Homemade solar sensor, Looking for advice on making my own solar sensor Options

[Spamiam]

That is interesting that your ISS's range tops out at a lower value! That might be due to the A/D's reference being different.

Well, remember, I am driving the ISS from an op-amp. It will not get fully to the rails. So, I think that the 1765 is actually a pretty good match to your experimentally derived data.

Well mabey the .1u might be close... remember that a 100 ohm is going to be charging it and the fet is not a perfect switch. The ISS has it's own 3 volt regulator and the regulator's max current should not be a problem when the fet turns on to charge the cap.

Well, a 0.1 uF cap alone will charge fully in about 70uS. Even allowing for massive other issues, the power to the op-amp should be up to 100% in about 1mS. Probably, in another mS the output is fully "on" too. This leaves quite a bit of time for the ISS to do some sampling.

This is becoming an intereting topic... getting to know the differences between units. Have you been able to find out the lowest input for your ISS? and what the console's lowest value displayed (before 0) is?

No I didn't because No matter what that lowest value is, I will not be able to do much about it. Once I get the sensor's resistor values set for the overall gain, then I can start to think about offsets.
-Tony

[pomonabill220]

Well, remember, I am driving the ISS from an op-amp. It will not get fully to the rails. So, I think that the 1765 is actually a pretty good match to your experimentally derived data.



Well, a 0.1 uF cap alone will charge fully in about 70uS. Even allowing for massive other issues, the power to the op-amp should be up to 100% in about 1mS. Probably, in another mS the output is fully "on" too. This leaves quite a bit of time for the ISS to do some sampling.
Yes, realativley speaking, there should be more than enough time!



No I didn't because No matter what that lowest value is, I will not be able to do much about it. Once I get the sensor's resistor values set for the overall gain, then I can start to think about offsets.
True! Getting any kind of active drive to sink to 0.0v might be kinda tough (but not impossible).... mabey we could hang a lead weight on the signal wire to "sink" it to zero? Bill M
-Tony

[Spamiam]

True! Getting any kind of active drive to sink to 0.0v might be kinda tough (but not impossible).... mabey we could hang a lead weight on the signal wire to "sink" it to zero?

Well, I actually want to have a little positive offset so that at the first glimmering of light, I start to read more than zero.

I have my test circuit sitting outside. It reads zero while it is still quite light enough to see. I have not had a REALLY BRIGHT day to test the gain. Hopefully this weekend will be quite bright.

This AM before I left for work (7:45AM), it was reading about 5*90/1024 volts. This was with a moderate overcast and looking like it was considering raining.


I wonder how visually bright it can be before the Davis solar sensor starts registering more than zero.

-Tony

[Spamiam]

Well it was quite bright later in the day. My naked "test" photodiode and amplifier circuit got just over 3 volts output (using a 1.00K input resistor, 10.0K feedback resistor and a 1K resistor between non-inverting and ground). I was powering the op-amp from 5v so I can see precisely how high it will go, aiming to go no higher than 3V output.

I then modified my actual sensor (in a housing) to the same specs except it has a 2.2K resistor between the output and the ISS, and it has ground directly to the non-inverting input. I expect the housing to drop the peak reading to something below 3.0V.

It is fairly dark now (9:15PM) and it reads 19W/m2. I do not mind having a slight positive offset, btw.

-Tony

[pomonabill220]

Well it was quite bright later in the day. My naked "test" photodiode and amplifier circuit got just over 3 volts output (using a 1.00K input resistor, 10.0K feedback resistor and a 1K resistor between non-inverting and ground). I was powering the op-amp from 5v so I can see precisely how high it will go, aiming to go no higher than 3V output.

I then modified my actual sensor (in a housing) to the same specs except it has a 2.2K resistor between the output and the ISS, and it has ground directly to the non-inverting input. I expect the housing to drop the peak reading to something below 3.0V.

It is fairly dark now (9:15PM) and it reads 19W/m2. I do not mind having a slight positive offset, btw.

-Tony
Are you using your light pipe poking out of your enclosure to the pdiode? Are you using any kind of diffuser / filter or is the pipe frosted or clear and polished flat and pointing striaght up?
Thanks... Bill

[Spamiam]

Are you using your light pipe poking out of your enclosure to the pdiode? Are you using any kind of diffuser / filter or is the pipe frosted or clear and polished flat and pointing striaght up?

