Detecting only daylight

[dcom12]

Hello everybody ,
What sensor do I need to detect only daylight like the sensors on the cars for automatic headlight ?

Thanks in advance

 

[prestonee]

you could use a simple photo diode.
basically its resistance changes as a function of light. This is typically used as the sensor in night lights, you notice that if there is a slight shadow the nightlight will kick on a little but not very bright. you can make it more switch like by controlling the voltage divider and amplifying and running to a comparator. hope this helps
-Pb

 

[dcom12]

Yes but i want not to detect the artificial light ... only daylight ?

 

[prestonee]

hmm if this is the case then what you need to do is a bit more complex, I think even automatic car lights will be triggered off with artificial light(that is why the sensor is usually on the top and front of the dash so it wont be triggered while driving, but if you drive into a lit tunnel, the lights will turn off because the sensor will detect the tunnel lights). You would need to replace the photo diode with a UV sensor.
A warning though: the reason auto car light sensors do not detect true sunlight is because UV does not pass through glass. so if yo make a true sunlight detector and encase it in glass or in a car, it wont be triggered.
Goodluck
-Pb

 

[dcom12]

What sensor to use for these purpose then ? More sens to daylight ?

 

[prestonee]

Sunlight is made of visible light (same as artificial), infrared(heat, but the sky is mostly absent of infrared due to the sub zero temperature of space), and UV(also from black lights).
Your only real option is to detect visible or UV. You might try to use a UV sensor that detects NUV, near ultraviolet, which passes with more ease through glass(350 to 450nm length).
-Pb

-fixed typo of um to nm-

 

[FvM]

I don't believe that the suggested detection methods work.

There's no easy way to distinguish daylight clearly from a collection of different artificial light sources, e.g. incandescend light, HID, white LED, CFL. They have all different spectral characteristics and NUV or NIR intensity may be lower as well as higher than daylight. A resolved spectrum can show some specific properties, e.g. absorption bands of the atmosphere only achieved with km optical path length. But I guess that a spectrometer is beyond the project scope.

 

[ads_ee]

infrared(heat, but the sky is mostly absent of infrared due to the sub zero temperature of space)

Most of the energy from the sun is infrared...
https://en.wikipedia.org/wiki/Infrared
This infrared is what drives a large portion of the weather on the earth.

 

[sky_123]

Or, you could just check the time, and assume there is daylight during certain hours.
I suspect that wasn't what you wanted though!
Or, a video camera and some smart intelligent software to interpret the image.

 

[prestonee]

Yes IR is majority of the suns energy we receive, but I worked in IR the sky is usually used as a cold reference because even though the sun is IR it is surrounded by cold space. its liek looking at the moon in the pitch of night, a sensor will only pick up the night not the moon. In other words the IR spectrum is overwhelmed by the space(as in outer space). If truely trying to detect sunlight vs visible your only affordable option is detecting UV. you can buy a UV sensor easily from digikey or mouser. jsut make not of its wv sensitivity and make sure its within the band of interest, as I said before glass blocks UV at frequencies below ~350nm..quartz however is much better at allowing UV to pass through...or you can buy yourself a spectrometer for a small fortune or use a video camera with software approach... your call.
-Pb

 

[FvM]

Most of the energy from the sun is infrared...

Roughly 50% of the radiation power arriving at sea level. But that's not specific for sunlight. Incandescendent lamps have an even larger IR share due to lower temperature.

If truely trying to detect sunlight vs visible your only affordable option is detecting UV.

How to distinguish it without detailed spectral measurements from CFL or HID light, which have considerable UV A radiation?

 

[pplus]

I'm not sure that the problem is put correctly. Characteristics of sunlight change a lot of time of day and weather conditions. Artificial light sources may have the characteristics of light similar to daylight. May be sufficient simply to measure the illumination environment?

 

[FvM]

I'm not sure that the problem is put correctly.

At least not clearly.

 

[prestonee]

How to distinguish it without detailed spectral measurements from CFL or HID light, which have considerable UV A radiation?

You cant, but the differentiation between sunlight and most regular lighting may be what he wants.
And I believe all HID and CFL lights are required to be designed with a UV shield around them, so they shouldnt be an issue.
You can verify this but because of the harmful nature of UV the only UV source we should be subject to is the sun and the tanning bed(thanks to health codes etc)
I think defective CFL will be the biggest worry, the curving of the glass causes imperfections with the coating allowing some UV to pass through, that is why there is a large push to LED, they do not create any UV.
but again the uv from the defective CFL should be less then that of the sun by a significant amount.
-Pb

 

[pplus]

Your only real option is to detect visible or UV. You might try to use a UV sensor that detects NUV, near ultraviolet, which passes with more ease through glass(350 to 450nm length).

It is a good idea
UV-B Monitoring and Research Program
Respect.

PS The length of daylight varies greatly.

 

[FvM]

And I believe all HID and CFL lights are required to be designed with a UV shield around them, so they shouldnt be an issue.

They are using regular (borosilicate) glass which still has limited UV-A transmission. Actually there's a big difference between CFL and HID. CFL have a low pressure mercury spectrum, and the dominant 254 nm line is completely absorbed by the glass tube. But there's still a less intense 370 nm line and possible NUV fluorescense. HID has a more solar like wide band spectrum, and it's surely exposing some UV-A. Indoor, or behind a car window, it won't be distinguishable from solar light very easily.

All-in-all I only want to abate wrong expectations. But we didn't hear a clear project specification yet, and apparently the O.P. isn't motivated to clarify things.

 

[prestonee]

I would be interested in seeing the results of this circuit.

 

[pplus]

Emission Spectra of Common Light Sources


I'm also working on this issue of self-interest, with no practical application. I think that in most cases, you can identify artificial or natural light. The question is, why it is needed and to whom.

 

[kripacharya]

after ALL this vast knowledge about spectrum etc etc has been discussed and evaluated, i would like to add my 2-bits.

I believe the answer is quite simple - the intensity of daylight vs any natural light are quite different altogether. So a detector simply needs to be calibrated for a luminosity level - and trigger (i.e. switch-on the headlights) if the lighting is below that level. Isn't that what headlights are for ?

Implementation will have to ensure that the sensor does not pick-up the light from the headlights themselves. Thats all.

cheers!!

 

[prestonee]

We do not know the application it is to be used in, also this level of detection would affect the response at say dusk and dawn and cloudy days etc. but yes if hes ok with some false triggers (ie headlights) or no triggers then dialing the threshold is pretty straight forward.( I think my 1st post was toward this manner)
-Pb