Login to AccuWeather.com Premium Login to AccuWeather.com Professional Login to AccuWeather.com RadarPlus AccuWeather.com

Welcome Guest ( Log In | Register )

2 Pages V   1 2 >  
Reply to this topicStart new topic
> Homemade solar sensor, Looking for advice on making my own solar sensor
Spamiam
post Jan 15 2008, 05:39 PM
Post #1




Rank: Whirlwind
*

Group: Member
Posts: 18
Joined: 13-January 08
Member No.: 12,430





I want to add a solar sensor to my Vantage 2. But I want to avoid the cost of the official Davis unit since I am not interested in absolutely accurate readings, just accurate and linear with respect to itself.

I have read the specifications of the solar sensor (3v supply, 2 ground wires and one output wire). I have the parts to make a nice op-amp unit, and I have a decent photodiode for the sensor. But, I am not sure about the specifics of the sensor voltage output vs. illumination

Has anyone made their own sensor? If I were designing this device, I would have the sensor output a logarithmic function of the illumination in order to expand the dynamic range. On the receiving side, I would exponentiate teh value to restore the original reading.

It is easy to have a logarithmic output from the sensor, but is this what Davis does?

I prototyped my own linear output and logarithmic output versions. By looking at the graph on the display, it seems to have a very strong peak during the day when I use the linear output. THis suggests to me that the Davis sensor has a logarithmic output.

Any ideas? If I have any success, a sensor that is "relatively calibrated" would cost less than $20.

-Tony
Go to the top of the page
 
+Quote Post
wxtech
post Jan 31 2008, 08:11 AM
Post #2




Rank: Whirlwind
*

Group: Founding Member
Posts: 7
Joined: 5-September 07
From: Lexington, Ga.
Member No.: 8,698





QUOTE(Spamiam @ Jan 15 2008, 05:39 PM) *
I want to add a solar sensor to my Vantage 2. But I want to avoid the cost of the official Davis unit since I am not interested in absolutely accurate readings, just accurate and linear with respect to itself.

I have read the specifications of the solar sensor (3v supply, 2 ground wires and one output wire). I have the parts to make a nice op-amp unit, and I have a decent photodiode for the sensor. But, I am not sure about the specifics of the sensor voltage output vs. illumination

Has anyone made their own sensor? If I were designing this device, I would have the sensor output a logarithmic function of the illumination in order to expand the dynamic range. On the receiving side, I would exponentiate teh value to restore the original reading.

It is easy to have a logarithmic output from the sensor, but is this what Davis does?

I prototyped my own linear output and logarithmic output versions. By looking at the graph on the display, it seems to have a very strong peak during the day when I use the linear output. THis suggests to me that the Davis sensor has a logarithmic output.

Any ideas? If I have any success, a sensor that is "relatively calibrated" would cost less than $20.

-Tony

I cringe at the thought of replacing my solar or UV sensors. My UV sensor failed within a month of purchase and Davis replaced it.
I have thought of building my own, but calibrating the homebuilt was my problem.
I could also build a unit and operate it alongside with the Davis sensors while checking compatibility.
My solar and UV graphs are on my site: www.alwashington.com
Go to the top of the page
 
+Quote Post
wxtech
post Feb 1 2008, 08:13 PM
Post #3




Rank: Whirlwind
*

Group: Founding Member
Posts: 7
Joined: 5-September 07
From: Lexington, Ga.
Member No.: 8,698





QUOTE(Spamiam @ Jan 15 2008, 05:39 PM) *
I have a decent photodiode for the sensor. -Tony


What type of sensor do you have/recommend?
Go to the top of the page
 
+Quote Post
Spamiam
post May 8 2008, 07:55 AM
Post #4




Rank: Whirlwind
*

Group: Member
Posts: 18
Joined: 13-January 08
Member No.: 12,430





QUOTE(wxtech @ Feb 1 2008, 09:13 PM) *
What type of sensor do you have/recommend?


I am using the CLD140 sensor (http://www.hobby-boards.com/catalog/product_info.php?cPath=26&products_id=102).

I used a MCP604 single supply rail-to-rail op amp (it is a quad device and I only used one channel. A single or dual device would be smaller). Almost any single supply rail-to-rail op amp will do.

I used a transistor in the feedback loop to give a logarithmic response.

