Making a phototransistor behave equally in sunlight

[ACharnley]

Just after some quick thoughts.

I've a photo transistor which was specifically chosen to block infrared. This it appears to do, however during day[sun]light the read is double that of a head-torch shining on the sensor. As a quick test I held a lighter flame close and it blows out the sensor read.

I'm thinking I need to filter the red and possibly yellow component, so a blue filter. Does this sound correct?

 

[Nigel Goodwin]

Just after some quick thoughts.

I've a photo transistor which was specifically chosen to block infrared. This it appears to do, however during day[sun]light the read is double that of a head-torch shining on the sensor. As a quick test I held a lighter flame close and it blows out the sensor read.

I'm thinking I need to filter the red and possibly yellow component, so a blue filter. Does this sound correct?

Sun light is massively brighter than a torch, so it sounds like it's behaving exactly as it's supposed to.

 

[schmitt trigger]

Why would you want to filter out some colors? The sunlight is full spectrum, whereas any artificial light, specifically the fluorescent and LED ones, will have significant spectral gaps.
Incandescent lamps are also continuous spectrum, but leaning very heavily towards the red.

 

[schmitt trigger]

Figure 4 you posted indicates that the detector’s maximum sensitivity occurs around 625 nm, which is a reddish-orange. Small wonder that a flame causes it to react.

 

[ACharnley]

Yeah, and it's not reacting the same under what the eye thinks is the same brightness indoors vs outdoors. Whence I thought to filter the redy orange out.

Cant recall having this with LDR's but then I checked and they usually react more towards the blue wavelength.

 

[ACharnley]

Sunlight is misleading, even overcast at dusk it's at 100% while in the garage under an led strip light it's at 50% ish. Both I would say are about equal brightness wise so it has to be spectrum.

 

[schmitt trigger]

Yes. It is the spectrum content. As a fact, the CIE defines the 6500K daylight white as the standard illuminant.

 

[rjenkinsgb]

Remember that your eye response is approximately logarithmic, not linear; subjective "twice as bright" is around 5 - 6x greater illumination.

 

[Nigel Goodwin]

Yeah, and it's not reacting the same under what the eye thinks is the same brightness indoors vs outdoors.

The eye is wrong, sunlight is massively brighter than artificial light. The eye, like the ear, has automatic gain control.

There was a TV series a few ago called 'Techno Games'. It was for school teams, and in fact I recognised an adult I knew assisting one of the teams (a guy called John, a serious genius!). There were various different events that the teams had to build robots to compete in, and one was a solar challenge race.

Basically they were solar powered model electric cars, no batteries allowed - and raced along an indoor course. In order for it to work, as it wasn't outside in the sun, the entire space above the course was completely filled with bright floodlights. When turned on, it looked like if you walked underneath you would be shrivelled to a crisp

Just had a quick google:

Techno Games - Wikipedia

en.wikipedia.org

 

[ACharnley]

Indeed, so if the blue filter idea falls flat I'm not sure how I go about resolving it. Other than knowing that it's indoors, which is impossible.

 

[Nigel Goodwin]

Indeed, so if the blue filter idea falls flat I'm not sure how I go about resolving it. Other than knowing that it's indoors, which is impossible.

What are you trying to do?, outside is brighter than inside, why does that matter?.

 

[ZipZapOuch]

Indeed, so if the blue filter idea falls flat I'm not sure how I go about resolving it. Other than knowing that it's indoors, which is impossible.

Your phototransistor shows fluorescent and incandescent light response up to 10k lux, but, to go with what Nigel is saying, direct sunlight will be in the 40k-100k lux range. Google: solar lux, to get relative values for overcast, indirect sunlight, etc. almost all are well above average room light levels. For reference, OSHA recommends the light level in an office be in the range of 300 to 500 lux. Anything over 1000 would be a noticeably bright room.

An option would be a signal compression circuit or an automatic gain control circuit.

A blue or other filter will attenuate everything and squash your indoor signal. Also, the spectral response of blue light is low. "Cool" (bluish 6500K) LEDs give off a lot of blue light while most Western Europeans and Americans tend to like incandescent-like "warm" (2700K-4200K) LEDs which give off very little blue light.

 

[Tony Stewart]

If you want a sensor to match CIE eye response then say so. There are many examples. Some are log gain analog others linear I2C

 

[ZipZapOuch]

If you want a sensor to match CIE eye response then say so.

Thats not what he was asking for.