the diode/panel temperature, nor report it...

(03-06-2015, 17:33)Werk_AG Wrote: [ -> ]Somebody published (many thanks for that) the schematic of a reputable brand solar radiation sensor, the first stage amplifier uses what seems to me an integrator configuration. hvalentim, could you please just calculate the used gain for me? Do you find the formula, or wish I put it here?

The deletion is a pity, because I had hoped to answer your questions (in the link above) and perhaps help to get closer to reproducing the results achieved by typical Davis sensors.

There seems to be quite lot of information and photos related to that post which I can no longer find on the forum. However, I had saved some to my own files, in particular the photo of the (quite old) PCB and related data sheets. I'll wait to see how/if you can tidy up or recover these threads, but I recall that the Davis design was quite "conventional" and I could perhaps start a new thread to discuss the "current feedback" design that was shown.

Cheers, Alan.

Allycat Wrote:The deletion is a pity(...)

That material seems to be here.

(03-06-2015, 17:33)Werk_AG Wrote: [ -> ]hvalentim, could you please just calculate the used gain for me? Do you find the formula, or wish I put it here?

Did not notice that question. Circuit design at that level and op-amp options was not something I brought up and is a matter I really have no opinion or knowledge about. Bearing in mind that there is a post > here < where someone claims success using an AD822 with a photodiode, among other options, including purchasing the kit scheduled to be available soon, I was happy to try the same myself with a BPW34-B. Perhaps applying it to the board instead of the solar cell, an idea which I since understood that it is perhaps not bright. In the previous example a 100 Ohm R1 is said to be used, which seems excessive.

But if it is inspiration for alternative designs that is being looked for, why not getting it from well documented sources instead of trying to guess values?

This fellow (p. 43) should know what he is talking about:

Harrison, G. (2015) Meteorological Measurements and Instrumentation. Oxford, Wiley-Blackwell.

[attachment=281]

"3.4.2 Photocurrent amplifier
An example of current measurement using an opamp is in the use of a photodiode (...) to measure solar radiation, with the signal-conditioning electronics employed shown schematically in Figure.
In the circuit of Figure the photodiode (D1) is effectively reverse-biased (i.e. its cathode voltage at the opamp’s inverting input is more positive than that of its anode, at 0 V), and allows a (photo)current iPD to flow out of the opamp to 0 V. This is converted to a voltage by the opamp and feedback resistor R2, and the output voltage is halved by the potential divider R3-R4, to ensure it falls within a 0 to 5 V range suitable for a subsequent measurement circuit. A subtlety is the use of a voltage reference, D2, to provide a fixed voltage (Vref) above ground at the non-inverting input, in this case, with Vref = 1.2 V. This serves to raise both opamp inputs to 1.2 V which both reverse biases D1 for correct photodiode operation and ensures that the opamp output with zero photocurrent remains positive. Effectively, this raises the opamp input to 1.2 V, making, in relative terms, 0 V appear as a negative supply rail, but without the complexity of a dual rail power supply. The response of the circuit to the photocurrent iPD is given by (forumula in figure) which, for a linear response in photocurrent to solar radiation, indicates a linear response in output voltage."

Components (referenced in the schema) seem affordable.

Hi,

(12-06-2015, 17:43)hvalentim Wrote: [ -> ]That material seems to be here.

Thanks hvalantim. Yes, that's the material I had seen (and I believe Werk is referring to below), but I couldn't find it again.

(03-06-2015, 17:33)Werk_AG Wrote: [ -> ]I know that I'm using on this, the most simplest OP amplifier circuit possible, a single stage non inverting amplifier, and that a lot better can be done, but this way, for start, almost everybody could understand how it works ......

Somebody published (many thanks for that) the schematic of a reputable brand solar radiation sensor, the first stage amplifier uses what seems to me an integrator configuration. ..... could you please just calculate the used gain for me? Do you find the formula, or wish I put it here?

Yes, you make some very important points in that post (#35), particularly concerning the circuit being easy to understand. But 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.

Almost the complete circuit (a supply decoupling capacitor would be wise) is as follows (cut from the data sheet Appendix that I linked in #26 above):

[attachment=282]

The first important detail is that the diode is connected "upside down" (anode to ground), so the Op-Amp inverting input is being pulled (very slightly) below ground by the photo-current. The output of the Op-Amp thus rises above ground and pushes current back through the resistor until it equals the photocurrent (because the Op-Amp has negligible input current and very high voltage gain, so the negative input must also be at almost exactly zero volts).