Right now my feedback resistor is too high and the output saturates. Since I do not need absolutely accurate measurements, all I need to do is to get the output to stay below 1700 on the brightest days.

I used a small box to house everything. I will paint the box silver to minimize solar heating, but I am trying to think of a good way to allow air circulation without allowing too much moisture or insects in.

I used a translucent plastic dowel to act as a "light pipe". I have it protruding the equivalent of one diameter out of the box. The diameter is about the same as the overall diameter of the photodiode.

It works well, and it only draws a small current for those few moments when the ISS is taking a reading. with the logarithmic response, I can tell the difference between an overcast night and a clear night. It reads about 200 at dusk, and about 20 when it is really dark. Right now, on a clear day it hits the maximum reading of about 1765 at about 10am and stays that way until at least 2pm. On rainy days it reads less than 1000.

-Tony
Go to the top of the page
 
+Quote Post
pomonabill220
post May 24 2008, 03:58 PM
Post #5




Rank: Whirlwind
*

Group: Member
Posts: 25
Joined: 23-May 08
From: Pomona, CA.
Member No.: 14,902





[quote name='Spamiam' date='May 8 2008, 05:55 AM' post='134246']
I am using the CLD140 sensor (http://www.hobby-boards.com/catalog/product_info.php?cPath=26&products_id=102).

I used a MCP604 single supply rail-to-rail op amp (it is a quad device and I only used one channel. A single or dual device would be smaller). Almost any single supply rail-to-rail op amp will do.

I used a transistor in the feedback loop to give a logarithmic response.

Right now my feedback resistor is too high and the output saturates. Since I do not need absolutely accurate measurements, all I need to do is to get the output to stay below 1700 on the brightest days.
----snip

How do you input the analog output into the ISS? Does the ISS expect an analog voltage?
Go to the top of the page
 
+Quote Post
CentralOhio
post May 28 2008, 03:20 PM
Post #6




Rank: Whirlwind
*

Group: Member
Posts: 4
Joined: 21-December 07
Member No.: 11,906





I am interested in adding this also. Can you provide more details?
Go to the top of the page
 
+Quote Post
Spamiam
post May 28 2008, 06:15 PM
Post #7




Rank: Whirlwind
*

Group: Member
Posts: 18
Joined: 13-January 08
Member No.: 12,430





QUOTE(CentralOhio @ May 28 2008, 04:20 PM) *
I am interested in adding this also. Can you provide more details?


I can post the info here. I do not see a way to attach a schematic, but I will investigate this further. I had been trying to design a circuit to have a logarithmic output, AND have a simple circuit. Well, I have been failing on this. It seems that I need to have 2 op-amps in the circuit. One to do the logarithmic conversion, and one to do the offset compensation and amplification. I am finding it hard to achieve the goal of simplicity because I do not have a negative voltage available (just +3.3v). There is an added complication that the temperature offset of the transistor used to make the logarithmic conversion will have a big amplification too (about 30x in the current configuration).

The Davis data sheet indicates that their sensor had 1.67mV per W/m2. So this seems to indicate that thier sensor has a linear output.... I had poor results when I tried to use a linear output. It took some considerable light to register ANY reading above zero. Maybe with my box which should improve sensitivity to oblique angles of illumination will fix this. I will make some tests of a linear system.

In a nutshell, I am using an MCP601 or MCP602 op-amp. It is a rail-to-rail, single supply op-amp. It will function at 3.3v.

I use a Clairex photodiode (CL-140?). I have the photodiode oriented so that it will send a NEGATIVE voltage to the inverting input through a 2.2K resistor. There is a 10K or so resistor in the feedback loop to the inverting input. The non-inverting input goes to ground through 2.2K resistor.

I have a 0.1uF decoupling capacitor across the supply at the op-amp. I have a 0.01uF capacitor between the output and the inverting input.

The feedback resistor is the variable to be changed to set the sensitivity. You want to get to >1000W/m2 in FULL FULL sunlight. If you anly want relative results, then set the maximum reading to close to 1500-1600 in the brightest light you can get. Now is a good time for this in the northern hemisphere.