I had on hand some translucent white plastic dowels, some sort of plastic stand-off. They are about 1/4" diameter and the section I used is solid. I cut it down with a razor saw so that there would be as much length protruding as its diameter. I.e. the length and width are equal. The sensor is centered under the light pipe. It is pointing straight up. The cut end is reasonably smooth, but I did not polish it or anything.

I modified the resistors yesterday and something is not right. I have to bench test the sensor to see what is wrong. I had been using a 1K resistor between the sensor and the inverting input. The feedback resistor was 10K, the non-inverting input was connected directly to ground and I had a 2.2K resistor between the output and the ISS.

I was finding that I was only getting to ~600w/m2 in full bright, and there was a very slight positive offset in the dark.

I deleted the 2.2K resistor, add a 1K resistor between the non-inverting input and ground and increased the feedback resistor to 12.1K.

Last night I saw a significant positive offset (244w/m2)! I am not sure what the peak reading will be today, but it sort of does not matter given the large offset!

-Tony

ADDENDUM:

I pulled out the circuitboard and cound a stupid error. The non-inverting input was getting ground directly, then the 1k resistor was passing the ground through it to the rest of the circuit. I.E. there was high impedance in the ground. I only had to re-route the ground to everything else directly, then thru the 1K resistor to the non-inverting input.

With the 12.1K resistor, I am getting just over 1000w/m2 at high noon.

This seems like a reasonable result. I have no idea of the temperature effects on this reading, unfortunately. I painted the top of the enclosure white yesterday. It seems to have lowered the heating of the enclosure somewhat.

-Tony

This post has been edited by Spamiam: Jun 8 2008, 10:53 AM

[pomonabill220]

I had on hand some translucent white plastic dowels, some sort of plastic stand-off. They are about 1/4" diameter and the section I used is solid. I cut it down with a razor saw so that there would be as much length protruding as its diameter. I.e. the length and width are equal. The sensor is centered under the light pipe. It is pointing straight up. The cut end is reasonably smooth, but I did not polish it or anything.

I modified the resistors yesterday and something is not right. I have to bench test the sensor to see what is wrong. I had been using a 1K resistor between the sensor and the inverting input. The feedback resistor was 10K, the non-inverting input was connected directly to ground and I had a 2.2K resistor between the output and the ISS.

I was finding that I was only getting to ~600w/m2 in full bright, and there was a very slight positive offset in the dark.

I deleted the 2.2K resistor, add a 1K resistor between the non-inverting input and ground and increased the feedback resistor to 12.1K.

Last night I saw a significant positive offset (244w/m2)! I am not sure what the peak reading will be today, but it sort of does not matter given the large offset!

-Tony

ADDENDUM:

I pulled out the circuitboard and cound a stupid error. The non-inverting input was getting ground directly, then the 1k resistor was passing the ground through it to the rest of the circuit. I.E. there was high impedance in the ground. I only had to re-route the ground to everything else directly, then thru the 1K resistor to the non-inverting input.
WOOPS! been there...done that!!! Comes under the heading of "pay attention"
With the 12.1K resistor, I am getting just over 1000w/m2 at high noon.
now THAT'S more like it!!! It will be interesting to see what the dark reading is now! If it isn't 0 it will probably be real close!
This seems like a reasonable result. I have no idea of the temperature effects on this reading, unfortunately. I painted the top of the enclosure white yesterday. It seems to have lowered the heating of the enclosure somewhat.
IF nothing else, a lower temp. will help the life of the parts.

-Tony

[Spamiam]

I _am_ getting a dark reading of zero. It peaked today at 1209. I am attaching the graph as displayed on the console. It is a flat-topped curve. The readings I have seen from real solar sensors are more peaked. I think this is because they are looking for a "sine response", which, I guess, corresponds to actual power per square meter.

I designed the little transulcent light pipe to be as tall as it is wide so that it does NOT have a sine response. I was intending to have a measure of "brightness" rather than total power. I would have thought that it was the same, but not necessarily? The power is brightness * sine(theta)....

To get a more "sine" response, I think that a shorter light pipe would be better. Maybe VERY short, like virtually flush?