You may be surprised that the answer to your question about the (voltage) gain is "perhaps a million" (the open loop gain of the Op-Amp), but remember that we are trying to measure the "short-circuit" current from the diode and you don't get much voltage across a short-circuit! The point is that this amplifier is converting a (small) current to a voltage (as does a resistor, of course) in a very predictable manner. The "conversion gain" is given simply by the resistance, for example a 2k resistor would give a 2 volt output for a 1 mA photocurrent (i.e. the "gain" is 2 V/mA).

Therefore, this amplifier is NOT (fundamentally) an "integrator", the capacitor is primarily to stabilise the Op-Amp (i.e. to prevent it oscillating at high frequency because it has a high voltage gain). However, a much larger capacitor value than "necessary" may be used to act as a low pass filter to stabilise (average) the output signal.

(12-06-2015, 17:43)hvalentim Wrote: [ -> ]This fellow (p. 43) should know what he is talking about:
Harrison, G. (2015) Meteorological Measurements and Instrumentation. Oxford, Wiley-Blackwell.

Yes his circuit is basically the same, I'll let you decide whose design and explanation you prefer. But IMHO (and perhaps Davis') it may be unnecessarily complicated! The key is the Op-Amp: If you use a "Single Rail" Op-Amp, they have a statement such as:

"Input voltage range extends below ground, Output swings rail-to-rail" (AD822),

so it is really not necessary to bias up the inputs if a suitable Op-Amp is chosen. Admittedly the output might not pull completely down to ground (at zero photocurrent), but alternatively, if the inputs are biassed higher than ground, then the bias voltage needs to be measured (or calculated) and subtracted from the Output signal - very probably giving a greater "zero" error than the slight loss of swing (with a grounded input).

Cheers, Alan.

My previous post seems to have disappeared;

Hi looking at these solar radiation sensors, I attach a very basic schematic (as far as I can see it.) and a few pictures of the layout of a Davis 7821 sensor. There are a few other things to consider.
a) The diffuser (any ideas on material.?)
b) The double housing which allows an air flow around the sensor inner housing. Presumably reducing the need for temp compensation.
Other info:
http://www.davisnet.com/support/weather/...sp?dtype=2
7821
Can this, or similar work on weatherduino?

---

just two more pictures, sketch of schematic for davis as far as I can see it and photo of inner housing, note another housing goes over this; which allows a flow of air through effectively a "double skin" housing, also note "diffuser"

Hi,

Yes, something seems to have gone wrong with the forum links, but you should find it here.

Cheers, Alan.

(12-06-2015, 13:34)AllyCat Wrote: [ -> ]Hi Werk_AG,

Thanks. Is it the link (now broken) to #12 in my post (and I think referred to by you), both below, that has been deleted? Maybe Davis has complained?

Hi Allan

Quote:Maybe it's possible to retain or create a second thread, since the original thread did rather appear to be wandering Off-Topic...

As sugested, I splitted some posts from the original thread and create a new one, but I forgot to publish any info about that. Sorry!
I have corrected the broken link in your post, hope you don't mind.


Usually, I read all your posts (not only here but in other forums too), with great attention, they are a source of learning to me, and my motivation to all this is learning... What I do for living is designing lighting for television shows, but since I were young I ever liked programing and electronics, I'm just trying to return to origins

Thanks for the explanations about how Davis used the OP in their Solar radiation sensor. In fact, since the member JT118, published the circuit I did some research trying to understand it.

Hi Werk_AG,

Thanks for your kind words. My background is a (retired) electronics engineer, primarily hardware, but I also spent many years writing microcontroller program code in Assembler. However, I was never "comfortable" moving to higher level languages such as C.

One of my present (long-term) "projects" is a microcontroller to "Add Solar logging to any FO station", which started before the WH-308x Solar Pods were even released! So it also might be relevant to Weatherduino. But my presence on this (and the Cumulus) forum might best be summarised as "To help make FO sensors better".

However, I'll just add a little to explain why the "Inverted input" Transimpedance configuration just discussed above might NOT be appropriate to this particular thread.

Personally, I distinguish between two types of "Photo Cell", which I would call a "Photo-Diode" and a "PV Panel". The first is basically a "detector" (with an active area up to perhaps 10 mm2), whilst the second is intended to generate power (with an area greater than say 10 mm2). But in principle, both can act like a photodiode and/or generate a current.

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.