-Tony
Go to the top of the page
 
+Quote Post
pomonabill220
post May 28 2008, 11:08 PM
Post #8




Rank: Whirlwind
*

Group: Member
Posts: 25
Joined: 23-May 08
From: Pomona, CA.
Member No.: 14,902





If anyone is interested, I profiled the UV input voltage vs. VP2 indication. I used a precision 10 turn pot, a sampling DVM and waited 50 seconds between samples. The ISS turns on the +3 volts every 50 seconds for about 2 seconds when it takes a sample. The voltage for 0.0 reading starts at abou 70 mV and full scale tops out at 2.400 volts.
Anything lower than 70 mV still reads 0.0 and above 2.400 still reads 16.0.
The points were odd voltages and UV indications, but when I enter them into an Excel spreadsheet, I can plot the overall response.
I would imagine that the A/D converter in the ISS between different units might be slightly different (tolerance), but this is a starting point.
I will post my results when I get them tabulate. I took 36 data points so the results should be fairly accurate.
I imagine the solar input might be similar, but for sunburn and skin damage, I thought the UV was more important, but the solar is next.
I am also finding several different companies make UV detectors that are visible and IR blind so no filter is needed!!! Perfect!!!
More later.........
Go to the top of the page
 
+Quote Post
Spamiam
post May 29 2008, 08:22 AM
Post #9




Rank: Whirlwind
*

Group: Member
Posts: 18
Joined: 13-January 08
Member No.: 12,430





I, for one, will be VERY interested to see what the solar sensor output graph looks like! Since I have an old oscilloscope, and a digital voltmeter, I had a hard time determining exactly what the voltage input to the sensor was. I assumed 3.3V or just slightly less.

It is helpful to know that there looks to be about a 70mV offset in the A/D conversion. I can build-in a 70mV offset if necessary.

I am assuming that the sensor's output is roughly linear (rather than logarithmic). If I were desinging it, I would probably have used a log output so that twilight would be read-able too.

I will work on a linear amplifier and attach the schematic when I have something working reasonably well. I have designed a functional log converter, but it is too complicated and subject to too much temperature variability, AND it appears that Davis uses a linear response anyway.

-Tony

QUOTE(pomonabill220 @ May 29 2008, 12:08 AM) *
If anyone is interested, I profiled the UV input voltage vs. VP2 indication. I used a precision 10 turn pot, a sampling DVM and waited 50 seconds between samples. The ISS turns on the +3 volts every 50 seconds for about 2 seconds when it takes a sample. The voltage for 0.0 reading starts at abou 70 mV and full scale tops out at 2.400 volts.
Anything lower than 70 mV still reads 0.0 and above 2.400 still reads 16.0.
The points were odd voltages and UV indications, but when I enter them into an Excel spreadsheet, I can plot the overall response.
I would imagine that the A/D converter in the ISS between different units might be slightly different (tolerance), but this is a starting point.
I will post my results when I get them tabulate. I took 36 data points so the results should be fairly accurate.
I imagine the solar input might be similar, but for sunburn and skin damage, I thought the UV was more important, but the solar is next.
I am also finding several different companies make UV detectors that are visible and IR blind so no filter is needed!!! Perfect!!!
More later.........

Go to the top of the page
 
+Quote Post
CentralOhio
post May 29 2008, 11:38 AM
Post #10




Rank: Whirlwind
*

Group: Member
Posts: 4
Joined: 21-December 07
Member No.: 11,906





QUOTE(Spamiam @ May 29 2008, 09:22 AM) *
I, for one, will be VERY interested to see what the solar sensor output graph looks like! Since I have an old oscilloscope, and a digital voltmeter, I had a hard time determining exactly what the voltage input to the sensor was. I assumed 3.3V or just slightly less.

It is helpful to know that there looks to be about a 70mV offset in the A/D conversion. I can build-in a 70mV offset if necessary.

I am assuming that the sensor's output is roughly linear (rather than logarithmic). If I were desinging it, I would probably have used a log output so that twilight would be read-able too.

I will work on a linear amplifier and attach the schematic when I have something working reasonably well. I have designed a functional log converter, but it is too complicated and subject to too much temperature variability, AND it appears that Davis uses a linear response anyway.

-Tony


Keep us updated. I am quite interested in both the solar and UV sensors.
Go to the top of the page
 
+Quote Post
Spamiam
post May 29 2008, 09:19 PM
Post #11




Rank: Whirlwind
*

Group: Member
Posts: 18
Joined: 13-January 08
Member No.: 12,430





Here is a general schematic for the linear amplifier. R1 and R2 need to be adjusted to suit the overall gain. Increasing R1 will increase the gain. Increasing R2 will decrease the gain. For the most part the gain is equal to R1/R2. The actual gain may not be precisely equal to that ratio, but it will be pretty close.