-Tony

Attached thumbnail(s)

 

[pomonabill220]

I _am_ getting a dark reading of zero. It peaked today at 1209. I am attaching the graph as displayed on the console. It is a flat-topped curve. The readings I have seen from real solar sensors are more peaked. I think this is because they are looking for a "sine response", which, I guess, corresponds to actual power per square meter.

I designed the little transulcent light pipe to be as tall as it is wide so that it does NOT have a sine response. I was intending to have a measure of "brightness" rather than total power. I would have thought that it was the same, but not necessarily? The power is brightness * sine(theta)....

To get a more "sine" response, I think that a shorter light pipe would be better. Maybe VERY short, like virtually flush?

-Tony

I just wonder if sunlight is being captured through the sides and being detected during the mid morning and afternoon, giving you the flat top? You could try and wrap some opaque tape around the sides of the light pipe that is sticking above the top of your enclosure just to try it, and see what the response looks like.
Was the sky clear all day... i.e. you were getting full sunlight?
I STILL need to get my detectors so I can play too! I did scavange a photo diode from an old laser printer (the scan detector diode for timing). I just don't know what the spectral response is. I get about 350 mV in full sunlight so an amp with about a gain of 10 (approx.) would do it. I might give it a try just to see.
Bill M

[Spamiam]

I just wonder if sunlight is being captured through the sides and being detected during the mid morning and afternoon, giving you the flat top?

Yes, that is correct. I specifically intended for low angle light to be measured the same as high angles by making the "pipe" as tall as it is wide. I suppose I should shorten it, but I am nit oositive how much. I presule it shouod be rather short. so that light coming in from the side is not going to cause and reading.

-Tony

[pomonabill220]

Yes, that is correct. I specifically intended for low angle light to be measured the same as high angles by making the "pipe" as tall as it is wide. I suppose I should shorten it, but I am nit oositive how much. I presule it shouod be rather short. so that light coming in from the side is not going to cause and reading.

-Tony

Hi Tony... been away for a while but NOT forgotten!!!
I just bought a VP2 for my dad (he should get it Wednesday 6/25) BUT he doesn't know it !!! hehehe
I will find out if he wants to upgrade to a UV and/or Solar sensor and will get the detectors and build a couple of boards for the both of us.
How are your's going? Had any further luck with the low end?
Bill

[Spamiam]

Yes, that is correct. I specifically intended for low angle light to be measured the same as high angles by making the "pipe" as tall as it is wide. I suppose I should shorten it, but I am nit oositive how much. I presule it shouod be rather short. so that light coming in from the side is not going to cause and reading.

-Tony

Sorry for not getting back to you sooner. I did not get any email notice that there was a new posting on this thread.

Anyway, the solar sensor is working very well. I get up to a reading of 1400 on a really bright day. The low end is fine.

Overall I am very happy with the functioning of the home-made sensor. There does not appear to be any problem with very high interior tamperatures of the sensor. It is a sealed box, painted white, though.

-Tony

[pomonabill220]

Sorry for not getting back to you sooner. I did not get any email notice that there was a new posting on this thread.

Anyway, the solar sensor is working very well. I get up to a reading of 1400 on a really bright day. The low end is fine.

Overall I am very happy with the functioning of the home-made sensor. There does not appear to be any problem with very high interior tamperatures of the sensor. It is a sealed box, painted white, though.

-Tony

Hi Tony;
Well I have been away since my last post, and kept meaning to get back and see how your sensor is working. Good to hear that it is working! and that heat has not caused a problem.
I finally broke down and purchased a couple of UV and visible photodiodes in SMT form, and requested some samples of dual rail to rail opamps from Linear Devices.
Now I need to make the PC board and try it. I am going to make a set for my father for his VP2. He is estatic about his station! I didn't know that he has been "drooling" (his words) for a Davis station, for about 20 years!!!! AND NOW HE HAS ONE!!! Made points there, huh!
Let you know how things turn out with the sensors and upload some pics too. It will be a while because I have things to do around the house. I live alone, but they qualify for honey do's.
Keep in touch!
Bill

[Jim18655]

What were the final results? I've been working with this circuit this week and can't get the night time readings below 250. Have a 100k trim pot R1 and 470 for the R2. Still trying to get the output over 350 in the day. It has been cloudy all week. Gap in data is from having the sensor on the work bench.

This post has been edited by Jim18655: May 24 2011, 09:39 PM