Probably R2 should be greater than 1K. R3 should be equal to R2.

The schmatic shows the photodiode as the CLD240E, but you can use whatever diode you want, depending on your needs and what is available. Different diodes will need different gains.

THe schematic shows the MCP604 as the opamp. THis is what I had on hand. If I were going to buy an opamp specifically for this purpose, I would use a dual op amp, instead of a quad. Short the unused inputs to ground to reduce power consumption and noise.

Use a small capacitor (e.g. 0.1uF ceramic) across the power terminals of the opamp. Use a small capacitor across the output and inverting input. It helps reduce noise. You probably want something no larger than 0.01uF. The minimum size should be 10pF.

I will get back with my final resistor selection.

-Tony
Attached Image
Go to the top of the page
 
+Quote Post
pomonabill220
post May 30 2008, 04:51 PM
Post #12




Rank: Whirlwind
*

Group: Member
Posts: 25
Joined: 23-May 08
From: Pomona, CA.
Member No.: 14,902





THANKS ALOT FOR THE SCHEMATIC!!!!! Spamiam!!
I had to reacquire the data because my other DVM was being a real B***** (very frustarting after all the time I spent).
I am putting together an Excel Spreadsheet with that data, but FYI...
The input IS linear...
The minimum display value for UV index is 0.5 at 0.070 volts.
The step voltage (from say 1.0 to 1.1 UV index) is 0.0143 volts per step.
The MAX input for 16.0 UV index is 2.395 volts.

The short circuit current from the ISS is 28 mA.
With a 2.67 mA load, the ISS voltage is 2.723 volts
With a 1.383 mA load, the ISS voltage is 2.858 volts
Open ckt. ISS voltage is 3.003 volts.
The ISS samples the UV sensor every 50 seconds for 2 seconds (the 3 volts IS switched NOT constant).
PM me if you would like a copy of the spreadsheet.
Bill
Go to the top of the page
 
+Quote Post
pomonabill220
post May 30 2008, 04:55 PM
Post #13




Rank: Whirlwind
*

Group: Member
Posts: 25
Joined: 23-May 08
From: Pomona, CA.
Member No.: 14,902





Forgot to add that I am going to characterize the solar input next, but I bet it is the same (as far as inearity and range).
Bill
Go to the top of the page
 
+Quote Post
pomonabill220
post May 31 2008, 04:24 PM
Post #14




Rank: Whirlwind
*

Group: Member
Posts: 25
Joined: 23-May 08
From: Pomona, CA.
Member No.: 14,902





So Spamiam.... was this schematic suggested by Clarex for use with there sensor?
I have been looking at others as well and from what I gather from their datasheets, when the sensor is used in photovoltaic mode, the output is very small and you need alot of gain, possibly two opamps: one to invert and two to give the big gain.
The output from the sensor goes to the noninverting input (with the gain) and the other amp inverts.... (from what I remember).
If your matic' works, all the better because it is simpler!
Also, mabey a larger cap in the feedback (like 1uF) could be used to REALLY reduce any HF noise since the change for the UV will be relativley slow?? Whatcha think?
I am still working on the spreadsheet and characterizing the solar input. The solar is a little harder because the 3 volts is only on for 120mSec !!!! and there is a small offset current coming out of the input that has to be delt with because the console starts reading solar at about 0.015 volts!!!
This might be a tough one but I still need to plot my results for linearity.

Let you all know what the outcome is!
Go to the top of the page
 
+Quote Post
pomonabill220
post Jun 2 2008, 09:03 PM
Post #15




Rank: Whirlwind
*

Group: Member
Posts: 25
Joined: 23-May 08
From: Pomona, CA.
Member No.: 14,902





biggrin.gif
Well boys and girls, I have the solar and the UV data all tabulated in Excel, plotted and I reverse engineered the ISS's inputs for solar and UV.
It is pretty simple but there is a gotcha!
The signal input line is diode clamped to gnd and 3.3 volts and there is a 100K pullup to the 3.3volts! This requires a rail to rail amp so the input (most importantly the solar input) gets pulled to gnd.
If you would like the data, please feel free to PM me and I will send you the file (free of viruses, spam, nasty stuff).
Bill
Go to the top of the page
 
+Quote Post
Spamiam
post Jun 3 2008, 05:35 PM
Post #16




Rank: Whirlwind
*

Group: Member
Posts: 18
Joined: 13-January 08
Member No.: 12,430





QUOTE(pomonabill220 @ Jun 2 2008, 10:03 PM) *
biggrin.gif
Well boys and girls, I have the solar and the UV data all tabulated in Excel, plotted and I reverse engineered the ISS's inputs for solar and UV.
It is pretty simple but there is a gotcha!
The signal input line is diode clamped to gnd and 3.3 volts and there is a 100K pullup to the 3.3volts! This requires a rail to rail amp so the input (most importantly the solar input) gets pulled to gnd.
If you would like the data, please feel free to PM me and I will send you the file (free of viruses, spam, nasty stuff).
Bill


Interesting data. The 100K pull-up on the solar sensor output is not a problem. I just want to be clear about the info. The ISS has the 100K pull-up the solar signal internally?

This is a "weak" pull-up. It is not a problem to deal with. Since the op-amp is a reasonably low impedance current source, the pull-up will have no influence. In my original schematic I had a current limiting resistor on the op-amp output. Since it is clear, based on your info, the ISS has sufficient input protection to make that resistor unnecessary. The schematic above does not have that limiting resistor.

Your question about the need for high gain on the op-amp, all I can say is that I have NOT needed a 2 stage amplifier, and I have not needed excessively large resistors.

The CLD240 is rated at 40uA for a 1mW/sq.cm. illumination. This is the equivalent of 10W/m2. Since maximum illumination might be as high as 1000W/m2 (100 times greater), then the current might be as high as 4000uA (4mA). At 3 volts, this would imply a feedback resistor of as little as 750 ohms. Since I would be using an input resistor of 1K to 2.2K, the feedback resistor would be larger. THis is an easily feasible job for 1 op-amp.

The schematic did not come from Clairex, it is just a simple photovoltaic mode op-amp schematic. THe resistor values are the numbers I used in my initial settings for a linear sensor. I have had no problem getting useful voltages with the Clairex photodiode and a single op-amp using resistors of this magnitude.

I just need to fine tune the resistor values.

As for the feedback capacitor, You can't go too large because the output needs to settle within the 120mS that the power is applied. I would want it to settle in 1/10th that time since we can assume that the signal is being sampled repetetively (for smoothing and noise filtration) by the ISS. The 0.01uF cap seems fine. THe sensor readings seem stable. I would hesitate to go with a larger cap. I am not positive of the time constant of that one, except it seems to eliminate the hum from fluorescent lights quite effectively, so it is longer than 1/60th of a second be a good margin.

I will PM you for the data.

WHat was the voltage input to the ISS that gave the maximum reading on the solar reading?

-Tony

This post has been edited by Spamiam: Jun 3 2008, 05:40 PM
Go to the top of the page
 
+Quote Post
pomonabill220
post Jun 3 2008, 11:38 PM
Post #17




Rank: Whirlwind
*

Group: Member
Posts: 25
Joined: 23-May 08
From: Pomona, CA.
Member No.: 14,902





QUOTE(CentralOhio @ May 29 2008, 09:38 AM) *
Keep us updated. I am quite interested in both the solar and UV sensors.


Well the system will not let me upload this file so I will have to come up with another way to get the info to you.... sorry
Go to the top of the page
 
+Quote Post
pomonabill220
post Jun 3 2008, 11:57 PM
Post #18




Rank: Whirlwind
*

Group: Member
Posts: 25
Joined: 23-May 08
From: Pomona, CA.
Member No.: 14,902





QUOTE(Spamiam @ Jun 3 2008, 03:35 PM) *
Interesting data. The 100K pull-up on the solar sensor output is not a problem. I just want to be clear about the info. The ISS has the 100K pull-up the solar signal internally?
Yes, the pullup is switched along with the 3 volts. A p channel fet is used to source the 3 volts, and the 100 k pullup (to the signal input) is taken off the 3 volts switched. The gate of the fet comes from the processor in the ISS through a 1 meg resistor, and the processor's output is normally high and goes low to turn on the fet. Each input has it's own fet / input protection network.

This is a "weak" pull-up. It is not a problem to deal with. Since the op-amp is a reasonably low impedance current source, the pull-up will have no influence. In my original schematic I had a current limiting resistor on the op-amp output. Since it is clear, based on your info, the ISS has sufficient input protection to make that resistor unnecessary. The schematic above does not have that limiting resistor.
Yes the 3 volt output, and the signal input is "protected" by a 100 ohm resistor so it is not a "hard" line and semi-shortcircuit proofed. Really there to help isolate the outside world.

Your question about the need for high gain on the op-amp, all I can say is that I have NOT needed a 2 stage amplifier, and I have not needed excessively large resistors.

The CLD240 is rated at 40uA for a 1mW/sq.cm. illumination. This is the equivalent of 10W/m2. Since maximum illumination might be as high as 1000W/m2 (100 times greater), then the current might be as high as 4000uA (4mA). At 3 volts, this would imply a feedback resistor of as little as 750 ohms. Since I would be using an input resistor of 1K to 2.2K, the feedback resistor would be larger. THis is an easily feasible job for 1 op-amp.
That's true... MORE than enough output from the photodiode in photovoltaic mode for reasonable gain amp.

The schematic did not come from Clairex, it is just a simple photovoltaic mode op-amp schematic. THe resistor values are the numbers I used in my initial settings for a linear sensor. I have had no problem getting useful voltages with the Clairex photodiode and a single op-amp using resistors of this magnitude.

I just need to fine tune the resistor values.

As for the feedback capacitor, You can't go too large because the output needs to settle within the 120mS that the power is applied. I would want it to settle in 1/10th that time since we can assume that the signal is being sampled repetetively (for smoothing and noise filtration) by the ISS. The 0.01uF cap seems fine. THe sensor readings seem stable. I would hesitate to go with a larger cap. I am not positive of the time constant of that one, except it seems to eliminate the hum from fluorescent lights quite effectively, so it is longer than 1/60th of a second be a good margin.
True. Just long enough time constant to attenuate any 60Hz "noise" from being amplified and detected. Seems kinda strange that the on time for the UV is at least 2 seconds (maybe 2.5), and the solar is only 120 mSec. Must be at the end of a A/D mux cycle.
I will PM you for the data.

WHat was the voltage input to the ISS that gave the maximum reading on the solar reading?
Max solar was 2.977 for 1782 displayed... never went any higher.
Max UV was 2.395 for 16.0 UV index display.
These readings were taken with a HP 3466A DVM using a 10 turn pot to simulate the input voltages and wwaaaatiinng for the console to take a sample (at least 2). There was no integrating of either input so the displayed values were instant. If the input changed drastically, the console would take 2 sample cycles before it displayed them.
What I can do is scan my schematic and spreadsheets (since the system will not let me attach the files) and send them to you. If you can let me know how to attach the files and send them to you, I would appreciate it.
Thanks.... Bill
-Tony
Go to the top of the page
 
+Quote Post
Spamiam
post Jun 4 2008, 12:37 PM
Post #19




Rank: Whirlwind
*

Group: Member
Posts: 18
Joined: 13-January 08
Member No.: 12,430





QUOTE(pomonabill220)
Yes, the pullup is switched along with the 3 volts. A p channel fet is used to source the 3 volts, and the 100 k pullup (to the signal input) is taken off the 3 volts switched. The gate of the fet comes from the processor in the ISS through a 1 meg resistor, and the processor's output is normally high and goes low >to turn on the fet. Each input has it's own fet / input protection network.

Yes the 3 volt output, and the signal input is "protected" by a 100 ohm resistor so it is not a "hard" line and semi-shortcircuit proofed. Really there to help isolate the outside world.


I am glad that you opened up the ISS to look at the innards. I was hesitant to do so. This is really useful information to be used when making plans on the final design of the solar sensor. E.G. A current limiting resistor on the output is not necessary because the ISS already has one.

It is also nice to know that the Solar Sensor power is strobed for 120mS. I was not sure of the actual time, but I knew it was pretty short because my meter wouldn't even settle on the voltage before the power was cut back off! I am not sure whether the 0.1uF bypass cap on the op-amp is going to be a problem. Based on your description of the ISS, I think the bypass cap will charge up fully in the required time. I just wonder if all the energy lost in the cap is a bad thing for the ISS power budget. Maybe a smaller bypass cap would be adequate, but 0.1uF is a pretty standard component and I hesitate to change it.

QUOTE
Max solar was 2.977 for 1782 displayed... never went any higher.
Max UV was 2.395 for 16.0 UV index display.


Isn't the highest possible insolation about 1000W/m2? If so, there is a lot uf unused range on the ISS solar input. 1000W equates to a sensor output of about 1.67v. I wonder what the resolution is for the D-A converter. You had found the UV input has a step voltage of 0.0143 volts. This would give only 210 steps over a 3V range. From this I would suspect that the ISS uses nothing better than a 10 bit A-D converter. This would give 1024 steps which may get divided down to 512 or 256 steps. Assuming 512 steps in the full 3V range, then the 0 to 1000W per sq. meter insolation range would use 287 steps. Not a whole lot considering that bright light is well over 1000x times brigher than dim light. This is why I was interested in a logarithmic response curve for the sensor, it mimics the response of the eye more closely than a linear curve....

Since I am not interested in absolute accuracy of the sensor, just relatively accurate compared to itself, I will make use of the full 0 - 1782 range. BTW, the largest value I was getting was about 1765.

I will try to make some changes to the sensor today to eliminate the transistor (to make a linear response), and get some initial resistors in the circuit.

-Tony
Go to the top of the page
 
+Quote Post
pomonabill220
post Jun 4 2008, 10:07 PM
Post #20




Rank: Whirlwind
*

Group: Member
Posts: 25
Joined: 23-May 08
From: Pomona, CA.
Member No.: 14,902





QUOTE(Spamiam @ Jun 4 2008, 10:37 AM) *
I am glad that you opened up the ISS to look at the innards. I was hesitant to do so. This is really useful information to be used when making plans on the final design of the solar sensor. E.G. A current limiting resistor on the output is not necessary because the ISS already has one.

It is also nice to know that the Solar Sensor power is strobed for 120mS. I was not sure of the actual time, but I knew it was pretty short because my meter wouldn't even settle on the voltage before the power was cut back off! I am not sure whether the 0.1uF bypass cap on the op-amp is going to be a problem. Based on your description of the ISS, I think the bypass cap will charge up fully in the required time. I just wonder if all the energy lost in the cap is a bad thing for the ISS power budget. Maybe a smaller bypass cap would be adequate, but 0.1uF is a pretty standard component and I hesitate to change it.
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. I would guess the cap in the ISS is a 0.01uF... I did not remove it to test it as all the parts are SMT parts and the board is conformal coated.
I should clarify the 120 mSec. There was a small change between strobe cycles. I sort of remember the shortest time around 110 mSec and the longest was mabey 135 mSec. I was not too concerned at getting the absolute times, but I do remeber a small change. I was using a Tek storage scope (7834) so I was able to store the times, just didn't delve into it.
Isn't the highest possible insolation about 1000W/m2? If so, there is a lot uf unused range on the ISS solar input. 1000W equates to a sensor output of about 1.67v. I wonder what the resolution is for the D-A converter. You had found the UV input has a step voltage of 0.0143 volts. This would give only 210 steps over a 3V range. From this I would suspect that the ISS uses nothing better than a 10 bit A-D converter. This would give 1024 steps which may get divided down to 512 or 256 steps. Assuming 512 steps in the full 3V range, then the 0 to 1000W per sq. meter insolation range would use 287 steps. Not a whole lot considering that bright light is well over 1000x times brigher than dim light. This is why I was interested in a logarithmic response curve for the sensor, it mimics the response of the eye more closely than a linear curve....
It could be that the designers limited the overall range so that unit to unit's tolerance would cover a "usable range" and individual calibration would not be necessary.
Since I am not interested in absolute accuracy of the sensor, just relatively accurate compared to itself, I will make use of the full 0 - 1782 range. BTW, the largest value I was getting was about 1765.
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.
I will try to make some changes to the sensor today to eliminate the transistor (to make a linear response), and get some initial resistors in the circuit.
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?
Bill
-Tony
Go to the top of the page
 
+Quote Post

2 Pages V   1 2 >
Reply to this topicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 

RSS Lo-Fi Version Time is now: 20th August 2014 - 03:44 AM